From e216094e2192c05ddbd0988458eb8c0012e7baf8 Mon Sep 17 00:00:00 2001 From: Vicent Marti Date: Tue, 2 Sep 2014 01:10:54 +0200 Subject: lol --- src/buffer.h | 119 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 119 insertions(+) create mode 100644 src/buffer.h (limited to 'src/buffer.h') diff --git a/src/buffer.h b/src/buffer.h new file mode 100644 index 0000000..2581ee3 --- /dev/null +++ b/src/buffer.h @@ -0,0 +1,119 @@ +#ifndef INCLUDE_buffer_h__ +#define INCLUDE_buffer_h__ + +#include +#include +#include +#include + +typedef struct { + unsigned char *ptr; + int asize, size; +} gh_buf; + +extern unsigned char gh_buf__initbuf[]; +extern unsigned char gh_buf__oom[]; + +#define GH_BUF_INIT { gh_buf__initbuf, 0, 0 } + +/** + * Initialize a gh_buf structure. + * + * For the cases where GH_BUF_INIT cannot be used to do static + * initialization. + */ +extern void gh_buf_init(gh_buf *buf, int initial_size); + +static inline void gh_buf_static(gh_buf *buf, unsigned char *source) +{ + buf->ptr = source; + buf->size = strlen(source); + buf->asize = -1; +} + +/** + * Attempt to grow the buffer to hold at least `target_size` bytes. + * + * If the allocation fails, this will return an error. If mark_oom is true, + * this will mark the buffer as invalid for future operations; if false, + * existing buffer content will be preserved, but calling code must handle + * that buffer was not expanded. + */ +extern int gh_buf_try_grow(gh_buf *buf, int target_size, bool mark_oom); + +/** + * Grow the buffer to hold at least `target_size` bytes. + * + * If the allocation fails, this will return an error and the buffer will be + * marked as invalid for future operations, invaliding contents. + * + * @return 0 on success or -1 on failure + */ +static inline int gh_buf_grow(gh_buf *buf, int target_size) +{ + return gh_buf_try_grow(buf, target_size, true); +} + +extern void gh_buf_free(gh_buf *buf); +extern void gh_buf_swap(gh_buf *buf_a, gh_buf *buf_b); + +/** + * Test if there have been any reallocation failures with this gh_buf. + * + * Any function that writes to a gh_buf can fail due to memory allocation + * issues. If one fails, the gh_buf will be marked with an OOM error and + * further calls to modify the buffer will fail. Check gh_buf_oom() at the + * end of your sequence and it will be true if you ran out of memory at any + * point with that buffer. + * + * @return false if no error, true if allocation error + */ +static inline bool gh_buf_oom(const gh_buf *buf) +{ + return (buf->ptr == gh_buf__oom); +} + + +static inline size_t gh_buf_len(const gh_buf *buf) +{ + return buf->size; +} + +extern int gh_buf_cmp(const gh_buf *a, const gh_buf *b); + +extern void gh_buf_attach(gh_buf *buf, char *ptr, int asize); +extern char *gh_buf_detach(gh_buf *buf); +extern void gh_buf_copy_cstr(char *data, int datasize, const gh_buf *buf); + +static inline const char *gh_buf_cstr(const gh_buf *buf) +{ + return buf->ptr; +} + +#define gh_buf_at(buf, n) ((buf)->ptr[n]) + +/* + * Functions below that return int value error codes will return 0 on + * success or -1 on failure (which generally means an allocation failed). + * Using a gh_buf where the allocation has failed with result in -1 from + * all further calls using that buffer. As a result, you can ignore the + * return code of these functions and call them in a series then just call + * gh_buf_oom at the end. + */ +extern int gh_buf_set(gh_buf *buf, const char *data, int len); +extern int gh_buf_sets(gh_buf *buf, const char *string); +extern int gh_buf_putc(gh_buf *buf, char c); +extern int gh_buf_put(gh_buf *buf, const char *data, int len); +extern int gh_buf_puts(gh_buf *buf, const char *string); +extern int gh_buf_printf(gh_buf *buf, const char *format, ...) + __attribute__((format (printf, 2, 3))); +extern int gh_buf_vprintf(gh_buf *buf, const char *format, va_list ap); +extern void gh_buf_clear(gh_buf *buf); + +int gh_buf_strchr(const gh_buf *buf, int c, int pos); +int gh_buf_strrchr(const gh_buf *buf, int c, int pos); +void gh_buf_truncate(gh_buf *buf, int len); +void gh_buf_ltruncate(gh_buf *buf, int len); +void gh_buf_trim(gh_buf *buf); + +#endif -- cgit v1.2.3 From 582674e662d1f8757350c51486a5e0a837195e15 Mon Sep 17 00:00:00 2001 From: Vicent Marti Date: Tue, 2 Sep 2014 13:18:04 +0200 Subject: ffffix --- Makefile | 11 ++- src/blocks.c | 58 +++++++---- src/buffer.c | 69 +++++-------- src/buffer.h | 19 ++-- src/html.c | 276 ---------------------------------------------------- src/inlines.c | 4 +- src/main.c | 142 ++++++++++++--------------- src/print.c | 307 ++++++++++++++++++++++++++++++---------------------------- src/stmd.h | 13 +-- src/utf8.c | 6 +- src/utf8.h | 6 -- 11 files changed, 304 insertions(+), 607 deletions(-) delete mode 100644 src/html.c delete mode 100644 src/utf8.h (limited to 'src/buffer.h') diff --git a/Makefile b/Makefile index cb5938d..d14a928 100644 --- a/Makefile +++ b/Makefile @@ -6,7 +6,7 @@ DATADIR=data PROG=./stmd .PHONY: all oldtests test spec benchjs testjs -all: $(SRCDIR)/case_fold_switch.c $(PROG) +all: $(SRCDIR)/case_fold_switch.inc $(PROG) README.html: README.md template.html pandoc --template template.html -S -s -t html5 -o $@ $< @@ -41,13 +41,16 @@ testjs: spec.txt benchjs: node js/bench.js ${BENCHINP} -$(PROG): $(SRCDIR)/main.c $(SRCDIR)/inlines.o $(SRCDIR)/buffer.o $(SRCDIR)/blocks.o $(SRCDIR)/scanners.c $(SRCDIR)/print.o $(SRCDIR)/html.o $(SRCDIR)/utf8.o +HTML_OBJ=$(SRCDIR)/html/html.o $(SRCDIR)/html/houdini_href_e.o $(SRCDIR)/html/houdini_html_e.o +STMD_OBJ=$(SRCDIR)/inlines.o $(SRCDIR)/buffer.o $(SRCDIR)/blocks.o $(SRCDIR)/scanners.c $(SRCDIR)/print.o $(SRCDIR)/utf8.o + +$(PROG): $(SRCDIR)/main.c $(HTML_OBJ) $(STMD_OBJ) $(CC) $(LDFLAGS) -o $@ $^ $(SRCDIR)/scanners.c: $(SRCDIR)/scanners.re re2c --case-insensitive -bis $< > $@ || (rm $@ && false) -$(SRCDIR)/case_fold_switch.inc $(DATADIR)/CaseFolding-3.2.0.txt +$(SRCDIR)/case_fold_switch.inc: $(DATADIR)/CaseFolding-3.2.0.txt perl mkcasefold.pl < $< > $@ .PHONY: leakcheck clean fuzztest dingus upload @@ -72,7 +75,7 @@ update-site: spec.html narrative.html (cd _site ; git pull ; git commit -a -m "Updated site for latest spec, narrative, js" ; git push; cd ..) clean: - -rm -f test $(SRCDIR)/*.o $(SRCDIR)/scanners.c + -rm -f test $(SRCDIR)/*.o $(SRCDIR)/scanners.c $(SRCDIR)/html/*.o -rm -rf *.dSYM -rm -f README.html -rm -f spec.md fuzz.txt spec.html diff --git a/src/blocks.c b/src/blocks.c index eabac03..71dc830 100644 --- a/src/blocks.c +++ b/src/blocks.c @@ -3,11 +3,12 @@ #include #include #include -#include "bstrlib.h" + #include "stmd.h" -#include "uthash.h" -#include "debug.h" #include "scanners.h" +#include "uthash.h" + +static void finalize(block* b, int line_number); static block* make_block(int tag, int start_line, int start_column) { @@ -140,7 +141,7 @@ static int break_out_of_lists(block ** bptr, int line_number) } -extern void finalize(block* b, int line_number) +static void finalize(block* b, int line_number) { int firstlinelen; int pos; @@ -364,7 +365,7 @@ static int lists_match(struct ListData list_data, list_data.bullet_char == item_data.bullet_char); } -static void expand_tabs(gh_buf *ob, const char *line, size_t size) +static void expand_tabs(gh_buf *ob, const unsigned char *line, size_t size) { size_t i = 0, tab = 0; @@ -389,13 +390,43 @@ static void expand_tabs(gh_buf *ob, const char *line, size_t size) } } -extern block *stmd_parse_document(const char *buffer, size_t len) +static block *finalize_parsing(block *document, int linenum) { - gh_buf line = GH_BUF_INIT; + while (document != document->top) { + finalize(document, linenum); + document = document->parent; + } + + finalize(document, linenum); + process_inlines(document, document->attributes.refmap); + + return document; +} +extern block *stmd_parse_file(FILE *f) +{ + gh_buf line = GH_BUF_INIT; + unsigned char buffer[4096]; + int linenum = 1; block *document = make_document(); + + while (fgets((char *)buffer, sizeof(buffer), f)) { + expand_tabs(&line, buffer, strlen(buffer)); + incorporate_line(&line, linenum, &document); + gh_buf_clear(&line); + linenum++; + } + + gh_buf_free(&line); + return finalize_document(document, linenum); +} + +extern block *stmd_parse_document(const unsigned char *buffer, size_t len) +{ + gh_buf line = GH_BUF_INIT; int linenum = 1; - const char *end = buffer + len; + const unsigned char *end = buffer + len; + block *document = make_document(); while (buffer < end) { const char *eol = memchr(buffer, '\n', end - buffer); @@ -414,16 +445,7 @@ extern block *stmd_parse_document(const char *buffer, size_t len) } gh_buf_free(&line); - - while (document != document->top) { - finalize(document, linenum); - document = document->parent; - } - - finalize(document, linenum); - process_inlines(document, document->attributes.refmap); - - return document; + return finalize_document(document, linenum); } // Process one line at a time, modifying a block. diff --git a/src/buffer.c b/src/buffer.c index b81e7fa..17dc864 100644 --- a/src/buffer.c +++ b/src/buffer.c @@ -31,10 +31,10 @@ void gh_buf_init(gh_buf *buf, int initial_size) int gh_buf_try_grow(gh_buf *buf, int target_size, bool mark_oom) { - char *new_ptr; - size_t new_size; + unsigned char *new_ptr; + int new_size; - if (buf->ptr == gh_buf__oom || buf->asize < 0) + if (buf->ptr == gh_buf__oom) return -1; if (target_size <= buf->asize) @@ -79,7 +79,7 @@ void gh_buf_free(gh_buf *buf) { if (!buf) return; - if (buf->asize > 0 && buf->ptr != gh_buf__initbuf && buf->ptr != gh_buf__oom) + if (buf->ptr != gh_buf__initbuf && buf->ptr != gh_buf__oom) free(buf->ptr); gh_buf_init(buf, 0); @@ -91,14 +91,9 @@ void gh_buf_clear(gh_buf *buf) if (buf->asize > 0) buf->ptr[0] = '\0'; - - if (buf->asize < 0) { - buf->ptr = gh_buf__initbuf; - buf->asize = 0; - } } -int gh_buf_set(gh_buf *buf, const char *data, int len) +int gh_buf_set(gh_buf *buf, const unsigned char *data, int len) { if (len == 0 || data == NULL) { gh_buf_clear(buf); @@ -115,10 +110,12 @@ int gh_buf_set(gh_buf *buf, const char *data, int len) int gh_buf_sets(gh_buf *buf, const char *string) { - return gh_buf_set(buf, string, string ? strlen(string) : 0); + return gh_buf_set(buf, + (const unsigned char *)string, + string ? strlen(string) : 0); } -int gh_buf_putc(gh_buf *buf, char c) +int gh_buf_putc(gh_buf *buf, int c) { ENSURE_SIZE(buf, buf->size + 2); buf->ptr[buf->size++] = c; @@ -126,7 +123,7 @@ int gh_buf_putc(gh_buf *buf, char c) return 0; } -int gh_buf_put(gh_buf *buf, const char *data, int len) +int gh_buf_put(gh_buf *buf, const unsigned char *data, int len) { ENSURE_SIZE(buf, buf->size + len + 1); memmove(buf->ptr + buf->size, data, len); @@ -137,8 +134,7 @@ int gh_buf_put(gh_buf *buf, const char *data, int len) int gh_buf_puts(gh_buf *buf, const char *string) { - assert(string); - return gh_buf_put(buf, string, strlen(string)); + return gh_buf_put(buf, (const unsigned char *)string, strlen(string)); } int gh_buf_vprintf(gh_buf *buf, const char *format, va_list ap) @@ -153,7 +149,7 @@ int gh_buf_vprintf(gh_buf *buf, const char *format, va_list ap) va_copy(args, ap); len = vsnprintf( - buf->ptr + buf->size, + (char *)buf->ptr + buf->size, buf->asize - buf->size, format, args ); @@ -187,9 +183,9 @@ int gh_buf_printf(gh_buf *buf, const char *format, ...) return r; } -void gh_buf_copy_cstr(char *data, size_t datasize, const gh_buf *buf) +void gh_buf_copy_cstr(char *data, int datasize, const gh_buf *buf) { - size_t copylen; + int copylen; assert(data && datasize && buf); @@ -212,9 +208,9 @@ void gh_buf_swap(gh_buf *buf_a, gh_buf *buf_b) *buf_b = t; } -char *gh_buf_detach(gh_buf *buf) +unsigned char *gh_buf_detach(gh_buf *buf) { - char *data = buf->ptr; + unsigned char *data = buf->ptr; if (buf->asize == 0 || buf->ptr == gh_buf__oom) return NULL; @@ -224,13 +220,13 @@ char *gh_buf_detach(gh_buf *buf) return data; } -void gh_buf_attach(gh_buf *buf, char *ptr, int asize) +void gh_buf_attach(gh_buf *buf, unsigned char *ptr, int asize) { gh_buf_free(buf); if (ptr) { buf->ptr = ptr; - buf->size = strlen(ptr); + buf->size = strlen((char *)ptr); if (asize) buf->asize = (asize < buf->size) ? buf->size + 1 : asize; else /* pass 0 to fall back on strlen + 1 */ @@ -249,11 +245,11 @@ int gh_buf_cmp(const gh_buf *a, const gh_buf *b) int gh_buf_strchr(const gh_buf *buf, int c, int pos) { - const char *p = memchr(buf->ptr + pos, c, buf->size - pos); - if (!p) - return -1; + const char *p = memchr(buf->ptr + pos, c, buf->size - pos); + if (!p) + return -1; - return (int)(p - p->ptr); + return (int)(p - buf->ptr); } int gh_buf_strrchr(const gh_buf *buf, int c, int pos) @@ -270,36 +266,21 @@ int gh_buf_strrchr(const gh_buf *buf, int c, int pos) void gh_buf_truncate(gh_buf *buf, size_t len) { - assert(buf->asize >= 0); - if (len < buf->size) { buf->size = len; buf->ptr[buf->size] = '\0'; } } -void gh_buf_ltruncate(gh_buf *buf, size_t len) -{ - assert(buf->asize >= 0); - - if (len && len < buf->size) { - memmove(buf->ptr, buf->ptr + len, buf->size - len); - buf->size -= len; - buf->ptr[buf->size] = '\0'; - } -} - void gh_buf_trim(gh_buf *buf) { - size_t i = 0; - - assert(buf->asize >= 0); - - /* ltrim */ + /* TODO: leading whitespace? */ + /* while (i < buf->size && isspace(buf->ptr[i])) i++; gh_buf_truncate(buf, i); + */ /* rtrim */ while (buf->size > 0) { diff --git a/src/buffer.h b/src/buffer.h index 2581ee3..422ef02 100644 --- a/src/buffer.h +++ b/src/buffer.h @@ -24,13 +24,6 @@ extern unsigned char gh_buf__oom[]; */ extern void gh_buf_init(gh_buf *buf, int initial_size); -static inline void gh_buf_static(gh_buf *buf, unsigned char *source) -{ - buf->ptr = source; - buf->size = strlen(source); - buf->asize = -1; -} - /** * Attempt to grow the buffer to hold at least `target_size` bytes. * @@ -81,13 +74,13 @@ static inline size_t gh_buf_len(const gh_buf *buf) extern int gh_buf_cmp(const gh_buf *a, const gh_buf *b); -extern void gh_buf_attach(gh_buf *buf, char *ptr, int asize); -extern char *gh_buf_detach(gh_buf *buf); +extern void gh_buf_attach(gh_buf *buf, unsigned char *ptr, int asize); +extern unsigned char *gh_buf_detach(gh_buf *buf); extern void gh_buf_copy_cstr(char *data, int datasize, const gh_buf *buf); static inline const char *gh_buf_cstr(const gh_buf *buf) { - return buf->ptr; + return (char *)buf->ptr; } #define gh_buf_at(buf, n) ((buf)->ptr[n]) @@ -100,10 +93,10 @@ static inline const char *gh_buf_cstr(const gh_buf *buf) * return code of these functions and call them in a series then just call * gh_buf_oom at the end. */ -extern int gh_buf_set(gh_buf *buf, const char *data, int len); +extern int gh_buf_set(gh_buf *buf, const unsigned char *data, int len); extern int gh_buf_sets(gh_buf *buf, const char *string); -extern int gh_buf_putc(gh_buf *buf, char c); -extern int gh_buf_put(gh_buf *buf, const char *data, int len); +extern int gh_buf_putc(gh_buf *buf, int c); +extern int gh_buf_put(gh_buf *buf, const unsigned char *data, int len); extern int gh_buf_puts(gh_buf *buf, const char *string); extern int gh_buf_printf(gh_buf *buf, const char *format, ...) __attribute__((format (printf, 2, 3))); diff --git a/src/html.c b/src/html.c deleted file mode 100644 index aeec5f1..0000000 --- a/src/html.c +++ /dev/null @@ -1,276 +0,0 @@ -#include -#include -#include -#include "bstrlib.h" -#include "stmd.h" -#include "debug.h" -#include "scanners.h" - -// Functions to convert block and inline lists to HTML strings. - -// Escape special characters in HTML. More efficient than -// three calls to bfindreplace. If preserve_entities is set, -// existing entities are left alone. -static bstring escape_html(bstring inp, bool preserve_entities) -{ - int pos = 0; - int match; - char c; - bstring escapable = blk2bstr("&<>\"", 4); - bstring ent; - bstring s = bstrcpy(inp); - while ((pos = binchr(s, pos, escapable)) != BSTR_ERR) { - c = bchar(s,pos); - switch (c) { - case '<': - bdelete(s, pos, 1); - ent = blk2bstr("<", 4); - binsert(s, pos, ent, ' '); - bdestroy(ent); - pos += 4; - break; - case '>': - bdelete(s, pos, 1); - ent = blk2bstr(">", 4); - binsert(s, pos, ent, ' '); - bdestroy(ent); - pos += 4; - break; - case '&': - if (preserve_entities && (match = scan_entity(s, pos))) { - pos += match; - } else { - bdelete(s, pos, 1); - ent = blk2bstr("&", 5); - binsert(s, pos, ent, ' '); - bdestroy(ent); - pos += 5; - } - break; - case '"': - bdelete(s, pos, 1); - ent = blk2bstr(""", 6); - binsert(s, pos, ent, ' '); - bdestroy(ent); - pos += 6; - break; - default: - bdelete(s, pos, 1); - log_err("unexpected character %02x", c); - } - } - bdestroy(escapable); - return s; -} - -static inline void cr(bstring buffer) -{ - int c = bchar(buffer, blength(buffer) - 1); - if (c != '\n' && c) { - bconchar(buffer, '\n'); - } -} - -// Convert a block list to HTML. Returns 0 on success, and sets result. -extern int blocks_to_html(block* b, bstring* result, bool tight) -{ - bstring contents = NULL; - bstring escaped, escaped2; - struct bstrList * info_words; - struct ListData * data; - bstring mbstart; - bstring html = blk2bstr("", 0); - - while(b != NULL) { - switch(b->tag) { - case document: - check(blocks_to_html(b->children, &contents, false) == 0, - "error converting blocks to html"); - bformata(html, "%s", contents->data); - bdestroy(contents); - break; - case paragraph: - check(inlines_to_html(b->inline_content, &contents) == 0, - "error converting inlines to html"); - if (tight) { - bformata(html, "%s", contents->data); - } else { - cr(html); - bformata(html, "

