| Age | Commit message (Collapse) | Author |
|---|
|
|
|
I understand the need to avoid chdir when running git_parse_changes
for receive now. At that point, the changes have not been pushed to
the srcdir's repo yet. When running the same code for preprevert,
chdir to the srcdir is ok, and necessary.
|
|
|
|
|
|
I removed the IkiWiki::rcs_ stubs for the revert hooks. Instead
recentchanges tests to see if the hooks are available and calls
them directly.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Need checkconfig hook; examples don't become default values.
|
|
|
|
|
|
|
|
|
|
|
|
plovs reported a crash when templates were not installed properly,
with a non-useful error about the template object not being defined.
I've audited all uses of template_depends(), and template(), and it makes
sense for them to throw an error if the template cannot be found. All code
with a user-supplied template catches errors already, to handle template
parse failures.
It did not make sense for template_file to throw errors, as some code uses
it to probe if a template file is available.
|
|
That template is user-controlled.
|
|
|
|
In the last version, the ikiwiki script stopped setting PATH.
But that leads to gcc failing when run from websetup. See
http://www.branchable.com/bugs/Crashes_when_rebuilding_wiki_after_setup_change/
|
|
The HTML::Tree changelog says:
[THINGS THAT MAY BREAK YOUR CODE OR TESTS]
...
* Attribute names are now validated in as_XML and invalid names will
cause an error.
and indeed the regression tests do get an error.
|
|
|
|
With a relative xrds-location, the openid perl client module will fail.
I haven't checked the specs to see if it needs to be absolute, but all
examples I've seen are absolute, so it seems a very good idea.
|
|
|
|
pasted on a page before being cut.
|
|
errors from conflicting obsolete remote branches.
|
|
that contained only a number, fixing a longstanding crash of the rst plugin.
|
|
I also tried setting RPC::XML::ENCODING but that did not prevent the crash,
and it seems that blogspam.net doesn't like getting xml encoded in unicode,
since it mis-flagged comments as spammy that way that are normally allowed
through.
|
|
|
|
second parameter, to allow for plugins that needs access to this information earlier than the delete hook.
|
|
|
|
This could happen if checkconfig was run twice, I think.
|
|
|
|
be configured via the web.
|
|
The only unsafe thing should be that enabling it with some languages will
generate po files.
|
|
If I am not mistaking all source files in ikiwiki are encoded in Unicode UTF-8.
Adding `\usepackage[utf8]{inputenc}` enables LaTeX to deal with the encoding.
As a consequence some special characters like umlauts can be used in the source
code which is useful for foreign languages.
[[!teximg code="a = b \text{ für alle } b \neq 2"]]
But for example »≠« cannot be used in LaTeX right now. One has to use other TeX
systems like XeTeX or LuaTeX featuring native UTF-8 support or use additional
nonstandard packages like uniinput [1].
I used the package `inputenc` (`texdoc inputenc`) and not `inputenx` (`texdoc
inputenx`), because I have not used `inputenx` that much and using the option
`math` is not supported in Debian (and I guess other distributions too) since
`inpmath` is not included in CTAN.
[1] http://wiki.neo-layout.org/browser/latex/Standard-LaTeX
Signed-off-by: Paul Menzel <paulepanter@users.sourceforge.net>
|
|
array of things that need built. (Backwards compatability code keeps plugins using the old interface working.)
|
|
These return codes are not currently used, but might be later.
|
|
a login session.
|
|
Abstraction violation. I now think the problem should be treated as a bug
in httpauth.
|
|
|
|
Avoid the generic "you are not allowed to change" message,
and instead allow check_canedit to propigate out useful error messages.
Went back to calling check_canedit in fatal mode, but added a parameter to
avoid calling the troublesome subs that might cause a login attempt.
|
|
|
