summaryrefslogtreecommitdiff
path: root/doc
diff options
context:
space:
mode:
authorjoey <joey@0fa5a96a-9a0e-0410-b3b2-a0fd24251071>2007-02-15 09:23:49 +0000
committerjoey <joey@0fa5a96a-9a0e-0410-b3b2-a0fd24251071>2007-02-15 09:23:49 +0000
commitf1f2d92ae19c5928215160d0ee231ba065626a3e (patch)
treebc76b5bd90a5f859a7f36b836262939122bd043a /doc
parent82178d8e8f3a9cda17ab13a82f5b66b03d58339f (diff)
followup
Diffstat (limited to 'doc')
-rw-r--r--doc/bugs/CGI_showed_HTML_when_perl_error.mdwn22
1 files changed, 21 insertions, 1 deletions
diff --git a/doc/bugs/CGI_showed_HTML_when_perl_error.mdwn b/doc/bugs/CGI_showed_HTML_when_perl_error.mdwn
index 1cb9b94bc..b222b297d 100644
--- a/doc/bugs/CGI_showed_HTML_when_perl_error.mdwn
+++ b/doc/bugs/CGI_showed_HTML_when_perl_error.mdwn
@@ -1,4 +1,8 @@
-I didn't have Time/Duration.pm installed when I clicked RecentChanges. The perl failed. The CGI outputed the Content-type: text/html and the complete HTML which included the error in side of the paragraph tags. Maybe a newline was sent before that Content-type line. The web browser didn't render the HTML but just showed the source.
+I didn't have Time/Duration.pm installed when I clicked RecentChanges. The
+perl failed. The CGI outputed the Content-type: text/html and the complete
+HTML which included the error in side of the paragraph tags. Maybe a newline
+was sent before that Content-type line. The web browser didn't render the HTML
+but just showed the source.
> I can't reproduce this, I get a properly formatted error page.
> If you'd like to send me the page, I can try to figure out what
@@ -18,3 +22,19 @@ I didn't have Time/Duration.pm installed when I clicked RecentChanges. The perl
>>
>> --JeremyReed
+Here's what I see, taking the web server out of the picture:
+
+ joey@kodama:~>~/html/ikiwiki.cgi 2>/dev/null |hexdump -C|head -1
+ 00000000 43 6f 6e 74 65 6e 74 2d 74 79 70 65 3a 20 74 65 |Content-type: te|
+
+No spurious 0a. With apache:
+
+ 0100 75 6e 6b 65 64 0d 0a 43 6f 6e 74 65 6e 74 2d 54 unked..C ontent-T
+
+Here the 0d 0a is a CRLF, and note that it's output by the web server, not
+ikiwiki. It's perfectly valid, while a lone 0a, just a linefeed, is not valid
+HTTP. Conclusion, this was your web server; it's not uncommon for hacky
+little web servers to not use proper CRLF's, and it works _some_ of the time,
+depending on how strict the browser is.
+
+I'm calling this [[bugs/done]] --[[Joey]]