%s

", contents->data); - cr(html); - } - bdestroy(contents); - break; - case block_quote: - check(blocks_to_html(b->children, &contents, false) == 0, - "error converting blocks to html"); - cr(html); - bformata(html, "
\n%s
", contents->data); - cr(html); - bdestroy(contents); - break; - case list_item: - check(blocks_to_html(b->children, &contents, tight) == 0, - "error converting blocks to html"); - brtrimws(contents); - cr(html); - bformata(html, "
  • %s
    • ", contents->data); - cr(html); - bdestroy(contents); - break; - case list: - // make sure a list starts at the beginning of the line: - cr(html); - data = &(b->attributes.list_data); - check(blocks_to_html(b->children, &contents, data->tight) == 0, - "error converting blocks to html"); - mbstart = bformat(" start=\"%d\"", data->start); - bformata(html, "<%s%s>\n%s", - data->list_type == bullet ? "ul" : "ol", - data->start == 1 ? "" : (char*) mbstart->data, - contents->data, - data->list_type == bullet ? "ul" : "ol"); - cr(html); - bdestroy(contents); - bdestroy(mbstart); - break; - case atx_header: - case setext_header: - check(inlines_to_html(b->inline_content, &contents) == 0, - "error converting inlines to html"); - cr(html); - bformata(html, "%s", - b->attributes.header_level, - contents->data, - b->attributes.header_level); - cr(html); - bdestroy(contents); - break; - case indented_code: - escaped = escape_html(b->string_content, false); - cr(html); - bformata(html, "
    %s
    ", escaped->data); - cr(html); - bdestroy(escaped); - break; - case fenced_code: - escaped = escape_html(b->string_content, false); - cr(html); - bformata(html, "
    attributes.fenced_code_data.info) > 0) {
-        escaped2 = escape_html(b->attributes.fenced_code_data.info, true);
-        info_words = bsplit(escaped2, ' ');
-        bformata(html, " class=\"language-%s\"", info_words->entry[0]->data);
-        bdestroy(escaped2);
-        bstrListDestroy(info_words);
-      }
-      bformata(html, ">%s
    ", escaped->data); - cr(html); - bdestroy(escaped); - break; - case html_block: - bformata(html, "%s", b->string_content->data); - break; - case hrule: - bformata(html, "
    "); - cr(html); - break; - case reference_def: - break; - default: - log_warn("block type %d not implemented\n", b->tag); - break; - } - b = b->next; - } - *result = html; - return 0; - error: - return -1; -} - -// Convert an inline list to HTML. Returns 0 on success, and sets result. -extern int inlines_to_html(inl* ils, bstring* result) -{ - bstring contents = NULL; - bstring html = blk2bstr("", 0); - bstring mbtitle, escaped, escaped2; - - while(ils != NULL) { - switch(ils->tag) { - case str: - escaped = escape_html(ils->content.literal, false); - bformata(html, "%s", escaped->data); - bdestroy(escaped); - break; - case linebreak: - bformata(html, "
    \n"); - break; - case softbreak: - bformata(html, "\n"); - break; - case code: - escaped = escape_html(ils->content.literal, false); - bformata(html, "%s", escaped->data); - bdestroy(escaped); - break; - case raw_html: - case entity: - bformata(html, "%s", ils->content.literal->data); - break; - case link: - check(inlines_to_html(ils->content.inlines, &contents) == 0, - "error converting inlines to html"); - if (blength(ils->content.linkable.title) > 0) { - escaped = escape_html(ils->content.linkable.title, true); - mbtitle = bformat(" title=\"%s\"", escaped->data); - bdestroy(escaped); - } else { - mbtitle = blk2bstr("",0); - } - escaped = escape_html(ils->content.linkable.url, true); - bformata(html, "%s", - escaped->data, - mbtitle->data, - contents->data); - bdestroy(escaped); - bdestroy(mbtitle); - bdestroy(contents); - break; - case image: - check(inlines_to_html(ils->content.inlines, &contents) == 0, - "error converting inlines to html"); - escaped = escape_html(ils->content.linkable.url, true); - escaped2 = escape_html(contents, false); - bdestroy(contents); - bformata(html, "\"%s\"",data, escaped2->data); - bdestroy(escaped); - bdestroy(escaped2); - if (blength(ils->content.linkable.title) > 0) { - escaped = escape_html(ils->content.linkable.title, true); - bformata(html, " title=\"%s\"", escaped->data); - bdestroy(escaped); - } - bformata(html, " />"); - break; - case strong: - check(inlines_to_html(ils->content.inlines, &contents) == 0, - "error converting inlines to html"); - bformata(html, "%s", contents->data); - bdestroy(contents); - break; - case emph: - check(inlines_to_html(ils->content.inlines, &contents) == 0, - "error converting inlines to html"); - bformata(html, "%s", contents->data); - bdestroy(contents); - break; - } - ils = ils->next; - } - *result = html; - return 0; - error: - return -1; -} diff --git a/src/inlines.c b/src/inlines.c index 4ff45ad..82c7219 100644 --- a/src/inlines.c +++ b/src/inlines.c @@ -6,9 +6,7 @@ #include "stmd.h" #include "uthash.h" -#include "debug.h" #include "scanners.h" -#include "utf8.h" typedef struct Subject { const gh_buf *buffer; @@ -119,7 +117,7 @@ inline static inl* make_linkable(int t, inl* label, chunk url, chunk title) e->tag = t; e->content.linkable.label = label; e->content.linkable.url = chunk_to_cstr(&url); - e->content.linkable.title = chunk_to_cstr(&title); + e->content.linkable.title = url.len ? chunk_to_cstr(&title) : NULL; e->next = NULL; return e; } diff --git a/src/main.c b/src/main.c index 9e0a3c8..e1abedc 100644 --- a/src/main.c +++ b/src/main.c @@ -1,99 +1,77 @@ #include #include -#include "bstrlib.h" +#include #include "stmd.h" #include "debug.h" void print_usage() { - printf("Usage: stmd [FILE*]\n"); - printf("Options: --help, -h Print usage information\n"); - printf(" --ast Print AST instead of HTML\n"); - printf(" --version Print version\n"); + printf("Usage: stmd [FILE*]\n"); + printf("Options: --help, -h Print usage information\n"); + printf(" --ast Print AST instead of HTML\n"); + printf(" --version Print version\n"); } -int main(int argc, char *argv[]) { - int i; - bool ast = false; - int g = 0; - int numfps = 0; - int files[argc]; +static void print_document(block *document, bool ast) +{ + gh_buf html = GH_BUF_INIT; + + if (ast) { + print_blocks(document, 0); + } else { + blocks_to_html(&html, document, false); + printf("%s", html.ptr); + gh_buf_free(&html); + } +} - for (i=1; i < argc; i++) { - if (strcmp(argv[i], "--version") == 0) { - printf("stmd %s", VERSION); - printf(" - CommonMark converter (c) 2014 John MacFarlane\n"); - exit(0); - } else if ((strcmp(argv[i], "--help") == 0) || - (strcmp(argv[i], "-h") == 0)) { - print_usage(); - exit(0); - } else if (strcmp(argv[i], "--ast") == 0) { - ast = true; - } else if (*argv[i] == '-') { - print_usage(); - exit(1); - } else { // treat as file argument - files[g] = i; - g++; - } - } +int main(int argc, char *argv[]) +{ + int i, numfps = 0; + bool ast = false; + int files[argc]; + block *document = NULL; - numfps = g; - bstring s = NULL; - bstring html; - g = 0; - block * cur = make_document(); - int linenum = 1; - extern int errno; - FILE * fp = NULL; + for (i = 1; i < argc; i++) { + if (strcmp(argv[i], "--version") == 0) { + printf("stmd %s", VERSION); + printf(" - CommonMark converter (c) 2014 John MacFarlane\n"); + exit(0); + } else if ((strcmp(argv[i], "--help") == 0) || + (strcmp(argv[i], "-h") == 0)) { + print_usage(); + exit(0); + } else if (strcmp(argv[i], "--ast") == 0) { + ast = true; + } else if (*argv[i] == '-') { + print_usage(); + exit(1); + } else { // treat as file argument + files[numfps++] = i; + } + } - if (numfps == 0) { - // read from stdin - while ((s = bgets((bNgetc) fgetc, stdin, '\n'))) { - check(incorporate_line(s, linenum, &cur) == 0, - "error incorporating line %d", linenum); - bdestroy(s); - linenum++; - } - } else { - // iterate over input file pointers - for (g=0; g < numfps; g++) { + if (numfps == 0) { + document = stmd_parse_file(stdin); + print_document(document, ast); + free_blocks(document); + } else { + for (i = 0; i < numfps; i++) { + FILE *fp = fopen(argv[files[i]], "r"); - fp = fopen(argv[files[g]], "r"); - if (fp == NULL) { - fprintf(stderr, "Error opening file %s: %s\n", - argv[files[g]], strerror(errno)); - exit(1); - } + if (fp == NULL) { + fprintf(stderr, "Error opening file %s: %s\n", + argv[files[i]], strerror(errno)); + exit(1); + } - while ((s = bgets((bNgetc) fgetc, fp, '\n'))) { - check(incorporate_line(s, linenum, &cur) == 0, - "error incorporating line %d", linenum); - bdestroy(s); - linenum++; - } - fclose(fp); - } - } + document = stmd_parse_file(fp); + print_document(document, ast); + free_blocks(document); + fclose(fp); + } + } - while (cur != cur->top) { - finalize(cur, linenum); - cur = cur->parent; - } - check(cur == cur->top, "problems finalizing open containers"); - finalize(cur, linenum); - process_inlines(cur, cur->attributes.refmap); - if (ast) { - print_blocks(cur, 0); - } else { - check(blocks_to_html(cur, &html, false) == 0, "could not format as HTML"); - // printf("%s", html->data); - bdestroy(html); - } - free_blocks(cur); - return 0; -error: - return -1; + return 0; } diff --git a/src/print.c b/src/print.c index a924870..3ebde16 100644 --- a/src/print.c +++ b/src/print.c @@ -1,168 +1,175 @@ #include #include -#include "bstrlib.h" +#include #include "stmd.h" #include "debug.h" -static bstring format_str(bstring s) +static void print_str(const unsigned char *s, int len) { - int pos = 0; - int len = blength(s); - bstring result = bfromcstr(""); - char c; - bformata(result, "\""); - while (pos < len) { - c = bchar(s, pos); - switch (c) { - case '\n': - bformata(result, "\\n"); - break; - case '"': - bformata(result, "\\\""); - break; - case '\\': - bformata(result, "\\\\"); - break; - default: - bformata(result, "%c", c); - } - pos++; - } - bformata(result, "\""); - return result; + int i; + + if (len < 0) + len = strlen(s); + + putchar('"'); + for (i = 0; i < len; ++i) { + unsigned char c = s[i]; + + switch (c) { + case '\n': + printf("\\n"); + break; + case '"': + printf("\\\""); + break; + case '\\': + printf("\\\\"); + break; + default: + putchar((int)c); + } + } + putchar('"'); } // Functions to pretty-print inline and block lists, for debugging. // Prettyprint an inline list, for debugging. extern void print_blocks(block* b, int indent) { - struct ListData * data; - while(b != NULL) { - // printf("%3d %3d %3d| ", b->start_line, b->start_column, b->end_line); - for (int i=0; i < indent; i++) { - putchar(' '); - } - switch(b->tag) { - case document: - printf("document\n"); - print_blocks(b->children, indent + 2); - break; - case block_quote: - printf("block_quote\n"); - print_blocks(b->children, indent + 2); - break; - case list_item: - data = &(b->attributes.list_data); - printf("list_item\n"); - print_blocks(b->children, indent + 2); - break; - case list: - data = &(b->attributes.list_data); - if (data->list_type == ordered) { - printf("list (type=ordered tight=%s start=%d delim=%s)\n", - (data->tight ? "true" : "false"), - data->start, - (data->delimiter == parens ? "parens" : "period")); - } else { - printf("list (type=bullet tight=%s bullet_char=%c)\n", - (data->tight ? "true" : "false"), - data->bullet_char); - } - print_blocks(b->children, indent + 2); - break; - case atx_header: - printf("atx_header (level=%d)\n", b->attributes.header_level); - print_inlines(b->inline_content, indent + 2); - break; - case setext_header: - printf("setext_header (level=%d)\n", b->attributes.header_level); - print_inlines(b->inline_content, indent + 2); - break; - case paragraph: - printf("paragraph\n"); - print_inlines(b->inline_content, indent + 2); - break; - case hrule: - printf("hrule\n"); - break; - case indented_code: - printf("indented_code %s\n", format_str(b->string_content)->data); - break; - case fenced_code: - printf("fenced_code length=%d info=%s %s\n", - b->attributes.fenced_code_data.fence_length, - format_str(b->attributes.fenced_code_data.info)->data, - format_str(b->string_content)->data); - break; - case html_block: - printf("html_block %s\n", format_str(b->string_content)->data); - break; - case reference_def: - printf("reference_def\n"); - break; - default: - log_warn("block type %d not implemented\n", b->tag); - break; - } - b = b->next; - } + struct ListData *data; + + while(b != NULL) { + // printf("%3d %3d %3d| ", b->start_line, b->start_column, b->end_line); + for (int i=0; i < indent; i++) { + putchar(' '); + } + + switch(b->tag) { + case document: + printf("document\n"); + print_blocks(b->children, indent + 2); + break; + case block_quote: + printf("block_quote\n"); + print_blocks(b->children, indent + 2); + break; + case list_item: + data = &(b->attributes.list_data); + printf("list_item\n"); + print_blocks(b->children, indent + 2); + break; + case list: + data = &(b->attributes.list_data); + if (data->list_type == ordered) { + printf("list (type=ordered tight=%s start=%d delim=%s)\n", + (data->tight ? "true" : "false"), + data->start, + (data->delimiter == parens ? "parens" : "period")); + } else { + printf("list (type=bullet tight=%s bullet_char=%c)\n", + (data->tight ? "true" : "false"), + data->bullet_char); + } + print_blocks(b->children, indent + 2); + break; + case atx_header: + printf("atx_header (level=%d)\n", b->attributes.header_level); + print_inlines(b->inline_content, indent + 2); + break; + case setext_header: + printf("setext_header (level=%d)\n", b->attributes.header_level); + print_inlines(b->inline_content, indent + 2); + break; + case paragraph: + printf("paragraph\n"); + print_inlines(b->inline_content, indent + 2); + break; + case hrule: + printf("hrule\n"); + break; + case indented_code: + printf("indented_code "); + print_str(b->string_content.ptr, -1); + putchar('\n'); + break; + case fenced_code: + printf("fenced_code length=%d info=", + b->attributes.fenced_code_data.fence_length); + print_str(b->attributes.fenced_code_data.info.ptr, -1); + putchar(' '); + print_str(b->string_content.ptr, -1); + putchar('\n'); + break; + case html_block: + printf("html_block "); + print_str(b->string_content.ptr, -1); + putchar('\n'); + break; + case reference_def: + printf("reference_def\n"); + break; + default: + printf("# NOT IMPLEMENTED (%d)\n", b->tag); + break; + } + b = b->next; + } } // Prettyprint an inline list, for debugging. extern void print_inlines(inl* ils, int indent) { - while(ils != NULL) { - /* - // we add 11 extra spaces for the line/column info - for (int i=0; i < 11; i++) { - putchar(' '); - } - putchar('|'); - putchar(' '); - */ - for (int i=0; i < indent; i++) { - putchar(' '); - } - switch(ils->tag) { - case str: - printf("str %s\n", format_str(ils->content.literal)->data); - break; - case linebreak: - printf("linebreak\n"); - break; - case softbreak: - printf("softbreak\n"); - break; - case code: - printf("code %s\n", format_str(ils->content.literal)->data); - break; - case raw_html: - printf("html %s\n", format_str(ils->content.literal)->data); - break; - case entity: - printf("entity %s\n", format_str(ils->content.literal)->data); - break; - case link: - printf("link url=%s title=%s\n", - format_str(ils->content.linkable.url)->data, - format_str(ils->content.linkable.title)->data); - print_inlines(ils->content.linkable.label, indent + 2); - break; - case image: - printf("image url=%s title=%s\n", - format_str(ils->content.linkable.url)->data, - format_str(ils->content.linkable.title)->data); - print_inlines(ils->content.linkable.label, indent + 2); - break; - case strong: - printf("strong\n"); - print_inlines(ils->content.linkable.label, indent + 2); - break; - case emph: - printf("emph\n"); - print_inlines(ils->content.linkable.label, indent + 2); - break; - } - ils = ils->next; - } + while(ils != NULL) { + for (int i=0; i < indent; i++) { + putchar(' '); + } + switch(ils->tag) { + case str: + printf("str "); + print_str(ils->content.literal.data, ils->content.literal.len); + putchar('\n'); + break; + case linebreak: + printf("linebreak\n"); + break; + case softbreak: + printf("softbreak\n"); + break; + case code: + printf("code "); + print_str(ils->content.literal.data, ils->content.literal.len); + putchar('\n'); + break; + case raw_html: + printf("html "); + print_str(ils->content.literal.data, ils->content.literal.len); + putchar('\n'); + break; + case entity: + printf("entity "); + print_str(ils->content.literal.data, ils->content.literal.len); + putchar('\n'); + break; + case link: + case image: + printf("%s url=", ils->tag == link ? "link" : "image"); + print_str(ils->content.linkable.url, -1); + if (ils->content.linkable.title) { + printf(" title="); + print_str(ils->content.linkable.title, -1); + } + putchar('\n'); + print_inlines(ils->content.linkable.label, indent + 2); + break; + case strong: + printf("strong\n"); + print_inlines(ils->content.linkable.label, indent + 2); + break; + case emph: + printf("emph\n"); + print_inlines(ils->content.linkable.label, indent + 2); + break; + } + ils = ils->next; + } } diff --git a/src/stmd.h b/src/stmd.h index eb1b989..dc24235 100644 --- a/src/stmd.h +++ b/src/stmd.h @@ -105,19 +105,14 @@ extern block* add_child(block* parent, int block_type, int start_line, int start_column); void free_blocks(block* e); -block *stmd_parse_document(const char *buffer, size_t len); - -// FOR NOW: -void