X-Git-Url: http://git.vanrenterghem.biz/git.ikiwiki.info.git/blobdiff_plain/55dda5d59166c935a96bcd56515e3a158838c9bc..34115a34e0593e999b7a279e07293b090012082a:/doc/index/discussion.mdwn

diff --git a/doc/index/discussion.mdwn b/doc/index/discussion.mdwn
index ef7337c89..4cf38e9cb 100644
--- a/doc/index/discussion.mdwn
+++ b/doc/index/discussion.mdwn
@@ -405,6 +405,9 @@ I'm playing around with various ways that I can use subversion with ikiwiki.
 > away without running the post-commit wrapper on commit, and all you lose
 > is the ability to send commit notification emails.
 
+> (And now that [[recentchanges]] includes rss, you can just subscribe to
+> that, no need to worry about commit notification emails anymore.)
+
 * Is it possible / sensible to have ikiwiki share a subversion repository with other data (either completely unrelated files or another ikiwiki instance)?  This works in part but again the post-commit hook seems problematic.
 
 --[[AdamShand]]
@@ -413,3 +416,25 @@ I'm playing around with various ways that I can use subversion with ikiwiki.
 > in the same repo. If you have two wikis in one repository, you will need
 > to write a post-commit script that calls the post-commit wrappers for each
 > wiki.
+
+----
+
+# Regex for Valid Characters in Filenames
+
+I'm sure that this is documented somewhere but I've ransacked the wiki and I can't find it.  :-(  What are the allowed characters in an ikiwiki page name?  I'm writing a simple script to make updating my blog easier and need to filter invalid characters (so far I've found that # and , aren't allowed ;-)).  Thanks for any pointers.  -- [[AdamShand]]
+
+> The default `wiki_file_regexp` matches filenames containing only
+> `[-[:alnum:]_.:/+]`
+> 
+> The IkiWiki::titlepage() function will convert freeform text to a valid
+> page name. See [[todo/should_use_a_standard_encoding_for_utf_chars_in_filenames]]
+> for an example. --[[Joey]]
+
+>> Perfect, thanks! 
+>>
+>> In the end I decided that I didn't need any special characters in filenames and replaced everything but alphanumeric characters with underscores.  In addition to replacing bad characters I also collapse multiple underscores into a single one, and strip off trailing and leading underscores to make tidy filenames.  If it's useful to anybody else here's a sed example:
+>>
+>>     # echo "++ Bad: ~@#$%^&*()_=}{[];,? Iki: +_-:./ Num: 65.5 ++" | sed -e 's/[^A-Za-z0-9_]/_/g' -e 's/__*/_/g' -e 's/^_//g' -e 's/_$//g'
+>>     Bad_Iki_Num_65_5
+>>
+>>--[[AdamShand]]