1 files changed, 18 insertions, 0 deletions
diff --git a/doc/bugs/Not_all_comments_are_listed_by___33__map_or___33__inline.mdwn b/doc/bugs/Not_all_comments_are_listed_by___33__map_or___33__inline.mdwn
index bfdc8d046..1032d7e1b 100644
--- a/doc/bugs/Not_all_comments_are_listed_by___33__map_or___33__inline.mdwn
+++ b/doc/bugs/Not_all_comments_are_listed_by___33__map_or___33__inline.mdwn
@@ -19,6 +19,18 @@ While working on our forum I was surprised to see that some of the comments were
 >>> So I suppose I'd need a testcase in a tarball to reproduce 
 >>> any problem. --[[Joey]]
 
+>>>> Once I set `rcs: git` and `gitorigin_branch: ''`, I also get a
+>>>> perfect RSS feed that contains the items git log makes me expect,
+>>>> in the correct order. So this is not a ikiwiki bug after all,
+>>>> sorry for the annoyance. (For the record, I think we have two
+>>>> problems: first, our ikiwiki.setup does not enable a RCS, mainly
+>>>> to avoid local refresh to create ugly "updated PO files" Git
+>>>> commits; this explains the issue sajolida noticed while locally
+>>>> building the wiki. Second, as far as our online ikiwiki is
+>>>> concerned, the .ikiwiki may be broken, may lack some times
+>>>> extracted from Git, who knows => we'll see if a rebuild cures
+>>>> it.) --[[intrigeri]]
+
 Then I found out that a map directive such as `!map pages="forum/* or
 comment(forum/*)"` was bringing a weird result too. The output is a map
 with quite a few broken links.
@@ -33,6 +45,12 @@ with quite a few broken links.
 >>> I checked the first 50% or so of the comments, and every one was a
 >>> broken link. --[[Joey]]
 
+>>>> I now observe the same behaviour as Joey, which seems totally
+>>>> logical to me after all given that `forum/*/comment_*.html` are
+>>>> not generated. I wonder how we could have observed map generating
+>>>> working links to comments in the first place; sajolida, can you
+>>>> please try reproducing it? --[[intrigeri]]
+
 Plus, some broken links in the map do match the comments missing on the RSS feed but some others do not.
 
 Unfortunately, I couldn't find an obvious pattern for this failure.