process_inlines(block* cur, reference** refmap); -void incorporate_line(gh_buf *ln, int line_number, block** curptr); -void finalize(block* b, int line_number); +extern block *stmd_parse_document(const unsigned char *buffer, size_t len); +extern block *stmd_parse_file(FILE *f); void print_inlines(inl* ils, int indent); void print_blocks(block* blk, int indent); -/* TODO */ -// int blocks_to_html(block* b, bstring* result, bool tight); -// int inlines_to_html(inl* b, bstring* result); +void blocks_to_html(gh_buf *html, block *b, bool tight); +void inlines_to_html(gh_buf *html, inl *b); void utf8proc_case_fold(gh_buf *dest, const unsigned char *str, int len); diff --git a/src/utf8.c b/src/utf8.c index 1a5df9e..e3f8dd3 100644 --- a/src/utf8.c +++ b/src/utf8.c @@ -1,6 +1,8 @@ #include -#include "bstrlib.h" -#include "debug.h" +#include +#include + +#include "stmd.h" static const int8_t utf8proc_utf8class[256] = { 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, diff --git a/src/utf8.h b/src/utf8.h deleted file mode 100644 index fe59a90..0000000 --- a/src/utf8.h +++ /dev/null @@ -1,6 +0,0 @@ -#include -#include "bstrlib.h" - -extern unsigned char * from_utf8(unsigned char * s, unsigned int *n); -extern int to_utf8(unsigned int c, bstring dest); -extern bstring case_fold(bstring source); -- cgit v1.2.3 From a7314deae649646f1f7ce5ede972641b5b62538c Mon Sep 17 00:00:00 2001 From: Vicent Marti Date: Wed, 3 Sep 2014 03:40:23 +0200 Subject: 338/103 --- Makefile | 4 +- src/blocks.c | 173 +++++++++++++++++----------------- src/buffer.c | 26 ++++- src/buffer.h | 2 +- src/html/houdini_href_e.c | 10 +- src/html/houdini_html_e.c | 10 +- src/html/html.c | 4 +- src/inlines.c | 235 ++++++++++++++++++---------------------------- src/print.c | 2 +- src/scanners.h | 28 +++--- src/scanners.re | 85 +++++++---------- src/stmd.h | 16 ++-- 12 files changed, 261 insertions(+), 334 deletions(-) (limited to 'src/buffer.h') diff --git a/Makefile b/Makefile index d14a928..89ec68c 100644 --- a/Makefile +++ b/Makefile @@ -1,5 +1,5 @@ -CFLAGS=-ggdb3 -O0 -Wall -Wextra -std=c99 -Isrc $(OPTFLAGS) -LDFLAGS=-ggdb3 -O0 -Wall -Werror +CFLAGS=-ggdb3 -O0 -Wall -Wextra -Wno-unused-variable -std=c99 -Isrc $(OPTFLAGS) +LDFLAGS=-ggdb3 -O0 -Wall -Wno-unused-variable # -Werror SRCDIR=src DATADIR=data diff --git a/src/blocks.c b/src/blocks.c index 42f20db..94ff986 100644 --- a/src/blocks.c +++ b/src/blocks.c @@ -8,6 +8,8 @@ #include "scanners.h" #include "uthash.h" +#define peek_at(i, n) (i)->data[n] + static void incorporate_line(gh_buf *ln, int line_number, block** curptr); static void finalize(block* b, int line_number); @@ -27,7 +29,6 @@ static block* make_block(int tag, int start_line, int start_column) e->top = NULL; e->attributes.refmap = NULL; gh_buf_init(&e->string_content, 32); - e->string_pos = 0; e->inline_content = NULL; e->next = NULL; e->prev = NULL; @@ -80,10 +81,10 @@ static inline bool accepts_lines(int block_type) block_type == fenced_code); } -static void add_line(block* block, gh_buf *ln, int offset) +static void add_line(block* block, chunk *ch, int offset) { assert(block->open); - gh_buf_put(&block->string_content, ln->ptr + offset, ln->size - offset); + gh_buf_put(&block->string_content, ch->data + offset, ch->len - offset); } static void remove_trailing_blank_lines(gh_buf *ln) @@ -104,7 +105,7 @@ static void remove_trailing_blank_lines(gh_buf *ln) i = gh_buf_strchr(ln, '\n', i); if (i >= 0) - gh_buf_truncate(ln, i + 1); + gh_buf_truncate(ln, i); } // Check to see if a block ends with a blank line, descending @@ -162,12 +163,12 @@ static void finalize(block* b, int line_number) switch (b->tag) { case paragraph: pos = 0; - while (gh_buf_at(&b->string_content, b->string_pos) == '[' && - (pos = parse_reference(&b->string_content, b->string_pos, - b->top->attributes.refmap))) { - b->string_pos = pos; + while (gh_buf_at(&b->string_content, 0) == '[' && + (pos = parse_reference(&b->string_content, b->top->attributes.refmap))) { + + gh_buf_drop(&b->string_content, pos); } - if (is_blank(&b->string_content, b->string_pos)) { + if (is_blank(&b->string_content, 0)) { b->tag = reference_def; } break; @@ -179,14 +180,16 @@ static void finalize(block* b, int line_number) case fenced_code: // first line of contents becomes info - firstlinelen = gh_buf_strchr(&b->string_content, '\n', b->string_pos); + firstlinelen = gh_buf_strchr(&b->string_content, '\n', 0); + + gh_buf_init(&b->attributes.fenced_code_data.info, 0); gh_buf_set( &b->attributes.fenced_code_data.info, - b->string_content.ptr + b->string_pos, + b->string_content.ptr, firstlinelen ); - b->string_pos = firstlinelen + 1; + gh_buf_drop(&b->string_content, firstlinelen + 1); gh_buf_trim(&b->attributes.fenced_code_data.info); unescape_buffer(&b->attributes.fenced_code_data.info); @@ -281,7 +284,7 @@ void process_inlines(block* cur, reference** refmap) case paragraph: case atx_header: case setext_header: - cur->inline_content = parse_inlines(&cur->string_content, cur->string_pos, refmap); + cur->inline_content = parse_inlines(&cur->string_content, refmap); // MEM // gh_buf_free(&cur->string_content); break; @@ -300,19 +303,18 @@ void process_inlines(block* cur, reference** refmap) // Attempts to parse a list item marker (bullet or enumerated). // On success, returns length of the marker, and populates // data with the details. On failure, returns 0. -static int parse_list_marker(gh_buf *ln, int pos, - struct ListData ** dataptr) +static int parse_list_marker(chunk *input, int pos, struct ListData ** dataptr) { - char c; + unsigned char c; int startpos; struct ListData * data; startpos = pos; - c = gh_buf_at(ln, pos); + c = peek_at(input, pos); - if ((c == '*' || c == '-' || c == '+') && !scan_hrule(ln, pos)) { + if ((c == '*' || c == '-' || c == '+') && !scan_hrule(input, pos)) { pos++; - if (!isspace(gh_buf_at(ln, pos))) { + if (!isspace(peek_at(input, pos))) { return 0; } data = malloc(sizeof(struct ListData)); @@ -327,14 +329,14 @@ static int parse_list_marker(gh_buf *ln, int pos, int start = 0; do { - start = (10 * start) + (gh_buf_at(ln, pos) - '0'); + start = (10 * start) + (peek_at(input, pos) - '0'); pos++; - } while (isdigit(gh_buf_at(ln, pos))); + } while (isdigit(peek_at(input, pos))); - c = gh_buf_at(ln, pos); + c = peek_at(input, pos); if (c == '.' || c == ')') { pos++; - if (!isspace(gh_buf_at(ln, pos))) { + if (!isspace(peek_at(input, pos))) { return 0; } data = malloc(sizeof(struct ListData)); @@ -449,8 +451,26 @@ extern block *stmd_parse_document(const unsigned char *buffer, size_t len) return finalize_document(document, linenum); } +static void chop_trailing_hashtags(chunk *ch) +{ + int n; + + chunk_rtrim(ch); + n = ch->len - 1; + + // if string ends in #s, remove these: + while (n >= 0 && peek_at(ch, n) == '#') + n--; + + // the last # was escaped, so we include it. + if (n >= 0 && peek_at(ch, n) == '\\') + n++; + + ch->len = n + 1; +} + // Process one line at a time, modifying a block. -static void incorporate_line(gh_buf *ln, int line_number, block** curptr) +static void incorporate_line(gh_buf *line, int line_number, block** curptr) { block* last_matched_container; int offset = 0; @@ -464,6 +484,10 @@ static void incorporate_line(gh_buf *ln, int line_number, block** curptr) bool blank = false; int first_nonspace; int indent; + chunk input; + + input.data = line->ptr; + input.len = line->size; // container starts at the document root. container = cur->top; @@ -475,21 +499,19 @@ static void incorporate_line(gh_buf *ln, int line_number, block** curptr) container = container->last_child; first_nonspace = offset; - while (gh_buf_at(ln, first_nonspace) == ' ') { + while (peek_at(&input, first_nonspace) == ' ') { first_nonspace++; } indent = first_nonspace - offset; - blank = gh_buf_at(ln, first_nonspace) == '\n'; + blank = peek_at(&input, first_nonspace) == '\n'; if (container->tag == block_quote) { - - matched = indent <= 3 && gh_buf_at(ln, first_nonspace) == '>'; + matched = indent <= 3 && peek_at(&input, first_nonspace) == '>'; if (matched) { offset = first_nonspace + 1; - if (gh_buf_at(ln, offset) == ' ') { + if (peek_at(&input, offset) == ' ') offset++; - } } else { all_matched = false; } @@ -526,7 +548,7 @@ static void incorporate_line(gh_buf *ln, int line_number, block** curptr) // skip optional spaces of fence offset i = container->attributes.fenced_code_data.fence_offset; - while (i > 0 && gh_buf_at(ln, offset) == ' ') { + while (i > 0 && peek_at(&input, offset) == ' ') { offset++; i--; } @@ -564,15 +586,13 @@ static void incorporate_line(gh_buf *ln, int line_number, block** curptr) container->tag != html_block) { first_nonspace = offset; - while (gh_buf_at(ln, first_nonspace) == ' ') { + while (peek_at(&input, first_nonspace) == ' ') first_nonspace++; - } indent = first_nonspace - offset; - blank = gh_buf_at(ln, first_nonspace) == '\n'; + blank = peek_at(&input, first_nonspace) == '\n'; if (indent >= CODE_INDENT) { - if (cur->tag != paragraph && !blank) { offset += CODE_INDENT; container = add_child(container, indented_code, line_number, offset + 1); @@ -580,76 +600,70 @@ static void incorporate_line(gh_buf *ln, int line_number, block** curptr) break; } - } else if (gh_buf_at(ln, first_nonspace) == '>') { + } else if (peek_at(&input, first_nonspace) == '>') { offset = first_nonspace + 1; // optional following character - if (gh_buf_at(ln, offset) == ' ') { + if (peek_at(&input, offset) == ' ') offset++; - } container = add_child(container, block_quote, line_number, offset + 1); - } else if ((matched = scan_atx_header_start(ln, first_nonspace))) { + } else if ((matched = scan_atx_header_start(&input, first_nonspace))) { offset = first_nonspace + matched; container = add_child(container, atx_header, line_number, offset + 1); - int hashpos = gh_buf_strchr(ln, '#', first_nonspace); - assert(hashpos >= 0); - + int hashpos = chunk_strchr(&input, '#', first_nonspace); int level = 0; - while (gh_buf_at(ln, hashpos) == '#') { + + while (peek_at(&input, hashpos) == '#') { level++; hashpos++; } container->attributes.header_level = level; - } else if ((matched = scan_open_code_fence(ln, first_nonspace))) { + } else if ((matched = scan_open_code_fence(&input, first_nonspace))) { - container = add_child(container, fenced_code, line_number, - first_nonspace + 1); - container->attributes.fenced_code_data.fence_char = gh_buf_at(ln, - first_nonspace); + container = add_child(container, fenced_code, line_number, first_nonspace + 1); + container->attributes.fenced_code_data.fence_char = peek_at(&input, first_nonspace); container->attributes.fenced_code_data.fence_length = matched; - container->attributes.fenced_code_data.fence_offset = - first_nonspace - offset; + container->attributes.fenced_code_data.fence_offset = first_nonspace - offset; offset = first_nonspace + matched; - } else if ((matched = scan_html_block_tag(ln, first_nonspace))) { + } else if ((matched = scan_html_block_tag(&input, first_nonspace))) { - container = add_child(container, html_block, line_number, - first_nonspace + 1); + container = add_child(container, html_block, line_number, first_nonspace + 1); // note, we don't adjust offset because the tag is part of the text } else if (container->tag == paragraph && - (lev = scan_setext_header_line(ln, first_nonspace)) && + (lev = scan_setext_header_line(&input, first_nonspace)) && // check that there is only one line in the paragraph: gh_buf_strrchr(&container->string_content, '\n', gh_buf_len(&container->string_content) - 2) < 0) { container->tag = setext_header; container->attributes.header_level = lev; - offset = gh_buf_len(ln) - 1; + offset = input.len - 1; } else if (!(container->tag == paragraph && !all_matched) && - (matched = scan_hrule(ln, first_nonspace))) { + (matched = scan_hrule(&input, first_nonspace))) { // it's only now that we know the line is not part of a setext header: container = add_child(container, hrule, line_number, first_nonspace + 1); finalize(container, line_number); container = container->parent; - offset = gh_buf_len(ln) - 1; + offset = input.len - 1; - } else if ((matched = parse_list_marker(ln, first_nonspace, &data))) { + } else if ((matched = parse_list_marker(&input, first_nonspace, &data))) { // compute padding: offset = first_nonspace + matched; i = 0; - while (i <= 5 && gh_buf_at(ln, offset + i) == ' ') { + while (i <= 5 && peek_at(&input, offset + i) == ' ') { i++; } // i = number of spaces after marker, up to 5 - if (i >= 5 || i < 1 || gh_buf_at(ln, offset) == '\n') { + if (i >= 5 || i < 1 || peek_at(&input, offset) == '\n') { data->padding = matched + 1; if (i > 0) { offset += 1; @@ -674,6 +688,7 @@ static void incorporate_line(gh_buf *ln, int line_number, block** curptr) // add the list item container = add_child(container, list_item, line_number, first_nonspace + 1); + /* TODO: static */ container->attributes.list_data = *data; free(data); @@ -691,12 +706,11 @@ static void incorporate_line(gh_buf *ln, int line_number, block** curptr) // appropriate container. first_nonspace = offset; - while (gh_buf_at(ln, first_nonspace) == ' ') { + while (peek_at(&input, first_nonspace) == ' ') first_nonspace++; - } indent = first_nonspace - offset; - blank = gh_buf_at(ln, first_nonspace) == '\n'; + blank = peek_at(&input, first_nonspace) == '\n'; // block quote lines are never blank as they start with > // and we don't count blanks in fenced code for purposes of tight/loose @@ -721,13 +735,12 @@ static void incorporate_line(gh_buf *ln, int line_number, block** curptr) cur->tag == paragraph && gh_buf_len(&cur->string_content) > 0) { - add_line(cur, ln, offset); + add_line(cur, &input, offset); } else { // not a lazy continuation // finalize any blocks that were not matched and set cur to container: while (cur != last_matched_container) { - finalize(cur, line_number); cur = cur->parent; assert(cur != NULL); @@ -735,58 +748,46 @@ static void incorporate_line(gh_buf *ln, int line_number, block** curptr) if (container->tag == indented_code) { - add_line(container, ln, offset); + add_line(container, &input, offset); } else if (container->tag == fenced_code) { matched = (indent <= 3 - && gh_buf_at(ln, first_nonspace) == container->attributes.fenced_code_data.fence_char) - && scan_close_code_fence(ln, first_nonspace, + && peek_at(&input, first_nonspace) == container->attributes.fenced_code_data.fence_char) + && scan_close_code_fence(&input, first_nonspace, container->attributes.fenced_code_data.fence_length); if (matched) { // if closing fence, don't add line to container; instead, close it: finalize(container, line_number); container = container->parent; // back up to parent } else { - add_line(container, ln, offset); + add_line(container, &input, offset); } } else if (container->tag == html_block) { - add_line(container, ln, offset); + add_line(container, &input, offset); } else if (blank) { // ??? do nothing } else if (container->tag == atx_header) { - // chop off trailing ###s...use a scanner? - gh_buf_trim(ln); - int p = gh_buf_len(ln) - 1; - - // if string ends in #s, remove these: - while (gh_buf_at(ln, p) == '#') { - p--; - } - if (gh_buf_at(ln, p) == '\\') { - // the last # was escaped, so we include it. - p++; - } - gh_buf_truncate(ln, p + 1); - add_line(container, ln, first_nonspace); + chop_trailing_hashtags(&input); + add_line(container, &input, first_nonspace); finalize(container, line_number); container = container->parent; } else if (accepts_lines(container->tag)) { - add_line(container, ln, first_nonspace); + add_line(container, &input, first_nonspace); } else if (container->tag != hrule && container->tag != setext_header) { // create paragraph container for line container = add_child(container, paragraph, line_number, first_nonspace + 1); - add_line(container, ln, first_nonspace); + add_line(container, &input, first_nonspace); } else { assert(false); diff --git a/src/buffer.c b/src/buffer.c index cfc6a7e..dc4a405 100644 --- a/src/buffer.c +++ b/src/buffer.c @@ -95,7 +95,7 @@ void gh_buf_clear(gh_buf *buf) int gh_buf_set(gh_buf *buf, const unsigned char *data, int len) { - if (len == 0 || data == NULL) { + if (len <= 0 || data == NULL) { gh_buf_clear(buf); } else { if (data != buf->ptr) { @@ -125,6 +125,9 @@ int gh_buf_putc(gh_buf *buf, int c) int gh_buf_put(gh_buf *buf, const unsigned char *data, int len) { + if (len <= 0) + return 0; + ENSURE_SIZE(buf, buf->size + len + 1); memmove(buf->ptr + buf->size, data, len); buf->size += len; @@ -272,15 +275,28 @@ void gh_buf_truncate(gh_buf *buf, int len) } } +void gh_buf_drop(gh_buf *buf, int n) +{ + if (n > 0) { + buf->size = buf->size - n; + if (buf->size) + memmove(buf->ptr, buf->ptr + n, buf->size); + + buf->ptr[buf->size] = '\0'; + } +} + void gh_buf_trim(gh_buf *buf) { - /* TODO: leading whitespace? */ - /* + int i = 0; + + if (!buf->size) + return; + while (i < buf->size && isspace(buf->ptr[i])) i++; - gh_buf_truncate(buf, i); - */ + gh_buf_drop(buf, i); /* rtrim */ while (buf->size > 0) { diff --git a/src/buffer.h b/src/buffer.h index 422ef02..0d5143e 100644 --- a/src/buffer.h +++ b/src/buffer.h @@ -105,8 +105,8 @@ extern void gh_buf_clear(gh_buf *buf); int gh_buf_strchr(const gh_buf *buf, int c, int pos); int gh_buf_strrchr(const gh_buf *buf, int c, int pos); +void gh_buf_drop(gh_buf *buf, int n); void gh_buf_truncate(gh_buf *buf, int len); -void gh_buf_ltruncate(gh_buf *buf, int len); void gh_buf_trim(gh_buf *buf); #endif diff --git a/src/html/houdini_href_e.c b/src/html/houdini_href_e.c index 59fe850..b2a7d79 100644 --- a/src/html/houdini_href_e.c +++ b/src/html/houdini_href_e.c @@ -62,16 +62,8 @@ houdini_escape_href(gh_buf *ob, const uint8_t *src, size_t size) while (i < size && HREF_SAFE[src[i]] != 0) i++; - if (likely(i > org)) { - if (unlikely(org == 0)) { - if (i >= size) - return 0; - - gh_buf_grow(ob, HOUDINI_ESCAPED_SIZE(size)); - } - + if (likely(i > org)) gh_buf_put(ob, src + org, i - org); - } /* escaping */ if (i >= size) diff --git a/src/html/houdini_html_e.c b/src/html/houdini_html_e.c index 316c5ce..5cdd3dd 100644 --- a/src/html/houdini_html_e.c +++ b/src/html/houdini_html_e.c @@ -54,16 +54,8 @@ houdini_escape_html0(gh_buf *ob, const uint8_t *src, size_t size, int secure) while (i < size && (esc = HTML_ESCAPE_TABLE[src[i]]) == 0) i++; - if (i > org) { - if (unlikely(org == 0)) { - if (i >= size) - return 0; - - gh_buf_grow(ob, HOUDINI_ESCAPED_SIZE(size)); - } - + if (i > org) gh_buf_put(ob, src + org, i - org); - } /* escaping */ if (unlikely(i >= size)) diff --git a/src/html/html.c b/src/html/html.c index 2f160ca..27ffe58 100644 --- a/src/html/html.c +++ b/src/html/html.c @@ -68,7 +68,7 @@ void blocks_to_html(gh_buf *html, block *b, bool tight) cr(html); gh_buf_puts(html, "
  • "); blocks_to_html(html, b->children, tight); - gh_buf_trim(html); + gh_buf_trim(html); /* TODO: rtrim */ gh_buf_puts(html, "
    • "); cr(html); break; @@ -106,7 +106,7 @@ void blocks_to_html(gh_buf *html, block *b, bool tight) cr(html); gh_buf_puts(html, "
    ");
 				escape_html(html, b->string_content.ptr, b->string_content.size);
-				gh_buf_puts(html, "
    "); + gh_buf_puts(html, ""); cr(html); break; diff --git a/src/inlines.c b/src/inlines.c index 7b48ad9..ef27a24 100644 --- a/src/inlines.c +++ b/src/inlines.c @@ -9,10 +9,10 @@ #include "scanners.h" typedef struct Subject { - const gh_buf *buffer; - int pos; - reference** reference_map; - int label_nestlevel; + chunk input; + int pos; + int label_nestlevel; + reference** reference_map; } subject; reference* lookup_reference(reference** refmap, chunk *label); @@ -27,12 +27,16 @@ inline static void chunk_trim(chunk *c); inline static chunk chunk_literal(const char *data); inline static chunk chunk_buf_detach(gh_buf *buf); -inline static chunk chunk_buf(const gh_buf *buf, int pos, int len); +inline static chunk chunk_dup(const chunk *ch, int pos, int len); static inl *parse_chunk_inlines(chunk *chunk, reference** refmap); static inl *parse_inlines_while(subject* subj, int (*f)(subject*)); static int parse_inline(subject* subj, inl ** last); +static void subject_from_chunk(subject *e, chunk *chunk, reference** refmap); +static void subject_from_buf(subject *e, gh_buf *buffer, reference** refmap); +static int subject_find_special_char(subject *subj); + extern void free_reference(reference *ref) { free(ref->label); free(ref->url); @@ -101,10 +105,12 @@ extern reference* make_reference(chunk *label, chunk *url, chunk *title) extern void add_reference(reference** refmap, reference* ref) { reference * t = NULL; - HASH_FIND(hh, *refmap, (char*)ref->label, (unsigned)strlen(ref->label), t); + const char *label = (const char *)ref->label; + + HASH_FIND(hh, *refmap, label, strlen(label), t); if (t == NULL) { - HASH_ADD_KEYPTR(hh, *refmap, (char*)ref->label, (unsigned)strlen(ref->label), ref); + HASH_ADD_KEYPTR(hh, *refmap, label, strlen(label), ref); } else { free_reference(ref); // we free this now since it won't be in the refmap } @@ -210,87 +216,49 @@ inline static inl* append_inlines(inl* a, inl* b) return a; } -// Make a 'subject' from an input string. -static void init_subject(subject *e, gh_buf *buffer, int input_pos, reference** refmap) +static void subject_from_buf(subject *e, gh_buf *buffer, reference** refmap) { - e->buffer = buffer; - e->pos = input_pos; + e->input.data = buffer->ptr; + e->input.len = buffer->size; + e->input.alloc = 0; + e->pos = 0; e->label_nestlevel = 0; e->reference_map = refmap; -} - -inline static int isbacktick(int c) -{ - return (c == '`'); -} - -inline static void chunk_free(chunk *c) -{ - if (c->alloc) - free((char *)c->data); - - c->data = NULL; - c->alloc = 0; - c->len = 0; -} - -inline static void chunk_trim(chunk *c) -{ - while (c->len && isspace(c->data[0])) { - c->data++; - c->len--; - } - - while (c->len > 0) { - if (!isspace(c->data[c->len - 1])) - break; - c->len--; - } + chunk_rtrim(&e->input); } -inline static unsigned char *chunk_to_cstr(chunk *c) +static void subject_from_chunk(subject *e, chunk *chunk, reference** refmap) { - unsigned char *str; - - str = malloc(c->len + 1); - memcpy(str, c->data, c->len); - str[c->len] = 0; + e->input.data = chunk->data; + e->input.len = chunk->len; + e->input.alloc = 0; + e->pos = 0; + e->label_nestlevel = 0; + e->reference_map = refmap; - return str; + chunk_rtrim(&e->input); } -inline static chunk chunk_literal(const char *data) +inline static int isbacktick(int c) { - chunk c = {data, data ? strlen(data) : 0, 0}; - return c; + return (c == '`'); } -inline static chunk chunk_buf(const gh_buf *buf, int pos, int len) +static inline unsigned char peek_char(subject *subj) { - chunk c = {buf->ptr + pos, len, 0}; - return c; + return (subj->pos < subj->input.len) ? subj->input.data[subj->pos] : 0; } -inline static chunk chunk_buf_detach(gh_buf *buf) +static inline unsigned char peek_at(subject *subj, int pos) { - chunk c; - - c.len = buf->size; - c.data = gh_buf_detach(buf); - c.alloc = 1; - - return c; + return subj->input.data[pos]; } -// Return the next character in the subject, without advancing. -// Return 0 if at the end of the subject. -#define peek_char(subj) gh_buf_at((subj)->buffer, (subj)->pos) - // Return true if there are more characters in the subject. inline static int is_eof(subject* subj) { - return (subj->pos >= gh_buf_len(subj->buffer)); + return (subj->pos >= subj->input.len); } // Advance the subject. Doesn't check for eof. @@ -308,7 +276,7 @@ inline static chunk take_while(subject* subj, int (*f)(int)) len++; } - return chunk_buf(subj->buffer, startpos, len); + return chunk_dup(&subj->input, startpos, len); } // Try to process a backtick code span that began with a @@ -388,7 +356,7 @@ static inl* handle_backticks(subject *subj) } else { gh_buf buf = GH_BUF_INIT; - gh_buf_set(&buf, subj->buffer->ptr + startpos, endpos - startpos - openticks.len); + gh_buf_set(&buf, subj->input.data + startpos, endpos - startpos - openticks.len); gh_buf_trim(&buf); normalize_whitespace(&buf); @@ -404,7 +372,7 @@ static int scan_delims(subject* subj, char c, bool * can_open, bool * can_close) char char_before, char_after; int startpos = subj->pos; - char_before = subj->pos == 0 ? '\n' : gh_buf_at(subj->buffer, subj->pos - 1); + char_before = subj->pos == 0 ? '\n' : peek_at(subj, subj->pos - 1); while (peek_char(subj) == c) { numdelims++; advance(subj); @@ -439,7 +407,7 @@ static inl* handle_strong_emph(subject* subj, char c) numdelims = scan_delims(subj, c, &can_open, &can_close); subj->pos += numdelims; - new = make_str(chunk_buf(subj->buffer, subj->pos - numdelims, numdelims)); + new = make_str(chunk_dup(&subj->input, subj->pos - numdelims, numdelims)); *last = new; first_head = new; result = new; @@ -488,7 +456,7 @@ static inl* handle_strong_emph(subject* subj, char c) numdelims = scan_delims(subj, c, &can_open, &can_close); if (can_close && numdelims >= 1 && numdelims <= 3 && numdelims != first_close_delims) { - new = make_str(chunk_buf(subj->buffer, subj->pos, numdelims)); + new = make_str(chunk_dup(&subj->input, subj->pos, numdelims)); append_inlines(*last, new); *last = new; if (first_close_delims == 1 && numdelims > 2) { @@ -554,7 +522,7 @@ static inl* handle_backslash(subject *subj) unsigned char nextchar = peek_char(subj); if (ispunct(nextchar)) { // only ascii symbols and newline can be escaped advance(subj); - return make_str(chunk_buf(subj->buffer, subj->pos - 1, 1)); + return make_str(chunk_dup(&subj->input, subj->pos - 1, 1)); } else if (nextchar == '\n') { advance(subj); return make_linebreak(); @@ -569,9 +537,9 @@ static inl* handle_entity(subject* subj) { int match; inl *result; - match = scan_entity(subj->buffer, subj->pos); + match = scan_entity(&subj->input, subj->pos); if (match) { - result = make_entity(chunk_buf(subj->buffer, subj->pos, match)); + result = make_entity(chunk_dup(&subj->input, subj->pos, match)); subj->pos += match; } else { advance(subj); @@ -584,15 +552,13 @@ static inl* handle_entity(subject* subj) // Returns an inline sequence consisting of str and entity elements. static inl *make_str_with_entities(chunk *content) { - inl * result = NULL; - inl * new; + inl *result = NULL; + inl *new; int searchpos; char c; subject subj; - gh_buf content_buf = GH_BUF_INIT; - gh_buf_set(&content_buf, content->data, content->len); - init_subject(&subj, &content_buf, 0, NULL); + subject_from_chunk(&subj, content, NULL); while ((c = peek_char(&subj))) { switch (c) { @@ -600,18 +566,13 @@ static inl *make_str_with_entities(chunk *content) new = handle_entity(&subj); break; default: - searchpos = gh_buf_strchr(subj.buffer, '&', subj.pos); - if (searchpos < 0) { - searchpos = gh_buf_len(subj.buffer); - } - - new = make_str(chunk_buf(subj.buffer, subj.pos, searchpos - subj.pos)); + searchpos = chunk_strchr(&subj.input, '&', subj.pos); + new = make_str(chunk_dup(&subj.input, subj.pos, searchpos - subj.pos)); subj.pos = searchpos; } result = append_inlines(result, new); } - gh_buf_free(&content_buf); return result; } @@ -678,9 +639,9 @@ static inl* handle_pointy_brace(subject* subj) advance(subj); // advance past first < // first try to match a URL autolink - matchlen = scan_autolink_uri(subj->buffer, subj->pos); + matchlen = scan_autolink_uri(&subj->input, subj->pos); if (matchlen > 0) { - contents = chunk_buf(subj->buffer, subj->pos, matchlen - 1); + contents = chunk_dup(&subj->input, subj->pos, matchlen - 1); subj->pos += matchlen; return make_link( @@ -691,11 +652,11 @@ static inl* handle_pointy_brace(subject* subj) } // next try to match an email autolink - matchlen = scan_autolink_email(subj->buffer, subj->pos); + matchlen = scan_autolink_email(&subj->input, subj->pos); if (matchlen > 0) { gh_buf mail_url = GH_BUF_INIT; - contents = chunk_buf(subj->buffer, subj->pos, matchlen - 1); + contents = chunk_dup(&subj->input, subj->pos, matchlen - 1); subj->pos += matchlen; gh_buf_puts(&mail_url, "mailto:"); @@ -709,9 +670,9 @@ static inl* handle_pointy_brace(subject* subj) } // finally, try to match an html tag - matchlen = scan_html_tag(subj->buffer, subj->pos); + matchlen = scan_html_tag(&subj->input, subj->pos); if (matchlen > 0) { - contents = chunk_buf(subj->buffer, subj->pos - 1, matchlen + 1); + contents = chunk_dup(&subj->input, subj->pos - 1, matchlen + 1); subj->pos += matchlen; return make_raw_html(contents); } @@ -776,12 +737,7 @@ static int link_label(subject* subj, chunk *raw_label) } } if (c == ']') { - *raw_label = chunk_buf( - subj->buffer, - startpos + 1, - subj->pos - (startpos + 1) - ); - + *raw_label = chunk_dup(&subj->input, startpos + 1, subj->pos - (startpos + 1)); subj->label_nestlevel = 0; advance(subj); // advance past ] return 1; @@ -813,25 +769,25 @@ static inl* handle_left_bracket(subject* subj) if (found_label) { if (peek_char(subj) == '(' && - ((sps = scan_spacechars(subj->buffer, subj->pos + 1)) > -1) && - ((n = scan_link_url(subj->buffer, subj->pos + 1 + sps)) > -1)) { + ((sps = scan_spacechars(&subj->input, subj->pos + 1)) > -1) && + ((n = scan_link_url(&subj->input, subj->pos + 1 + sps)) > -1)) { // try to parse an explicit link: starturl = subj->pos + 1 + sps; // after ( endurl = starturl + n; - starttitle = endurl + scan_spacechars(subj->buffer, endurl); + starttitle = endurl + scan_spacechars(&subj->input, endurl); // ensure there are spaces btw url and title endtitle = (starttitle == endurl) ? starttitle : - starttitle + scan_link_title(subj->buffer, starttitle); + starttitle + scan_link_title(&subj->input, starttitle); - endall = endtitle + scan_spacechars(subj->buffer, endtitle); + endall = endtitle + scan_spacechars(&subj->input, endtitle); - if (gh_buf_at(subj->buffer, endall) == ')') { + if (peek_at(subj, endall) == ')') { subj->pos = endall + 1; - url = chunk_buf(subj->buffer, starturl, endurl - starturl); - title = chunk_buf(subj->buffer, starttitle, endtitle - starttitle); + url = chunk_dup(&subj->input, starturl, endurl - starturl); + title = chunk_dup(&subj->input, starttitle, endtitle - starttitle); lab = parse_chunk_inlines(&rawlabel, NULL); return make_link(lab, url, title); @@ -850,7 +806,7 @@ static inl* handle_left_bracket(subject* subj) // Check for reference link. // First, see if there's another label: - subj->pos = subj->pos + scan_spacechars(subj->buffer, endlabel); + subj->pos = subj->pos + scan_spacechars(&subj->input, endlabel); reflabel = rawlabel; // if followed by a nonempty link label, we change reflabel to it: @@ -892,8 +848,8 @@ static inl* handle_newline(subject *subj) advance(subj); } if (nlpos > 1 && - gh_buf_at(subj->buffer, nlpos - 1) == ' ' && - gh_buf_at(subj->buffer, nlpos - 2) == ' ') { + peek_at(subj, nlpos - 1) == ' ' && + peek_at(subj, nlpos - 2) == ' ') { return make_linebreak(); } else { return make_softbreak(); @@ -917,30 +873,22 @@ extern inl* parse_inlines_while(subject* subj, int (*f)(subject*)) inl *parse_chunk_inlines(chunk *chunk, reference** refmap) { - inl *result; subject subj; - gh_buf full_chunk = GH_BUF_INIT; - - gh_buf_set(&full_chunk, chunk->data, chunk->len); - init_subject(&subj, &full_chunk, 0, refmap); - result = parse_inlines_while(&subj, not_eof); - - gh_buf_free(&full_chunk); - return result; + subject_from_chunk(&subj, chunk, refmap); + return parse_inlines_while(&subj, not_eof); } -static int find_special_char(subject *subj) +static int subject_find_special_char(subject *subj) { int n = subj->pos + 1; - int size = (int)gh_buf_len(subj->buffer); - while (n < size) { - if (strchr("\n\\`&_*[]buffer, n))) + while (n < subj->input.len) { + if (strchr("\n\\`&_*[]input.data[n])) return n; n++; } - return -1; + return subj->input.len; } // Parse an inline, advancing subject, and add it to last element. @@ -973,11 +921,13 @@ static int parse_inline(subject* subj, inl ** last) new = handle_pointy_brace(subj); break; case '_': - if (subj->pos > 0 && (isalnum(gh_buf_at(subj->buffer, subj->pos - 1)) || - gh_buf_at(subj->buffer, subj->pos - 1) == '_')) { - new = make_str(chunk_literal("_")); - advance(subj); - break; + if (subj->pos > 0) { + unsigned char prev = peek_at(subj, subj->pos - 1); + if (isalnum(prev) || prev == '_') { + new = make_str(chunk_literal("_")); + advance(subj); + break; + } } new = handle_strong_emph(subj, '_'); @@ -1002,18 +952,13 @@ static int parse_inline(subject* subj, inl ** last) } break; default: - text_literal: - endpos = find_special_char(subj); - if (endpos < 0) { - endpos = gh_buf_len(subj->buffer); - } - - contents = chunk_buf(subj->buffer, subj->pos, endpos - subj->pos); + endpos = subject_find_special_char(subj); + contents = chunk_dup(&subj->input, subj->pos, endpos - subj->pos); subj->pos = endpos; // if we're at a newline, strip trailing spaces. if (peek_char(subj) == '\n') { - chunk_trim(&contents); + chunk_rtrim(&contents); } new = make_str(contents); @@ -1026,10 +971,10 @@ static int parse_inline(subject* subj, inl ** last) return 1; } -extern inl* parse_inlines(gh_buf *input, int input_pos, reference** refmap) +extern inl* parse_inlines(gh_buf *input, reference** refmap) { subject subj; - init_subject(&subj, input, input_pos, refmap); + subject_from_buf(&subj, input, refmap); return parse_inlines_while(&subj, not_eof); } @@ -1048,7 +993,7 @@ void spnl(subject* subj) // Modify refmap if a reference is encountered. // Return 0 if no reference found, otherwise position of subject // after reference is parsed. -extern int parse_reference(gh_buf *input, int input_pos, reference** refmap) +extern int parse_reference(gh_buf *input, reference** refmap) { subject subj; @@ -1058,9 +1003,9 @@ extern int parse_reference(gh_buf *input, int input_pos, reference** refmap) int matchlen = 0; int beforetitle; - reference * new = NULL; + reference *new = NULL; - init_subject(&subj, input, input_pos, NULL); + subject_from_buf(&subj, input, NULL); // parse label: if (!link_label(&subj, &lab)) @@ -1075,9 +1020,9 @@ extern int parse_reference(gh_buf *input, int input_pos, reference** refmap) // parse link url: spnl(&subj); - matchlen = scan_link_url(subj.buffer, subj.pos); + matchlen = scan_link_url(&subj.input, subj.pos); if (matchlen) { - url = chunk_buf(subj.buffer, subj.pos, matchlen); + url = chunk_dup(&subj.input, subj.pos, matchlen); subj.pos += matchlen; } else { return 0; @@ -1086,9 +1031,9 @@ extern int parse_reference(gh_buf *input, int input_pos, reference** refmap) // parse optional link_title beforetitle = subj.pos; spnl(&subj); - matchlen = scan_link_title(subj.buffer, subj.pos); + matchlen = scan_link_title(&subj.input, subj.pos); if (matchlen) { - title = chunk_buf(subj.buffer, subj.pos, matchlen); + title = chunk_dup(&subj.input, subj.pos, matchlen); subj.pos += matchlen; } else { subj.pos = beforetitle; diff --git a/src/print.c b/src/print.c index 0a87925..c262995 100644 --- a/src/print.c +++ b/src/print.c @@ -9,7 +9,7 @@ static void print_str(const unsigned char *s, int len) int i; if (len < 0) - len = strlen(s); + len = strlen((char *)s); putchar('"'); for (i = 0; i < len; ++i) { diff --git a/src/scanners.h b/src/scanners.h index b6e586b..f96c42d 100644 --- a/src/scanners.h +++ b/src/scanners.h @@ -1,15 +1,15 @@ -#include "buffer.h" +#include "stmd.h" -int scan_autolink_uri(const gh_buf *s, int pos); -int scan_autolink_email(const gh_buf *s, int pos); -int scan_html_tag(const gh_buf *s, int pos); -int scan_html_block_tag(const gh_buf *s, int pos); -int scan_link_url(const gh_buf *s, int pos); -int scan_link_title(const gh_buf *s, int pos); -int scan_spacechars(const gh_buf *s, int pos); -int scan_atx_header_start(const gh_buf *s, int pos); -int scan_setext_header_line(const gh_buf *s, int pos); -int scan_hrule(const gh_buf *s, int pos); -int scan_open_code_fence(const gh_buf *s, int pos); -int scan_close_code_fence(const gh_buf *s, int pos, int len); -int scan_entity(const gh_buf *s, int pos); +int scan_autolink_uri(chunk *c, int offset); +int scan_autolink_email(chunk *c, int offset); +int scan_html_tag(chunk *c, int offset); +int scan_html_block_tag(chunk *c, int offset); +int scan_link_url(chunk *c, int offset); +int scan_link_title(chunk *c, int offset); +int scan_spacechars(chunk *c, int offset); +int scan_atx_header_start(chunk *c, int offset); +int scan_setext_header_line(chunk *c, int offset); +int scan_hrule(chunk *c, int offset); +int scan_open_code_fence(chunk *c, int offset); +int scan_close_code_fence(chunk *c, int offset, int len); +int scan_entity(chunk *c, int offset); diff --git a/src/scanners.re b/src/scanners.re index 7323ef9..5ac7c15 100644 --- a/src/scanners.re +++ b/src/scanners.re @@ -1,8 +1,15 @@ -#include "buffer.h" +#include "scanners.h" + +#define SCAN_DATA \ + const unsigned char *marker = NULL; \ + const unsigned char *p = c->data + offset; \ + const unsigned char *start = p; \ + const unsigned char *end = c->data + c->len /*!re2c re2c:define:YYCTYPE = "unsigned char"; re2c:define:YYCURSOR = p; + re2c:define:YYLIMIT = end; re2c:define:YYMARKER = marker; re2c:define:YYCTXMARKER = marker; re2c:yyfill:enable = 0; @@ -55,11 +62,9 @@ */ // Try to match URI autolink after first <, returning number of chars matched. -extern int scan_autolink_uri(const gh_buf *s, int pos) +extern int scan_autolink_uri(chunk *c, int offset) { - unsigned char * marker = NULL; - unsigned char * p = &(s->ptr[pos]); - unsigned char * start = p; + SCAN_DATA; /*!re2c scheme [:]([^\x00-\x20<>\\]|escaped_char)*[>] { return (p - start); } .? { return 0; } @@ -67,11 +72,9 @@ extern int scan_autolink_uri(const gh_buf *s, int pos) } // Try to match email autolink after first <, returning num of chars matched. -extern int scan_autolink_email(const gh_buf *s, int pos) +extern int scan_autolink_email(chunk *c, int offset) { - unsigned char * marker = NULL; - unsigned char * p = &(s->ptr[pos]); - unsigned char * start = p; + SCAN_DATA; /*!re2c [a-zA-Z0-9.!#$%&'*+/=?^_`{|}~-]+ [@] @@ -83,11 +86,9 @@ extern int scan_autolink_email(const gh_buf *s, int pos) } // Try to match an HTML tag after first <, returning num of chars matched. -extern int scan_html_tag(const gh_buf *s, int pos) +extern int scan_html_tag(chunk *c, int offset) { - unsigned char * marker = NULL; - unsigned char * p = &(s->ptr[pos]); - unsigned char * start = p; + SCAN_DATA; /*!re2c htmltag { return (p - start); } .? { return 0; } @@ -96,11 +97,9 @@ extern int scan_html_tag(const gh_buf *s, int pos) // Try to match an HTML block tag including first <, // returning num of chars matched. -extern int scan_html_block_tag(const gh_buf *s, int pos) +extern int scan_html_block_tag(chunk *c, int offset) { - unsigned char * marker = NULL; - unsigned char * p = &(s->ptr[pos]); - unsigned char * start = p; + SCAN_DATA; /*!re2c [<] [/] blocktagname (spacechar | [>]) { return (p - start); } [<] blocktagname (spacechar | [/>]) { return (p - start); } @@ -113,11 +112,9 @@ extern int scan_html_block_tag(const gh_buf *s, int pos) // This may optionally be contained in <..>; otherwise // whitespace and unbalanced right parentheses aren't allowed. // Newlines aren't ever allowed. -extern int scan_link_url(const gh_buf *s, int pos) +extern int scan_link_url(chunk *c, int offset) { - unsigned char * marker = NULL; - unsigned char * p = &(s->ptr[pos]); - unsigned char * start = p; + SCAN_DATA; /*!re2c [ \n]* [<] ([^<>\n\\\x00] | escaped_char | [\\])* [>] { return (p - start); } [ \n]* (reg_char+ | escaped_char | in_parens_nosp)* { return (p - start); } @@ -128,11 +125,9 @@ extern int scan_link_url(const gh_buf *s, int pos) // Try to match a link title (in single quotes, in double quotes, or // in parentheses), returning number of chars matched. Allow one // level of internal nesting (quotes within quotes). -extern int scan_link_title(const gh_buf *s, int pos) +extern int scan_link_title(chunk *c, int offset) { - unsigned char * marker = NULL; - unsigned char * p = &(s->ptr[pos]); - unsigned char * start = p; + SCAN_DATA; /*!re2c ["] (escaped_char|[^"\x00])* ["] { return (p - start); } ['] (escaped_char|[^'\x00])* ['] { return (p - start); } @@ -142,10 +137,9 @@ extern int scan_link_title(const gh_buf *s, int pos) } // Match space characters, including newlines. -extern int scan_spacechars(const gh_buf *s, int pos) +extern int scan_spacechars(chunk *c, int offset) { - unsigned char * p = &(s->ptr[pos]); - unsigned char * start = p; + SCAN_DATA; /*!re2c [ \t\n]* { return (p - start); } . { return 0; } @@ -153,11 +147,9 @@ extern int scan_spacechars(const gh_buf *s, int pos) } // Match ATX header start. -extern int scan_atx_header_start(const gh_buf *s, int pos) +extern int scan_atx_header_start(chunk *c, int offset) { - unsigned char * marker = NULL; - unsigned char * p = &(s->ptr[pos]); - unsigned char * start = p; + SCAN_DATA; /*!re2c [#]{1,6} ([ ]+|[\n]) { return (p - start); } .? { return 0; } @@ -166,10 +158,9 @@ extern int scan_atx_header_start(const gh_buf *s, int pos) // Match sexext header line. Return 1 for level-1 header, // 2 for level-2, 0 for no match. -extern int scan_setext_header_line(const gh_buf *s, int pos) +extern int scan_setext_header_line(chunk *c, int offset) { - unsigned char * marker = NULL; - unsigned char * p = &(s->ptr[pos]); + SCAN_DATA; /*!re2c [=]+ [ ]* [\n] { return 1; } [-]+ [ ]* [\n] { return 2; } @@ -180,11 +171,9 @@ extern int scan_setext_header_line(const gh_buf *s, int pos) // Scan a horizontal rule line: "...three or more hyphens, asterisks, // or underscores on a line by themselves. If you wish, you may use // spaces between the hyphens or asterisks." -extern int scan_hrule(const gh_buf *s, int pos) +extern int scan_hrule(chunk *c, int offset) { - unsigned char * marker = NULL; - unsigned char * p = &(s->ptr[pos]); - unsigned char * start = p; + SCAN_DATA; /*!re2c ([*][ ]*){3,} [ \t]* [\n] { return (p - start); } ([_][ ]*){3,} [ \t]* [\n] { return (p - start); } @@ -194,11 +183,9 @@ extern int scan_hrule(const gh_buf *s, int pos) } // Scan an opening code fence. -extern int scan_open_code_fence(const gh_buf *s, int pos) +extern int scan_open_code_fence(chunk *c, int offset) { - unsigned char * marker = NULL; - unsigned char * p = &(s->ptr[pos]); - unsigned char * start = p; + SCAN_DATA; /*!re2c [`]{3,} / [^`\n\x00]*[\n] { return (p - start); } [~]{3,} / [^~\n\x00]*[\n] { return (p - start); } @@ -207,11 +194,9 @@ extern int scan_open_code_fence(const gh_buf *s, int pos) } // Scan a closing code fence with length at least len. -extern int scan_close_code_fence(const gh_buf *s, int pos, int len) +extern int scan_close_code_fence(chunk *c, int offset, int len) { - unsigned char * marker = NULL; - unsigned char * p = &(s->ptr[pos]); - unsigned char * start = p; + SCAN_DATA; /*!re2c ([`]{3,} | [~]{3,}) / spacechar* [\n] { if (p - start > len) { @@ -225,11 +210,9 @@ extern int scan_close_code_fence(const gh_buf *s, int pos, int len) // Scans an entity. // Returns number of chars matched. -extern int scan_entity(const gh_buf *s, int pos) +extern int scan_entity(chunk *c, int offset) { - unsigned char * marker = NULL; - unsigned char * p = &(s->ptr[pos]); - unsigned char * start = p; + SCAN_DATA; /*!re2c [&] ([#] ([Xx][A-Fa-f0-9]{1,8}|[0-9]{1,8}) |[A-Za-z][A-Za-z0-9]{1,31} ) [;] { return (p - start); } diff --git a/src/stmd.h b/src/stmd.h index 3e284bd..4a3c399 100644 --- a/src/stmd.h +++ b/src/stmd.h @@ -1,17 +1,15 @@ +#ifndef _STDMD_H_ +#define _STDMD_H_ + #include #include #include "buffer.h" +#include "chunk.h" #include "uthash.h" #define VERSION "0.1" #define CODE_INDENT 4 -typedef struct { - const unsigned char *data; - int len; - int alloc; -} chunk; - typedef struct Inline { enum { INL_STRING, INL_SOFTBREAK, INL_LINEBREAK, INL_CODE, INL_RAW_HTML, INL_ENTITY, INL_EMPH, INL_STRONG, INL_LINK, INL_IMAGE } tag; @@ -79,7 +77,6 @@ typedef struct Block { struct Block* parent; struct Block* top; gh_buf string_content; - int string_pos; inl* inline_content; union { struct ListData list_data; @@ -91,10 +88,10 @@ typedef struct Block { struct Block * prev; } block; -inl* parse_inlines(gh_buf *input, int input_pos, reference** refmap); +inl* parse_inlines(gh_buf *input, reference** refmap); void free_inlines(inl* e); -int parse_reference(gh_buf *input, int input_pos, reference** refmap); +int parse_reference(gh_buf *input, reference** refmap); void free_reference(reference *ref); void free_reference_map(reference **refmap); @@ -117,3 +114,4 @@ void inlines_to_html(gh_buf *html, inl *b); void utf8proc_case_fold(gh_buf *dest, const unsigned char *str, int len); +#endif -- cgit v1.2.3 From 543c2c94d71adee42c7bd2f8027d75c87ed8120d Mon Sep 17 00:00:00 2001 From: Vicent Marti Date: Thu, 4 Sep 2014 18:38:14 +0200 Subject: Rename to strbuf --- src/blocks.c | 64 +++++++++++++++---------------- src/buffer.c | 86 ++++++++++++++++++++--------------------- src/buffer.h | 80 +++++++++++++++++++------------------- src/chunk.h | 4 +- src/html/houdini.h | 22 +++++------ src/html/houdini_href_e.c | 12 +++--- src/html/houdini_html_e.c | 10 ++--- src/html/html.c | 98 +++++++++++++++++++++++------------------------ src/inlines.c | 50 ++++++++++++------------ src/main.c | 4 +- src/stmd.h | 16 ++++---- src/utf8.c | 6 +-- 12 files changed, 226 insertions(+), 226 deletions(-) (limited to 'src/buffer.h') diff --git a/src/blocks.c b/src/blocks.c index cf0e9e4..9faccd9 100644 --- a/src/blocks.c +++ b/src/blocks.c @@ -10,7 +10,7 @@ #define peek_at(i, n) (i)->data[n] -static void incorporate_line(gh_buf *ln, int line_number, block** curptr); +static void incorporate_line(strbuf *ln, int line_number, block** curptr); static void finalize(block* b, int line_number); static block* make_block(int tag, int start_line, int start_column) @@ -28,7 +28,7 @@ static block* make_block(int tag, int start_line, int start_column) e->parent = NULL; e->top = NULL; e->attributes.refmap = NULL; - gh_buf_init(&e->string_content, 32); + strbuf_init(&e->string_content, 32); e->inline_content = NULL; e->next = NULL; e->prev = NULL; @@ -49,7 +49,7 @@ extern block* make_document() } // Returns true if line has only space characters, else false. -bool is_blank(gh_buf *s, int offset) +bool is_blank(strbuf *s, int offset) { while (offset < s->size) { switch (s->ptr[offset]) { @@ -85,10 +85,10 @@ static inline bool accepts_lines(int block_type) static void add_line(block* block, chunk *ch, int offset) { assert(block->open); - gh_buf_put(&block->string_content, ch->data + offset, ch->len - offset); + strbuf_put(&block->string_content, ch->data + offset, ch->len - offset); } -static void remove_trailing_blank_lines(gh_buf *ln) +static void remove_trailing_blank_lines(strbuf *ln) { int i; @@ -100,13 +100,13 @@ static void remove_trailing_blank_lines(gh_buf *ln) } if (i < 0) { - gh_buf_clear(ln); + strbuf_clear(ln); return; } - i = gh_buf_strchr(ln, '\n', i); + i = strbuf_strchr(ln, '\n', i); if (i >= 0) - gh_buf_truncate(ln, i); + strbuf_truncate(ln, i); } // Check to see if a block ends with a blank line, descending @@ -164,10 +164,10 @@ static void finalize(block* b, int line_number) switch (b->tag) { case paragraph: pos = 0; - while (gh_buf_at(&b->string_content, 0) == '[' && + while (strbuf_at(&b->string_content, 0) == '[' && (pos = parse_reference(&b->string_content, b->top->attributes.refmap))) { - gh_buf_drop(&b->string_content, pos); + strbuf_drop(&b->string_content, pos); } if (is_blank(&b->string_content, 0)) { b->tag = reference_def; @@ -176,23 +176,23 @@ static void finalize(block* b, int line_number) case indented_code: remove_trailing_blank_lines(&b->string_content); - gh_buf_putc(&b->string_content, '\n'); + strbuf_putc(&b->string_content, '\n'); break; case fenced_code: // first line of contents becomes info - firstlinelen = gh_buf_strchr(&b->string_content, '\n', 0); + firstlinelen = strbuf_strchr(&b->string_content, '\n', 0); - gh_buf_init(&b->attributes.fenced_code_data.info, 0); - gh_buf_set( + strbuf_init(&b->attributes.fenced_code_data.info, 0); + strbuf_set( &b->attributes.fenced_code_data.info, b->string_content.ptr, firstlinelen ); - gh_buf_drop(&b->string_content, firstlinelen + 1); + strbuf_drop(&b->string_content, firstlinelen + 1); - gh_buf_trim(&b->attributes.fenced_code_data.info); + strbuf_trim(&b->attributes.fenced_code_data.info); unescape_buffer(&b->attributes.fenced_code_data.info); break; @@ -265,9 +265,9 @@ extern void free_blocks(block* e) while (e != NULL) { next = e->next; free_inlines(e->inline_content); - gh_buf_free(&e->string_content); + strbuf_free(&e->string_content); if (e->tag == fenced_code) { - gh_buf_free(&e->attributes.fenced_code_data.info); + strbuf_free(&e->attributes.fenced_code_data.info); } else if (e->tag == document) { free_reference_map(e->attributes.refmap); } @@ -287,7 +287,7 @@ void process_inlines(block* cur, reference** refmap) case setext_header: cur->inline_content = parse_inlines(&cur->string_content, refmap); // MEM - // gh_buf_free(&cur->string_content); + // strbuf_free(&cur->string_content); break; default: @@ -369,7 +369,7 @@ static int lists_match(struct ListData list_data, list_data.bullet_char == item_data.bullet_char); } -static void expand_tabs(gh_buf *ob, const unsigned char *line, size_t size) +static void expand_tabs(strbuf *ob, const unsigned char *line, size_t size) { size_t i = 0, tab = 0; @@ -381,13 +381,13 @@ static void expand_tabs(gh_buf *ob, const unsigned char *line, size_t size) } if (i > org) - gh_buf_put(ob, line + org, i - org); + strbuf_put(ob, line + org, i - org); if (i >= size) break; do { - gh_buf_putc(ob, ' '); tab++; + strbuf_putc(ob, ' '); tab++; } while (tab % 4); i++; @@ -409,7 +409,7 @@ static block *finalize_document(block *document, int linenum) extern block *stmd_parse_file(FILE *f) { - gh_buf line = GH_BUF_INIT; + strbuf line = GH_BUF_INIT; unsigned char buffer[4096]; int linenum = 1; block *document = make_document(); @@ -417,17 +417,17 @@ extern block *stmd_parse_file(FILE *f) while (fgets((char *)buffer, sizeof(buffer), f)) { expand_tabs(&line, buffer, strlen((char *)buffer)); incorporate_line(&line, linenum, &document); - gh_buf_clear(&line); + strbuf_clear(&line); linenum++; } - gh_buf_free(&line); + strbuf_free(&line); return finalize_document(document, linenum); } extern block *stmd_parse_document(const unsigned char *buffer, size_t len) { - gh_buf line = GH_BUF_INIT; + strbuf line = GH_BUF_INIT; int linenum = 1; const unsigned char *end = buffer + len; block *document = make_document(); @@ -444,11 +444,11 @@ extern block *stmd_parse_document(const unsigned char *buffer, size_t len) } incorporate_line(&line, linenum, &document); - gh_buf_clear(&line); + strbuf_clear(&line); linenum++; } - gh_buf_free(&line); + strbuf_free(&line); return finalize_document(document, linenum); } @@ -471,7 +471,7 @@ static void chop_trailing_hashtags(chunk *ch) } // Process one line at a time, modifying a block. -static void incorporate_line(gh_buf *line, int line_number, block** curptr) +static void incorporate_line(strbuf *line, int line_number, block** curptr) { block* last_matched_container; int offset = 0; @@ -639,8 +639,8 @@ static void incorporate_line(gh_buf *line, int line_number, block** curptr) } else if (container->tag == paragraph && (lev = scan_setext_header_line(&input, first_nonspace)) && // check that there is only one line in the paragraph: - gh_buf_strrchr(&container->string_content, '\n', - gh_buf_len(&container->string_content) - 2) < 0) { + strbuf_strrchr(&container->string_content, '\n', + strbuf_len(&container->string_content) - 2) < 0) { container->tag = setext_header; container->attributes.header_level = lev; @@ -734,7 +734,7 @@ static void incorporate_line(gh_buf *line, int line_number, block** curptr) container == last_matched_container && !blank && cur->tag == paragraph && - gh_buf_len(&cur->string_content) > 0) { + strbuf_len(&cur->string_content) > 0) { add_line(cur, &input, offset); diff --git a/src/buffer.c b/src/buffer.c index dc4a405..90c2186 100644 --- a/src/buffer.c +++ b/src/buffer.c @@ -9,32 +9,32 @@ #include "buffer.h" -/* Used as default value for gh_buf->ptr so that people can always - * assume ptr is non-NULL and zero terminated even for new gh_bufs. +/* Used as default value for strbuf->ptr so that people can always + * assume ptr is non-NULL and zero terminated even for new strbufs. */ -unsigned char gh_buf__initbuf[1]; -unsigned char gh_buf__oom[1]; +unsigned char strbuf__initbuf[1]; +unsigned char strbuf__oom[1]; #define ENSURE_SIZE(b, d) \ - if ((d) > buf->asize && gh_buf_grow(b, (d)) < 0)\ + if ((d) > buf->asize && strbuf_grow(b, (d)) < 0)\ return -1; -void gh_buf_init(gh_buf *buf, int initial_size) +void strbuf_init(strbuf *buf, int initial_size) { buf->asize = 0; buf->size = 0; - buf->ptr = gh_buf__initbuf; + buf->ptr = strbuf__initbuf; if (initial_size) - gh_buf_grow(buf, initial_size); + strbuf_grow(buf, initial_size); } -int gh_buf_try_grow(gh_buf *buf, int target_size, bool mark_oom) +int strbuf_try_grow(strbuf *buf, int target_size, bool mark_oom) { unsigned char *new_ptr; int new_size; - if (buf->ptr == gh_buf__oom) + if (buf->ptr == strbuf__oom) return -1; if (target_size <= buf->asize) @@ -60,7 +60,7 @@ int gh_buf_try_grow(gh_buf *buf, int target_size, bool mark_oom) if (!new_ptr) { if (mark_oom) - buf->ptr = gh_buf__oom; + buf->ptr = strbuf__oom; return -1; } @@ -75,17 +75,17 @@ int gh_buf_try_grow(gh_buf *buf, int target_size, bool mark_oom) return 0; } -void gh_buf_free(gh_buf *buf) +void strbuf_free(strbuf *buf) { if (!buf) return; - if (buf->ptr != gh_buf__initbuf && buf->ptr != gh_buf__oom) + if (buf->ptr != strbuf__initbuf && buf->ptr != strbuf__oom) free(buf->ptr); - gh_buf_init(buf, 0); + strbuf_init(buf, 0); } -void gh_buf_clear(gh_buf *buf) +void strbuf_clear(strbuf *buf) { buf->size = 0; @@ -93,10 +93,10 @@ void gh_buf_clear(gh_buf *buf) buf->ptr[0] = '\0'; } -int gh_buf_set(gh_buf *buf, const unsigned char *data, int len) +int strbuf_set(strbuf *buf, const unsigned char *data, int len) { if (len <= 0 || data == NULL) { - gh_buf_clear(buf); + strbuf_clear(buf); } else { if (data != buf->ptr) { ENSURE_SIZE(buf, len + 1); @@ -108,14 +108,14 @@ int gh_buf_set(gh_buf *buf, const unsigned char *data, int len) return 0; } -int gh_buf_sets(gh_buf *buf, const char *string) +int strbuf_sets(strbuf *buf, const char *string) { - return gh_buf_set(buf, + return strbuf_set(buf, (const unsigned char *)string, string ? strlen(string) : 0); } -int gh_buf_putc(gh_buf *buf, int c) +int strbuf_putc(strbuf *buf, int c) { ENSURE_SIZE(buf, buf->size + 2); buf->ptr[buf->size++] = c; @@ -123,7 +123,7 @@ int gh_buf_putc(gh_buf *buf, int c) return 0; } -int gh_buf_put(gh_buf *buf, const unsigned char *data, int len) +int strbuf_put(strbuf *buf, const unsigned char *data, int len) { if (len <= 0) return 0; @@ -135,12 +135,12 @@ int gh_buf_put(gh_buf *buf, const unsigned char *data, int len) return 0; } -int gh_buf_puts(gh_buf *buf, const char *string) +int strbuf_puts(strbuf *buf, const char *string) { - return gh_buf_put(buf, (const unsigned char *)string, strlen(string)); + return strbuf_put(buf, (const unsigned char *)string, strlen(string)); } -int gh_buf_vprintf(gh_buf *buf, const char *format, va_list ap) +int strbuf_vprintf(strbuf *buf, const char *format, va_list ap) { const int expected_size = buf->size + (strlen(format) * 2); int len; @@ -159,7 +159,7 @@ int gh_buf_vprintf(gh_buf *buf, const char *format, va_list ap) if (len < 0) { free(buf->ptr); - buf->ptr = gh_buf__oom; + buf->ptr = strbuf__oom; return -1; } @@ -174,19 +174,19 @@ int gh_buf_vprintf(gh_buf *buf, const char *format, va_list ap) return 0; } -int gh_buf_printf(gh_buf *buf, const char *format, ...) +int strbuf_printf(strbuf *buf, const char *format, ...) { int r; va_list ap; va_start(ap, format); - r = gh_buf_vprintf(buf, format, ap); + r = strbuf_vprintf(buf, format, ap); va_end(ap); return r; } -void gh_buf_copy_cstr(char *data, int datasize, const gh_buf *buf) +void strbuf_copy_cstr(char *data, int datasize, const strbuf *buf) { int copylen; @@ -204,28 +204,28 @@ void gh_buf_copy_cstr(char *data, int datasize, const gh_buf *buf) data[copylen] = '\0'; } -void gh_buf_swap(gh_buf *buf_a, gh_buf *buf_b) +void strbuf_swap(strbuf *buf_a, strbuf *buf_b) { - gh_buf t = *buf_a; + strbuf t = *buf_a; *buf_a = *buf_b; *buf_b = t; } -unsigned char *gh_buf_detach(gh_buf *buf) +unsigned char *strbuf_detach(strbuf *buf) { unsigned char *data = buf->ptr; - if (buf->asize == 0 || buf->ptr == gh_buf__oom) + if (buf->asize == 0 || buf->ptr == strbuf__oom) return NULL; - gh_buf_init(buf, 0); + strbuf_init(buf, 0); return data; } -void gh_buf_attach(gh_buf *buf, unsigned char *ptr, int asize) +void strbuf_attach(strbuf *buf, unsigned char *ptr, int asize) { - gh_buf_free(buf); + strbuf_free(buf); if (ptr) { buf->ptr = ptr; @@ -235,18 +235,18 @@ void gh_buf_attach(gh_buf *buf, unsigned char *ptr, int asize) else /* pass 0 to fall back on strlen + 1 */ buf->asize = buf->size + 1; } else { - gh_buf_grow(buf, asize); + strbuf_grow(buf, asize); } } -int gh_buf_cmp(const gh_buf *a, const gh_buf *b) +int strbuf_cmp(const strbuf *a, const strbuf *b) { int result = memcmp(a->ptr, b->ptr, MIN(a->size, b->size)); return (result != 0) ? result : (a->size < b->size) ? -1 : (a->size > b->size) ? 1 : 0; } -int gh_buf_strchr(const gh_buf *buf, int c, int pos) +int strbuf_strchr(const strbuf *buf, int c, int pos) { const unsigned char *p = memchr(buf->ptr + pos, c, buf->size - pos); if (!p) @@ -255,7 +255,7 @@ int gh_buf_strchr(const gh_buf *buf, int c, int pos) return (int)(p - (const unsigned char *)buf->ptr); } -int gh_buf_strrchr(const gh_buf *buf, int c, int pos) +int strbuf_strrchr(const strbuf *buf, int c, int pos) { int i; @@ -267,7 +267,7 @@ int gh_buf_strrchr(const gh_buf *buf, int c, int pos) return -1; } -void gh_buf_truncate(gh_buf *buf, int len) +void strbuf_truncate(strbuf *buf, int len) { if (len < buf->size) { buf->size = len; @@ -275,7 +275,7 @@ void gh_buf_truncate(gh_buf *buf, int len) } } -void gh_buf_drop(gh_buf *buf, int n) +void strbuf_drop(strbuf *buf, int n) { if (n > 0) { buf->size = buf->size - n; @@ -286,7 +286,7 @@ void gh_buf_drop(gh_buf *buf, int n) } } -void gh_buf_trim(gh_buf *buf) +void strbuf_trim(strbuf *buf) { int i = 0; @@ -296,7 +296,7 @@ void gh_buf_trim(gh_buf *buf) while (i < buf->size && isspace(buf->ptr[i])) i++; - gh_buf_drop(buf, i); + strbuf_drop(buf, i); /* rtrim */ while (buf->size > 0) { diff --git a/src/buffer.h b/src/buffer.h index 0d5143e..6f45cbb 100644 --- a/src/buffer.h +++ b/src/buffer.h @@ -9,20 +9,20 @@ typedef struct { unsigned char *ptr; int asize, size; -} gh_buf; +} strbuf; -extern unsigned char gh_buf__initbuf[]; -extern unsigned char gh_buf__oom[]; +extern unsigned char strbuf__initbuf[]; +extern unsigned char strbuf__oom[]; -#define GH_BUF_INIT { gh_buf__initbuf, 0, 0 } +#define GH_BUF_INIT { strbuf__initbuf, 0, 0 } /** - * Initialize a gh_buf structure. + * Initialize a strbuf structure. * * For the cases where GH_BUF_INIT cannot be used to do static * initialization. */ -extern void gh_buf_init(gh_buf *buf, int initial_size); +extern void strbuf_init(strbuf *buf, int initial_size); /** * Attempt to grow the buffer to hold at least `target_size` bytes. @@ -32,7 +32,7 @@ extern void gh_buf_init(gh_buf *buf, int initial_size); * existing buffer content will be preserved, but calling code must handle * that buffer was not expanded. */ -extern int gh_buf_try_grow(gh_buf *buf, int target_size, bool mark_oom); +extern int strbuf_try_grow(strbuf *buf, int target_size, bool mark_oom); /** * Grow the buffer to hold at least `target_size` bytes. @@ -42,71 +42,71 @@ extern int gh_buf_try_grow(gh_buf *buf, int target_size, bool mark_oom); * * @return 0 on success or -1 on failure */ -static inline int gh_buf_grow(gh_buf *buf, int target_size) +static inline int strbuf_grow(strbuf *buf, int target_size) { - return gh_buf_try_grow(buf, target_size, true); + return strbuf_try_grow(buf, target_size, true); } -extern void gh_buf_free(gh_buf *buf); -extern void gh_buf_swap(gh_buf *buf_a, gh_buf *buf_b); +extern void strbuf_free(strbuf *buf); +extern void strbuf_swap(strbuf *buf_a, strbuf *buf_b); /** - * Test if there have been any reallocation failures with this gh_buf. + * Test if there have been any reallocation failures with this strbuf. * - * Any function that writes to a gh_buf can fail due to memory allocation - * issues. If one fails, the gh_buf will be marked with an OOM error and - * further calls to modify the buffer will fail. Check gh_buf_oom() at the + * Any function that writes to a strbuf can fail due to memory allocation + * issues. If one fails, the strbuf will be marked with an OOM error and + * further calls to modify the buffer will fail. Check strbuf_oom() at the * end of your sequence and it will be true if you ran out of memory at any * point with that buffer. * * @return false if no error, true if allocation error */ -static inline bool gh_buf_oom(const gh_buf *buf) +static inline bool strbuf_oom(const strbuf *buf) { - return (buf->ptr == gh_buf__oom); + return (buf->ptr == strbuf__oom); } -static inline size_t gh_buf_len(const gh_buf *buf) +static inline size_t strbuf_len(const strbuf *buf) { return buf->size; } -extern int gh_buf_cmp(const gh_buf *a, const gh_buf *b); +extern int strbuf_cmp(const strbuf *a, const strbuf *b); -extern void gh_buf_attach(gh_buf *buf, unsigned char *ptr, int asize); -extern unsigned char *gh_buf_detach(gh_buf *buf); -extern void gh_buf_copy_cstr(char *data, int datasize, const gh_buf *buf); +extern void strbuf_attach(strbuf *buf, unsigned char *ptr, int asize); +extern unsigned char *strbuf_detach(strbuf *buf); +extern void strbuf_copy_cstr(char *data, int datasize, const strbuf *buf); -static inline const char *gh_buf_cstr(const gh_buf *buf) +static inline const char *strbuf_cstr(const strbuf *buf) { return (char *)buf->ptr; } -#define gh_buf_at(buf, n) ((buf)->ptr[n]) +#define strbuf_at(buf, n) ((buf)->ptr[n]) /* * Functions below that return int value error codes will return 0 on * success or -1 on failure (which generally means an allocation failed). - * Using a gh_buf where the allocation has failed with result in -1 from + * Using a strbuf where the allocation has failed with result in -1 from * all further calls using that buffer. As a result, you can ignore the * return code of these functions and call them in a series then just call - * gh_buf_oom at the end. + * strbuf_oom at the end. */ -extern int gh_buf_set(gh_buf *buf, const unsigned char *data, int len); -extern int gh_buf_sets(gh_buf *buf, const char *string); -extern int gh_buf_putc(gh_buf *buf, int c); -extern int gh_buf_put(gh_buf *buf, const unsigned char *data, int len); -extern int gh_buf_puts(gh_buf *buf, const char *string); -extern int gh_buf_printf(gh_buf *buf, const char *format, ...) +extern int strbuf_set(strbuf *buf, const unsigned char *data, int len); +extern int strbuf_sets(strbuf *buf, const char *string); +extern int strbuf_putc(strbuf *buf, int c); +extern int strbuf_put(strbuf *buf, const unsigned char *data, int len); +extern int strbuf_puts(strbuf *buf, const char *string); +extern int strbuf_printf(strbuf *buf, const char *format, ...) __attribute__((format (printf, 2, 3))); -extern int gh_buf_vprintf(gh_buf *buf, const char *format, va_list ap); -extern void gh_buf_clear(gh_buf *buf); - -int gh_buf_strchr(const gh_buf *buf, int c, int pos); -int gh_buf_strrchr(const gh_buf *buf, int c, int pos); -void gh_buf_drop(gh_buf *buf, int n); -void gh_buf_truncate(gh_buf *buf, int len); -void gh_buf_trim(gh_buf *buf); +extern int strbuf_vprintf(strbuf *buf, const char *format, va_list ap); +extern void strbuf_clear(strbuf *buf); + +int strbuf_strchr(const strbuf *buf, int c, int pos); +int strbuf_strrchr(const strbuf *buf, int c, int pos); +void strbuf_drop(strbuf *buf, int n); +void strbuf_truncate(strbuf *buf, int len); +void strbuf_trim(strbuf *buf); #endif diff --git a/src/chunk.h b/src/chunk.h index f3841ed..f37a2f3 100644 --- a/src/chunk.h +++ b/src/chunk.h @@ -78,12 +78,12 @@ static inline chunk chunk_dup(const chunk *ch, int pos, int len) return c; } -static inline chunk chunk_buf_detach(gh_buf *buf) +static inline chunk chunk_buf_detach(strbuf *buf) { chunk c; c.len = buf->size; - c.data = gh_buf_detach(buf); + c.data = strbuf_detach(buf); c.alloc = 1; return c; diff --git a/src/html/houdini.h b/src/html/houdini.h index 31fe917..1e54d20 100644 --- a/src/html/houdini.h +++ b/src/html/houdini.h @@ -25,17 +25,17 @@ extern "C" { #define HOUDINI_ESCAPED_SIZE(x) (((x) * 12) / 10) #define HOUDINI_UNESCAPED_SIZE(x) (x) -extern int houdini_escape_html(gh_buf *ob, const uint8_t *src, size_t size); -extern int houdini_escape_html0(gh_buf *ob, const uint8_t *src, size_t size, int secure); -extern int houdini_unescape_html(gh_buf *ob, const uint8_t *src, size_t size); -extern int houdini_escape_xml(gh_buf *ob, const uint8_t *src, size_t size); -extern int houdini_escape_uri(gh_buf *ob, const uint8_t *src, size_t size); -extern int houdini_escape_url(gh_buf *ob, const uint8_t *src, size_t size); -extern int houdini_escape_href(gh_buf *ob, const uint8_t *src, size_t size); -extern int houdini_unescape_uri(gh_buf *ob, const uint8_t *src, size_t size); -extern int houdini_unescape_url(gh_buf *ob, const uint8_t *src, size_t size); -extern int houdini_escape_js(gh_buf *ob, const uint8_t *src, size_t size); -extern int houdini_unescape_js(gh_buf *ob, const uint8_t *src, size_t size); +extern int houdini_escape_html(strbuf *ob, const uint8_t *src, size_t size); +extern int houdini_escape_html0(strbuf *ob, const uint8_t *src, size_t size, int secure); +extern int houdini_unescape_html(strbuf *ob, const uint8_t *src, size_t size); +extern int houdini_escape_xml(strbuf *ob, const uint8_t *src, size_t size); +extern int houdini_escape_uri(strbuf *ob, const uint8_t *src, size_t size); +extern int houdini_escape_url(strbuf *ob, const uint8_t *src, size_t size); +extern int houdini_escape_href(strbuf *ob, const uint8_t *src, size_t size); +extern int houdini_unescape_uri(strbuf *ob, const uint8_t *src, size_t size); +extern int houdini_unescape_url(strbuf *ob, const uint8_t *src, size_t size); +extern int houdini_escape_js(strbuf *ob, const uint8_t *src, size_t size); +extern int houdini_unescape_js(strbuf *ob, const uint8_t *src, size_t size); #ifdef __cplusplus } diff --git a/src/html/houdini_href_e.c b/src/html/houdini_href_e.c index b2a7d79..12456ce 100644 --- a/src/html/houdini_href_e.c +++ b/src/html/houdini_href_e.c @@ -49,7 +49,7 @@ static const char HREF_SAFE[] = { }; int -houdini_escape_href(gh_buf *ob, const uint8_t *src, size_t size) +houdini_escape_href(strbuf *ob, const uint8_t *src, size_t size) { static const uint8_t hex_chars[] = "0123456789ABCDEF"; size_t i = 0, org; @@ -63,7 +63,7 @@ houdini_escape_href(gh_buf *ob, const uint8_t *src, size_t size) i++; if (likely(i > org)) - gh_buf_put(ob, src + org, i - org); + strbuf_put(ob, src + org, i - org); /* escaping */ if (i >= size) @@ -73,14 +73,14 @@ houdini_escape_href(gh_buf *ob, const uint8_t *src, size_t size) /* amp appears all the time in URLs, but needs * HTML-entity escaping to be inside an href */ case '&': - gh_buf_puts(ob, "&"); + strbuf_puts(ob, "&"); break; /* the single quote is a valid URL character * according to the standard; it needs HTML * entity escaping too */ case '\'': - gh_buf_puts(ob, "'"); + strbuf_puts(ob, "'"); break; /* the space can be escaped to %20 or a plus @@ -89,7 +89,7 @@ houdini_escape_href(gh_buf *ob, const uint8_t *src, size_t size) * when building GET strings */ #if 0 case ' ': - gh_buf_putc(ob, '+'); + strbuf_putc(ob, '+'); break; #endif @@ -97,7 +97,7 @@ houdini_escape_href(gh_buf *ob, const uint8_t *src, size_t size) default: hex_str[1] = hex_chars[(src[i] >> 4) & 0xF]; hex_str[2] = hex_chars[src[i] & 0xF]; - gh_buf_put(ob, hex_str, 3); + strbuf_put(ob, hex_str, 3); } i++; diff --git a/src/html/houdini_html_e.c b/src/html/houdini_html_e.c index 95b6c41..f2e86fe 100644 --- a/src/html/houdini_html_e.c +++ b/src/html/houdini_html_e.c @@ -45,7 +45,7 @@ static const char *HTML_ESCAPES[] = { }; int -houdini_escape_html0(gh_buf *ob, const uint8_t *src, size_t size, int secure) +houdini_escape_html0(strbuf *ob, const uint8_t *src, size_t size, int secure) { size_t i = 0, org, esc = 0; @@ -55,7 +55,7 @@ houdini_escape_html0(gh_buf *ob, const uint8_t *src, size_t size, int secure) i++; if (i > org) - gh_buf_put(ob, src + org, i - org); + strbuf_put(ob, src + org, i - org); /* escaping */ if (unlikely(i >= size)) @@ -63,9 +63,9 @@ houdini_escape_html0(gh_buf *ob, const uint8_t *src, size_t size, int secure) /* The forward slash is only escaped in secure mode */ if ((src[i] == '/' || src[i] == '\'') && !secure) { - gh_buf_putc(ob, src[i]); + strbuf_putc(ob, src[i]); } else { - gh_buf_puts(ob, HTML_ESCAPES[esc]); + strbuf_puts(ob, HTML_ESCAPES[esc]); } i++; @@ -75,7 +75,7 @@ houdini_escape_html0(gh_buf *ob, const uint8_t *src, size_t size, int secure) } int -houdini_escape_html(gh_buf *ob, const uint8_t *src, size_t size) +houdini_escape_html(strbuf *ob, const uint8_t *src, size_t size) { return houdini_escape_html0(ob, src, size, 1); } diff --git a/src/html/html.c b/src/html/html.c index 41b8fda..a9356dd 100644 --- a/src/html/html.c +++ b/src/html/html.c @@ -10,7 +10,7 @@ // Functions to convert block and inline lists to HTML strings. -static void escape_html(gh_buf *dest, const unsigned char *source, int length) +static void escape_html(strbuf *dest, const unsigned char *source, int length) { if (length < 0) length = strlen((char *)source); @@ -18,7 +18,7 @@ static void escape_html(gh_buf *dest, const unsigned char *source, int length) houdini_escape_html0(dest, source, (size_t)length, 0); } -static void escape_href(gh_buf *dest, const unsigned char *source, int length) +static void escape_href(strbuf *dest, const unsigned char *source, int length) { if (length < 0) length = strlen((char *)source); @@ -26,14 +26,14 @@ static void escape_href(gh_buf *dest, const unsigned char *source, int length) houdini_escape_href(dest, source, (size_t)length); } -static inline void cr(gh_buf *html) +static inline void cr(strbuf *html) { if (html->size && html->ptr[html->size - 1] != '\n') - gh_buf_putc(html, '\n'); + strbuf_putc(html, '\n'); } // Convert a block list to HTML. Returns 0 on success, and sets result. -void blocks_to_html(gh_buf *html, block *b, bool tight) +void blocks_to_html(strbuf *html, block *b, bool tight) { struct ListData *data; @@ -48,25 +48,25 @@ void blocks_to_html(gh_buf *html, block *b, bool tight) inlines_to_html(html, b->inline_content); } else { cr(html); - gh_buf_puts(html, "

    "); + strbuf_puts(html, "

    "); inlines_to_html(html, b->inline_content); - gh_buf_puts(html, "

    \n"); + strbuf_puts(html, "

    \n"); } break; case block_quote: cr(html); - gh_buf_puts(html, "
    \n"); + strbuf_puts(html, "
    \n"); blocks_to_html(html, b->children, false); - gh_buf_puts(html, "
    \n"); + strbuf_puts(html, "
    \n"); break; case list_item: cr(html); - gh_buf_puts(html, "
  • "); + strbuf_puts(html, "
  • "); blocks_to_html(html, b->children, tight); - gh_buf_trim(html); /* TODO: rtrim */ - gh_buf_puts(html, "
    • \n"); + strbuf_trim(html); /* TODO: rtrim */ + strbuf_puts(html, "\n"); break; case list: @@ -75,58 +75,58 @@ void blocks_to_html(gh_buf *html, block *b, bool tight) data = &(b->attributes.list_data); if (data->start > 1) { - gh_buf_printf(html, "<%s start=\"%d\">\n", + strbuf_printf(html, "<%s start=\"%d\">\n", data->list_type == bullet ? "ul" : "ol", data->start); } else { - gh_buf_puts(html, data->list_type == bullet ? "
      \n" : "
        \n"); + strbuf_puts(html, data->list_type == bullet ? "
          \n" : "
            \n"); } blocks_to_html(html, b->children, data->tight); - gh_buf_puts(html, data->list_type == bullet ? "
        " : "
      "); - gh_buf_putc(html, '\n'); + strbuf_puts(html, data->list_type == bullet ? "
    " : ""); + strbuf_putc(html, '\n'); break; case atx_header: case setext_header: cr(html); - gh_buf_printf(html, "", b->attributes.header_level); + strbuf_printf(html, "", b->attributes.header_level); inlines_to_html(html, b->inline_content); - gh_buf_printf(html, "\n", b->attributes.header_level); + strbuf_printf(html, "\n", b->attributes.header_level); break; case indented_code: case fenced_code: cr(html); - gh_buf_puts(html, "tag == fenced_code) { - gh_buf *info = &b->attributes.fenced_code_data.info; + strbuf *info = &b->attributes.fenced_code_data.info; - if (gh_buf_len(info) > 0) { - int first_tag = gh_buf_strchr(info, ' ', 0); + if (strbuf_len(info) > 0) { + int first_tag = strbuf_strchr(info, ' ', 0); if (first_tag < 0) - first_tag = gh_buf_len(info); + first_tag = strbuf_len(info); - gh_buf_puts(html, " class=\""); + strbuf_puts(html, " class=\""); escape_html(html, info->ptr, first_tag); - gh_buf_putc(html, '"'); + strbuf_putc(html, '"'); } } - gh_buf_puts(html, ">"); + strbuf_puts(html, ">"); escape_html(html, b->string_content.ptr, b->string_content.size); - gh_buf_puts(html, "\n"); + strbuf_puts(html, "\n"); break; case html_block: - gh_buf_put(html, b->string_content.ptr, b->string_content.size); + strbuf_put(html, b->string_content.ptr, b->string_content.size); break; case hrule: - gh_buf_puts(html, "
    \n"); + strbuf_puts(html, "
    \n"); break; case reference_def: @@ -141,9 +141,9 @@ void blocks_to_html(gh_buf *html, block *b, bool tight) } // Convert an inline list to HTML. Returns 0 on success, and sets result. -void inlines_to_html(gh_buf *html, inl* ils) +void inlines_to_html(strbuf *html, inl* ils) { - gh_buf scrap = GH_BUF_INIT; + strbuf scrap = GH_BUF_INIT; while(ils != NULL) { switch(ils->tag) { @@ -152,70 +152,70 @@ void inlines_to_html(gh_buf *html, inl* ils) break; case INL_LINEBREAK: - gh_buf_puts(html, "
    \n"); + strbuf_puts(html, "
    \n"); break; case INL_SOFTBREAK: - gh_buf_putc(html, '\n'); + strbuf_putc(html, '\n'); break; case INL_CODE: - gh_buf_puts(html, ""); + strbuf_puts(html, ""); escape_html(html, ils->content.literal.data, ils->content.literal.len); - gh_buf_puts(html, ""); + strbuf_puts(html, ""); break; case INL_RAW_HTML: case INL_ENTITY: - gh_buf_put(html, + strbuf_put(html, ils->content.literal.data, ils->content.literal.len); break; case INL_LINK: - gh_buf_puts(html, "content.linkable.url) escape_href(html, ils->content.linkable.url, -1); if (ils->content.linkable.title) { - gh_buf_puts(html, "\" title=\""); + strbuf_puts(html, "\" title=\""); escape_html(html, ils->content.linkable.title, -1); } - gh_buf_puts(html, "\">"); + strbuf_puts(html, "\">"); inlines_to_html(html, ils->content.inlines); - gh_buf_puts(html, ""); + strbuf_puts(html, ""); break; case INL_IMAGE: - gh_buf_puts(html, "content.linkable.url) escape_href(html, ils->content.linkable.url, -1); inlines_to_html(&scrap, ils->content.inlines); - gh_buf_puts(html, "\" alt=\""); + strbuf_puts(html, "\" alt=\""); if (scrap.size) escape_html(html, scrap.ptr, scrap.size); - gh_buf_clear(&scrap); + strbuf_clear(&scrap); if (ils->content.linkable.title) { - gh_buf_puts(html, "\" title=\""); + strbuf_puts(html, "\" title=\""); escape_html(html, ils->content.linkable.title, -1); } - gh_buf_puts(html, "\"/>"); + strbuf_puts(html, "\"/>"); break; case INL_STRONG: - gh_buf_puts(html, ""); + strbuf_puts(html, ""); inlines_to_html(html, ils->content.inlines); - gh_buf_puts(html, ""); + strbuf_puts(html, ""); break; case INL_EMPH: - gh_buf_puts(html, ""); + strbuf_puts(html, ""); inlines_to_html(html, ils->content.inlines); - gh_buf_puts(html, ""); + strbuf_puts(html, ""); break; } ils = ils->next; diff --git a/src/inlines.c b/src/inlines.c index 8e2e683..33973df 100644 --- a/src/inlines.c +++ b/src/inlines.c @@ -25,7 +25,7 @@ inline static void chunk_free(chunk *c); inline static void chunk_trim(chunk *c); inline static chunk chunk_literal(const char *data); -inline static chunk chunk_buf_detach(gh_buf *buf); +inline static chunk chunk_buf_detach(strbuf *buf); inline static chunk chunk_dup(const chunk *ch, int pos, int len); static inl *parse_chunk_inlines(chunk *chunk, reference** refmap); @@ -33,10 +33,10 @@ static inl *parse_inlines_while(subject* subj, int (*f)(subject*)); static int parse_inline(subject* subj, inl ** last); static void subject_from_chunk(subject *e, chunk *chunk, reference** refmap); -static void subject_from_buf(subject *e, gh_buf *buffer, reference** refmap); +static void subject_from_buf(subject *e, strbuf *buffer, reference** refmap); static int subject_find_special_char(subject *subj); -static void normalize_whitespace(gh_buf *s); +static void normalize_whitespace(strbuf *s); extern void free_reference(reference *ref) { free(ref->label); @@ -62,13 +62,13 @@ extern void free_reference_map(reference **refmap) { // remove leading/trailing whitespace, case fold static unsigned char *normalize_reference(chunk *ref) { - gh_buf normalized = GH_BUF_INIT; + strbuf normalized = GH_BUF_INIT; utf8proc_case_fold(&normalized, ref->data, ref->len); - gh_buf_trim(&normalized); + strbuf_trim(&normalized); normalize_whitespace(&normalized); - return gh_buf_detach(&normalized); + return strbuf_detach(&normalized); } // Returns reference if refmap contains a reference with matching @@ -218,7 +218,7 @@ inline static inl* append_inlines(inl* a, inl* b) return a; } -static void subject_from_buf(subject *e, gh_buf *buffer, reference** refmap) +static void subject_from_buf(subject *e, strbuf *buffer, reference** refmap) { e->input.data = buffer->ptr; e->input.len = buffer->size; @@ -309,7 +309,7 @@ static int scan_to_closing_backticks(subject* subj, int openticklength) // Destructively modify string, collapsing consecutive // space and newline characters into a single space. -static void normalize_whitespace(gh_buf *s) +static void normalize_whitespace(strbuf *s) { bool last_char_was_space = false; int r, w; @@ -331,7 +331,7 @@ static void normalize_whitespace(gh_buf *s) } } - gh_buf_truncate(s, w); + strbuf_truncate(s, w); } // Parse backtick code section or raw backticks, return an inline. @@ -346,10 +346,10 @@ static inl* handle_backticks(subject *subj) subj->pos = startpos; // rewind return make_str(openticks); } else { - gh_buf buf = GH_BUF_INIT; + strbuf buf = GH_BUF_INIT; - gh_buf_set(&buf, subj->input.data + startpos, endpos - startpos - openticks.len); - gh_buf_trim(&buf); + strbuf_set(&buf, subj->input.data + startpos, endpos - startpos - openticks.len); + strbuf_trim(&buf); normalize_whitespace(&buf); return make_code(chunk_buf_detach(&buf)); @@ -569,7 +569,7 @@ static inl *make_str_with_entities(chunk *content) } // Destructively unescape a string: remove backslashes before punctuation chars. -extern void unescape_buffer(gh_buf *buf) +extern void unescape_buffer(strbuf *buf) { int r, w; @@ -580,14 +580,14 @@ extern void unescape_buffer(gh_buf *buf) buf->ptr[w++] = buf->ptr[r]; } - gh_buf_truncate(buf, w); + strbuf_truncate(buf, w); } // Clean a URL: remove surrounding whitespace and surrounding <>, // and remove \ that escape punctuation. static unsigned char *clean_url(chunk *url, int is_email) { - gh_buf buf = GH_BUF_INIT; + strbuf buf = GH_BUF_INIT; chunk_trim(url); @@ -595,22 +595,22 @@ static unsigned char *clean_url(chunk *url, int is_email) return NULL; if (is_email) - gh_buf_puts(&buf, "mailto:"); + strbuf_puts(&buf, "mailto:"); if (url->data[0] == '<' && url->data[url->len - 1] == '>') { - gh_buf_put(&buf, url->data + 1, url->len - 2); + strbuf_put(&buf, url->data + 1, url->len - 2); } else { - gh_buf_put(&buf, url->data, url->len); + strbuf_put(&buf, url->data, url->len); } unescape_buffer(&buf); - return gh_buf_detach(&buf); + return strbuf_detach(&buf); } // Clean a title: remove surrounding quotes and remove \ that escape punctuation. static unsigned char *clean_title(chunk *title) { - gh_buf buf = GH_BUF_INIT; + strbuf buf = GH_BUF_INIT; unsigned char first, last; if (title->len == 0) @@ -623,13 +623,13 @@ static unsigned char *clean_title(chunk *title) if ((first == '\'' && last == '\'') || (first == '(' && last == ')') || (first == '"' && last == '"')) { - gh_buf_set(&buf, title->data + 1, title->len - 2); + strbuf_set(&buf, title->data + 1, title->len - 2); } else { - gh_buf_set(&buf, title->data, title->len); + strbuf_set(&buf, title->data, title->len); } unescape_buffer(&buf); - return gh_buf_detach(&buf); + return strbuf_detach(&buf); } // Parse an autolink or HTML tag. @@ -971,7 +971,7 @@ static int parse_inline(subject* subj, inl ** last) return 1; } -extern inl* parse_inlines(gh_buf *input, reference** refmap) +extern inl* parse_inlines(strbuf *input, reference** refmap) { subject subj; subject_from_buf(&subj, input, refmap); @@ -993,7 +993,7 @@ void spnl(subject* subj) // Modify refmap if a reference is encountered. // Return 0 if no reference found, otherwise position of subject // after reference is parsed. -extern int parse_reference(gh_buf *input, reference** refmap) +extern int parse_reference(strbuf *input, reference** refmap) { subject subj; diff --git a/src/main.c b/src/main.c index e1abedc..7cf67e2 100644 --- a/src/main.c +++ b/src/main.c @@ -14,14 +14,14 @@ void print_usage() static void print_document(block *document, bool ast) { - gh_buf html = GH_BUF_INIT; + strbuf html = GH_BUF_INIT; if (ast) { print_blocks(document, 0); } else { blocks_to_html(&html, document, false); printf("%s", html.ptr); - gh_buf_free(&html); + strbuf_free(&html); } } diff --git a/src/stmd.h b/src/stmd.h index 4a3c399..2e86f3a 100644 --- a/src/stmd.h +++ b/src/stmd.h @@ -50,7 +50,7 @@ struct FencedCodeData { int fence_length; int fence_offset; char fence_char; - gh_buf info; + strbuf info; }; typedef struct Block { @@ -76,7 +76,7 @@ typedef struct Block { struct Block* last_child; struct Block* parent; struct Block* top; - gh_buf string_content; + strbuf string_content; inl* inline_content; union { struct ListData list_data; @@ -88,15 +88,15 @@ typedef struct Block { struct Block * prev; } block; -inl* parse_inlines(gh_buf *input, reference** refmap); +inl* parse_inlines(strbuf *input, reference** refmap); void free_inlines(inl* e); -int parse_reference(gh_buf *input, reference** refmap); +int parse_reference(strbuf *input, reference** refmap); void free_reference(reference *ref); void free_reference_map(reference **refmap); void add_reference(reference** refmap, reference* ref); -void unescape_buffer(gh_buf *buf); +void unescape_buffer(strbuf *buf); extern block* make_document(); extern block* add_child(block* parent, @@ -109,9 +109,9 @@ extern block *stmd_parse_file(FILE *f); void print_inlines(inl* ils, int indent); void print_blocks(block* blk, int indent); -void blocks_to_html(gh_buf *html, block *b, bool tight); -void inlines_to_html(gh_buf *html, inl *b); +void blocks_to_html(strbuf *html, block *b, bool tight); +void inlines_to_html(strbuf *html, inl *b); -void utf8proc_case_fold(gh_buf *dest, const unsigned char *str, int len); +void utf8proc_case_fold(strbuf *dest, const unsigned char *str, int len); #endif diff --git a/src/utf8.c b/src/utf8.c index 32c78a4..cebd872 100644 --- a/src/utf8.c +++ b/src/utf8.c @@ -84,7 +84,7 @@ ssize_t utf8proc_iterate(const uint8_t *str, ssize_t str_len, int32_t *dst) return length; } -void utf8proc_encode_char(int32_t uc, gh_buf *buf) +void utf8proc_encode_char(int32_t uc, strbuf *buf) { unsigned char dst[4]; int len = 0; @@ -119,10 +119,10 @@ void utf8proc_encode_char(int32_t uc, gh_buf *buf) assert(false); } - gh_buf_put(buf, dst, len); + strbuf_put(buf, dst, len); } -void utf8proc_case_fold(gh_buf *dest, const unsigned char *str, int len) +void utf8proc_case_fold(strbuf *dest, const unsigned char *str, int len) { int32_t c; -- cgit v1.2.3 From 94a79a605f3e76a43f1f87a5044f6761b99e5ca5 Mon Sep 17 00:00:00 2001 From: Vicent Marti Date: Wed, 10 Sep 2014 18:33:27 +0200 Subject: Cleanup reference implementation --- Makefile | 2 +- src/blocks.c | 16 ++--- src/buffer.c | 43 ++++++++++++++ src/buffer.h | 2 + src/inlines.c | 176 +++++++------------------------------------------------ src/references.c | 109 ++++++++++++++++++++++++++++++++++ src/references.h | 27 +++++++++ src/stmd.h | 26 +++----- src/utf8.c | 10 ++-- src/utf8.h | 5 +- 10 files changed, 225 insertions(+), 191 deletions(-) create mode 100644 src/references.c create mode 100644 src/references.h (limited to 'src/buffer.h') diff --git a/Makefile b/Makefile index 5d13272..11e2141 100644 --- a/Makefile +++ b/Makefile @@ -42,7 +42,7 @@ benchjs: node js/bench.js ${BENCHINP} HTML_OBJ=$(SRCDIR)/html/html.o $(SRCDIR)/html/houdini_href_e.o $(SRCDIR)/html/houdini_html_e.o $(SRCDIR)/html/houdini_html_u.o -STMD_OBJ=$(SRCDIR)/inlines.o $(SRCDIR)/buffer.o $(SRCDIR)/blocks.o $(SRCDIR)/scanners.c $(SRCDIR)/print.o $(SRCDIR)/utf8.o +STMD_OBJ=$(SRCDIR)/inlines.o $(SRCDIR)/buffer.o $(SRCDIR)/blocks.o $(SRCDIR)/scanners.c $(SRCDIR)/print.o $(SRCDIR)/utf8.o $(SRCDIR)/references.c $(PROG): $(SRCDIR)/html/html_unescape.h $(SRCDIR)/case_fold_switch.inc $(HTML_OBJ) $(STMD_OBJ) $(SRCDIR)/main.c $(CC) $(LDFLAGS) -o $@ $(HTML_OBJ) $(STMD_OBJ) $(SRCDIR)/main.c diff --git a/src/blocks.c b/src/blocks.c index 72b2dc2..30a8284 100644 --- a/src/blocks.c +++ b/src/blocks.c @@ -8,7 +8,6 @@ #include "utf8.h" #include "html/houdini.h" #include "scanners.h" -#include "uthash.h" #define peek_at(i, n) (i)->data[n] @@ -36,12 +35,7 @@ static node_block* make_block(int tag, int start_line, int start_column) extern node_block* make_document() { node_block *e = make_block(BLOCK_DOCUMENT, 1, 1); - reference *map = NULL; - reference ** refmap; - - refmap = (reference**) malloc(sizeof(reference*)); - *refmap = map; - e->as.document.refmap = refmap; + e->as.document.refmap = reference_map_new(); e->top = e; return e; @@ -164,7 +158,7 @@ static void finalize(node_block* b, int line_number) case BLOCK_PARAGRAPH: pos = 0; while (strbuf_at(&b->string_content, 0) == '[' && - (pos = parse_reference(&b->string_content, b->top->as.document.refmap))) { + (pos = parse_reference_inline(&b->string_content, b->top->as.document.refmap))) { strbuf_drop(&b->string_content, pos); } @@ -192,7 +186,7 @@ static void finalize(node_block* b, int line_number) strbuf_drop(&b->string_content, firstlinelen + 1); strbuf_trim(&b->as.code.info); - unescape_buffer(&b->as.code.info); + strbuf_unescape(&b->as.code.info); break; case BLOCK_LIST: // determine tight/loose status @@ -268,7 +262,7 @@ extern void free_blocks(node_block* e) if (e->tag == BLOCK_FENCED_CODE) { strbuf_free(&e->as.code.info); } else if (e->tag == BLOCK_DOCUMENT) { - free_reference_map(e->as.document.refmap); + reference_map_free(e->as.document.refmap); } free_blocks(e->children); free(e); @@ -278,7 +272,7 @@ extern void free_blocks(node_block* e) // Walk through node_block and all children, recursively, parsing // string content into inline content where appropriate. -void process_inlines(node_block* cur, reference** refmap) +void process_inlines(node_block* cur, reference_map *refmap) { switch (cur->tag) { case BLOCK_PARAGRAPH: diff --git a/src/buffer.c b/src/buffer.c index 90c2186..cdf8ca0 100644 --- a/src/buffer.c +++ b/src/buffer.c @@ -308,3 +308,46 @@ void strbuf_trim(strbuf *buf) buf->ptr[buf->size] = '\0'; } + +// Destructively modify string, collapsing consecutive +// space and newline characters into a single space. +void strbuf_normalize_whitespace(strbuf *s) +{ + bool last_char_was_space = false; + int r, w; + + for (r = 0, w = 0; r < s->size; ++r) { + switch (s->ptr[r]) { + case ' ': + case '\n': + if (last_char_was_space) + break; + + s->ptr[w++] = ' '; + last_char_was_space = true; + break; + + default: + s->ptr[w++] = s->ptr[r]; + last_char_was_space = false; + } + } + + strbuf_truncate(s, w); +} + +// Destructively unescape a string: remove backslashes before punctuation chars. +extern void strbuf_unescape(strbuf *buf) +{ + int r, w; + + for (r = 0, w = 0; r < buf->size; ++r) { + if (buf->ptr[r] == '\\' && ispunct(buf->ptr[r + 1])) + continue; + + buf->ptr[w++] = buf->ptr[r]; + } + + strbuf_truncate(buf, w); +} + diff --git a/src/buffer.h b/src/buffer.h index 6f45cbb..1bc1eee 100644 --- a/src/buffer.h +++ b/src/buffer.h @@ -108,5 +108,7 @@ int strbuf_strrchr(const strbuf *buf, int c, int pos); void strbuf_drop(strbuf *buf, int n); void strbuf_truncate(strbuf *buf, int len); void strbuf_trim(strbuf *buf); +void strbuf_normalize_whitespace(strbuf *s); +void strbuf_unescape(strbuf *s); #endif diff --git a/src/inlines.c b/src/inlines.c index aa0e13e..3040f09 100644 --- a/src/inlines.c +++ b/src/inlines.c @@ -7,110 +7,23 @@ #include "stmd.h" #include "html/houdini.h" #include "utf8.h" -#include "uthash.h" #include "scanners.h" typedef struct Subject { chunk input; int pos; - int label_nestlevel; - reference** reference_map; + int label_nestlevel; + reference_map *refmap; } subject; -reference* lookup_reference(reference** refmap, chunk *label); -reference* make_reference(chunk *label, chunk *url, chunk *title); - -static unsigned char *clean_url(chunk *url); -static unsigned char *clean_title(chunk *title); -static unsigned char *clean_autolink(chunk *url, int is_email); - -inline static void chunk_free(chunk *c); -inline static void chunk_trim(chunk *c); - -inline static chunk chunk_literal(const char *data); -inline static chunk chunk_buf_detach(strbuf *buf); -inline static chunk chunk_dup(const chunk *ch, int pos, int len); - -static node_inl *parse_chunk_inlines(chunk *chunk, reference** refmap); +static node_inl *parse_chunk_inlines(chunk *chunk, reference_map *refmap); static node_inl *parse_inlines_while(subject* subj, int (*f)(subject*)); static int parse_inline(subject* subj, node_inl ** last); -static void subject_from_chunk(subject *e, chunk *chunk, reference** refmap); -static void subject_from_buf(subject *e, strbuf *buffer, reference** refmap); +static void subject_from_chunk(subject *e, chunk *chunk, reference_map *refmap); +static void subject_from_buf(subject *e, strbuf *buffer, reference_map *refmap); static int subject_find_special_char(subject *subj); -static void normalize_whitespace(strbuf *s); - -extern void free_reference(reference *ref) { - free(ref->label); - free(ref->url); - free(ref->title); - free(ref); -} - -extern void free_reference_map(reference **refmap) { - /* free the hash table contents */ - reference *s; - reference *tmp; - if (refmap != NULL) { - HASH_ITER(hh, *refmap, s, tmp) { - HASH_DEL(*refmap, s); - free_reference(s); - } - free(refmap); - } -} - -// normalize reference: collapse internal whitespace to single space, -// remove leading/trailing whitespace, case fold -static unsigned char *normalize_reference(chunk *ref) -{ - strbuf normalized = GH_BUF_INIT; - - utf8proc_case_fold(&normalized, ref->data, ref->len); - strbuf_trim(&normalized); - normalize_whitespace(&normalized); - - return strbuf_detach(&normalized); -} - -// Returns reference if refmap contains a reference with matching -// label, otherwise NULL. -extern reference* lookup_reference(reference** refmap, chunk *label) -{ - reference *ref = NULL; - unsigned char *norm = normalize_reference(label); - if (refmap != NULL) { - HASH_FIND_STR(*refmap, (char*)norm, ref); - } - free(norm); - return ref; -} - -extern reference* make_reference(chunk *label, chunk *url, chunk *title) -{ - reference *ref; - ref = malloc(sizeof(reference)); - ref->label = normalize_reference(label); - ref->url = clean_url(url); - ref->title = clean_title(title); - return ref; -} - -extern void add_reference(reference** refmap, reference* ref) -{ - reference * t = NULL; - const char *label = (const char *)ref->label; - - HASH_FIND(hh, *refmap, label, strlen(label), t); - - if (t == NULL) { - HASH_ADD_KEYPTR(hh, *refmap, label, strlen(label), ref); - } else { - free_reference(ref); // we free this now since it won't be in the refmap - } -} - static unsigned char *bufdup(const unsigned char *buf) { unsigned char *new = NULL; @@ -236,26 +149,26 @@ inline static node_inl* append_inlines(node_inl* a, node_inl* b) return a; } -static void subject_from_buf(subject *e, strbuf *buffer, reference** refmap) +static void subject_from_buf(subject *e, strbuf *buffer, reference_map *refmap) { e->input.data = buffer->ptr; e->input.len = buffer->size; e->input.alloc = 0; e->pos = 0; e->label_nestlevel = 0; - e->reference_map = refmap; + e->refmap = refmap; chunk_rtrim(&e->input); } -static void subject_from_chunk(subject *e, chunk *chunk, reference** refmap) +static void subject_from_chunk(subject *e, chunk *chunk, reference_map *refmap) { e->input.data = chunk->data; e->input.len = chunk->len; e->input.alloc = 0; e->pos = 0; e->label_nestlevel = 0; - e->reference_map = refmap; + e->refmap = refmap; chunk_rtrim(&e->input); } @@ -325,33 +238,6 @@ static int scan_to_closing_backticks(subject* subj, int openticklength) return (subj->pos); } -// Destructively modify string, collapsing consecutive -// space and newline characters into a single space. -static void normalize_whitespace(strbuf *s) -{ - bool last_char_was_space = false; - int r, w; - - for (r = 0, w = 0; r < s->size; ++r) { - switch (s->ptr[r]) { - case ' ': - case '\n': - if (last_char_was_space) - break; - - s->ptr[w++] = ' '; - last_char_was_space = true; - break; - - default: - s->ptr[w++] = s->ptr[r]; - last_char_was_space = false; - } - } - - strbuf_truncate(s, w); -} - // Parse backtick code section or raw backticks, return an inline. // Assumes that the subject has a backtick at the current position. static node_inl* handle_backticks(subject *subj) @@ -368,7 +254,7 @@ static node_inl* handle_backticks(subject *subj) strbuf_set(&buf, subj->input.data + startpos, endpos - startpos - openticks.len); strbuf_trim(&buf); - normalize_whitespace(&buf); + strbuf_normalize_whitespace(&buf); return make_code(chunk_buf_detach(&buf)); } @@ -575,24 +461,9 @@ static node_inl *make_str_with_entities(chunk *content) } } -// Destructively unescape a string: remove backslashes before punctuation chars. -extern void unescape_buffer(strbuf *buf) -{ - int r, w; - - for (r = 0, w = 0; r < buf->size; ++r) { - if (buf->ptr[r] == '\\' && ispunct(buf->ptr[r + 1])) - continue; - - buf->ptr[w++] = buf->ptr[r]; - } - - strbuf_truncate(buf, w); -} - // Clean a URL: remove surrounding whitespace and surrounding <>, // and remove \ that escape punctuation. -static unsigned char *clean_url(chunk *url) +unsigned char *clean_url(chunk *url) { strbuf buf = GH_BUF_INIT; @@ -607,11 +478,11 @@ static unsigned char *clean_url(chunk *url) houdini_unescape_html_f(&buf, url->data, url->len); } - unescape_buffer(&buf); + strbuf_unescape(&buf); return strbuf_detach(&buf); } -static unsigned char *clean_autolink(chunk *url, int is_email) +unsigned char *clean_autolink(chunk *url, int is_email) { strbuf buf = GH_BUF_INIT; @@ -628,7 +499,7 @@ static unsigned char *clean_autolink(chunk *url, int is_email) } // Clean a title: remove surrounding quotes and remove \ that escape punctuation. -static unsigned char *clean_title(chunk *title) +unsigned char *clean_title(chunk *title) { strbuf buf = GH_BUF_INIT; unsigned char first, last; @@ -648,7 +519,7 @@ static unsigned char *clean_title(chunk *title) houdini_unescape_html_f(&buf, title->data, title->len); } - unescape_buffer(&buf); + strbuf_unescape(&buf); return strbuf_detach(&buf); } @@ -810,7 +681,7 @@ static node_inl* handle_left_bracket(subject* subj) } else { // if we get here, we matched a label but didn't get further: subj->pos = endlabel; - lab = parse_chunk_inlines(&rawlabel, subj->reference_map); + lab = parse_chunk_inlines(&rawlabel, subj->refmap); result = append_inlines(make_str(chunk_literal("[")), append_inlines(lab, make_str(chunk_literal("]")))); @@ -834,13 +705,13 @@ static node_inl* handle_left_bracket(subject* subj) } // lookup rawlabel in subject->reference_map: - ref = lookup_reference(subj->reference_map, &reflabel); + ref = reference_lookup(subj->refmap, &reflabel); if (ref != NULL) { // found lab = parse_chunk_inlines(&rawlabel, NULL); result = make_ref_link(lab, ref); } else { subj->pos = endlabel; - lab = parse_chunk_inlines(&rawlabel, subj->reference_map); + lab = parse_chunk_inlines(&rawlabel, subj->refmap); result = append_inlines(make_str(chunk_literal("[")), append_inlines(lab, make_str(chunk_literal("]")))); } @@ -887,7 +758,7 @@ extern node_inl* parse_inlines_while(subject* subj, int (*f)(subject*)) return result; } -node_inl *parse_chunk_inlines(chunk *chunk, reference** refmap) +node_inl *parse_chunk_inlines(chunk *chunk, reference_map *refmap) { subject subj; subject_from_chunk(&subj, chunk, refmap); @@ -987,7 +858,7 @@ static int parse_inline(subject* subj, node_inl ** last) return 1; } -extern node_inl* parse_inlines(strbuf *input, reference** refmap) +extern node_inl* parse_inlines(strbuf *input, reference_map *refmap) { subject subj; subject_from_buf(&subj, input, refmap); @@ -1009,7 +880,7 @@ void spnl(subject* subj) // Modify refmap if a reference is encountered. // Return 0 if no reference found, otherwise position of subject // after reference is parsed. -extern int parse_reference(strbuf *input, reference** refmap) +int parse_reference_inline(strbuf *input, reference_map *refmap) { subject subj; @@ -1019,7 +890,6 @@ extern int parse_reference(strbuf *input, reference** refmap) int matchlen = 0; int beforetitle; - reference *new = NULL; subject_from_buf(&subj, input, NULL); @@ -1065,9 +935,7 @@ extern int parse_reference(strbuf *input, reference** refmap) return 0; } // insert reference into refmap - new = make_reference(&lab, &url, &title); - add_reference(refmap, new); - + reference_create(refmap, &lab, &url, &title); return subj.pos; } diff --git a/src/references.c b/src/references.c new file mode 100644 index 0000000..ff64b00 --- /dev/null +++ b/src/references.c @@ -0,0 +1,109 @@ +#include "stmd.h" +#include "utf8.h" +#include "references.h" + +static unsigned int +refhash(const unsigned char *link_ref) +{ + unsigned int hash = 0; + + while (*link_ref) + hash = (*link_ref++) + (hash << 6) + (hash << 16) - hash; + + return hash; +} + +// normalize reference: collapse internal whitespace to single space, +// remove leading/trailing whitespace, case fold +static unsigned char *normalize_reference(chunk *ref) +{ + strbuf normalized = GH_BUF_INIT; + + utf8proc_case_fold(&normalized, ref->data, ref->len); + strbuf_trim(&normalized); + strbuf_normalize_whitespace(&normalized); + + return strbuf_detach(&normalized); +} + +static void add_reference(reference_map *map, reference* ref) +{ + ref->next = map->table[ref->hash % REFMAP_SIZE]; + map->table[ref->hash % REFMAP_SIZE] = ref; +} + +extern reference *reference_create(reference_map *map, chunk *label, chunk *url, chunk *title) +{ + reference *ref; + ref = malloc(sizeof(reference)); + ref->label = normalize_reference(label); + ref->hash = refhash(ref->label); + ref->url = clean_url(url); + ref->title = clean_title(title); + ref->next = NULL; + + add_reference(map, ref); + + return ref; +} + +// Returns reference if refmap contains a reference with matching +// label, otherwise NULL. +reference* reference_lookup(reference_map *map, chunk *label) +{ + reference *ref = NULL; + unsigned char *norm; + unsigned int hash; + + if (map == NULL) + return NULL; + + norm = normalize_reference(label); + hash = refhash(norm); + ref = map->table[hash % REFMAP_SIZE]; + + while (ref) { + if (ref->label[0] == norm[0] && + !strcmp((char *)ref->label, (char *)norm)) + break; + ref = ref->next; + } + + free(norm); + return ref; +} + +static void reference_free(reference *ref) +{ + free(ref->label); + free(ref->url); + free(ref->title); + free(ref); +} + +void reference_map_free(reference_map *map) +{ + unsigned int i; + + for (i = 0; i < REFMAP_SIZE; ++i) { + reference *ref = map->table[i]; + reference *next; + + while (ref) { + next = ref->next; + reference_free(ref); + ref = next; + } + } + + free(map->table); + free(map); +} + +reference_map *reference_map_new(void) +{ + reference_map *map = malloc(sizeof(reference_map)); + memset(map, 0x0, sizeof(reference_map)); + return map; +} + diff --git a/src/references.h b/src/references.h new file mode 100644 index 0000000..78fffe7 --- /dev/null +++ b/src/references.h @@ -0,0 +1,27 @@ +#ifndef _REFERENCES_H_ +#define _REFERENCES_H_ + +#define REFMAP_SIZE 16 + +struct reference { + struct reference *next; + unsigned char *label; + unsigned char *url; + unsigned char *title; + unsigned int hash; +}; + +typedef struct reference reference; + +struct reference_map { + reference *table[REFMAP_SIZE]; +}; + +typedef struct reference_map reference_map; + +reference_map *reference_map_new(void); +void reference_map_free(reference_map *map); +reference* reference_lookup(reference_map *map, chunk *label); +extern reference *reference_create(reference_map *map, chunk *label, chunk *url, chunk *title); + +#endif diff --git a/src/stmd.h b/src/stmd.h index 21a86b0..4e21e6c 100644 --- a/src/stmd.h +++ b/src/stmd.h @@ -5,7 +5,7 @@ #include #include "buffer.h" #include "chunk.h" -#include "uthash.h" +#include "references.h" #define VERSION "0.1" #define CODE_INDENT 4 @@ -36,17 +36,7 @@ struct node_inl { typedef struct node_inl node_inl; -struct reference { - unsigned char *label; - unsigned char *url; - unsigned char *title; - UT_hash_handle hh; // used by uthash -}; - -typedef struct reference reference; - // Types for blocks - struct ListData { enum { bullet, @@ -104,7 +94,7 @@ struct node_block { int level; } header; struct { - reference** refmap; + reference_map *refmap; } document; } as; @@ -114,14 +104,10 @@ struct node_block { typedef struct node_block node_block; -node_inl* parse_inlines(strbuf *input, reference** refmap); +node_inl* parse_inlines(strbuf *input, reference_map *refmap); void free_inlines(node_inl* e); -int parse_reference(strbuf *input, reference** refmap); -void free_reference(reference *ref); -void free_reference_map(reference **refmap); - -void add_reference(reference** refmap, reference* ref); +int parse_reference_inline(strbuf *input, reference_map *refmap); void unescape_buffer(strbuf *buf); extern node_block* make_document(); @@ -138,4 +124,8 @@ void print_blocks(node_block* blk, int indent); void blocks_to_html(strbuf *html, node_block *b, bool tight); void inlines_to_html(strbuf *html, node_inl *b); +unsigned char *clean_url(chunk *url); +unsigned char *clean_autolink(chunk *url, int is_email); +unsigned char *clean_title(chunk *title); + #endif diff --git a/src/utf8.c b/src/utf8.c index 12d7ba5..c65aec6 100644 --- a/src/utf8.c +++ b/src/utf8.c @@ -25,7 +25,7 @@ static const int8_t utf8proc_utf8class[256] = { static void encode_unknown(strbuf *buf) { - static const unsigned char repl[] = {239, 191, 189}; + static const uint8_t repl[] = {239, 191, 189}; strbuf_put(buf, repl, 3); } @@ -52,9 +52,9 @@ ssize_t utf8proc_charlen(const uint8_t *str, ssize_t str_len) return length; } -void utf8proc_detab(strbuf *ob, const unsigned char *line, size_t size) +void utf8proc_detab(strbuf *ob, const uint8_t *line, size_t size) { - static const unsigned char whitespace[] = " "; + static const uint8_t whitespace[] = " "; size_t i = 0, tab = 0; @@ -132,7 +132,7 @@ ssize_t utf8proc_iterate(const uint8_t *str, ssize_t str_len, int32_t *dst) void utf8proc_encode_char(int32_t uc, strbuf *buf) { - unsigned char dst[4]; + uint8_t dst[4]; int len = 0; assert(uc >= 0); @@ -169,7 +169,7 @@ void utf8proc_encode_char(int32_t uc, strbuf *buf) strbuf_put(buf, dst, len); } -void utf8proc_case_fold(strbuf *dest, const unsigned char *str, int len) +void utf8proc_case_fold(strbuf *dest, const uint8_t *str, int len) { int32_t c; diff --git a/src/utf8.h b/src/utf8.h index 1e4e556..9506b75 100644 --- a/src/utf8.h +++ b/src/utf8.h @@ -1,12 +1,13 @@ #ifndef _H_STMD_UTF8_ #define _H_STMD_UTF8_ +#include #include "buffer.h" -void utf8proc_case_fold(strbuf *dest, const unsigned char *str, int len); +void utf8proc_case_fold(strbuf *dest, const uint8_t *str, int len); void utf8proc_encode_char(int32_t uc, strbuf *buf); ssize_t utf8proc_iterate(const uint8_t *str, ssize_t str_len, int32_t *dst); ssize_t utf8proc_charlen(const uint8_t *str, ssize_t str_len); -void utf8proc_detab(strbuf *dest, const unsigned char *line, size_t size); +void utf8proc_detab(strbuf *dest, const uint8_t *line, size_t size); #endif -- cgit v1.2.3