Re: [Pan-users] 0.95 woes

[Charles Kerr]

> On Tue, 02 May, 2006 at 22:01 +0200, Sami Haahtinen wrote:
>> First one is a bug: if you use the quicksearch to find an article
>> somewhere in the middle of a thread, then clear the search and press u
>> you will get a pretty constant segfault. My guess is that this is
>> related to the recent gtk fixes.
>
> Seems pretty reproducible. Backtrace attached for Charles.

Yes, this one's pretty easy, unlike the one Dave Chand's been driving
me nuts with. ;)  Just a one-liner in pan-tree.cc::iter_children:

@@ -141,7 +176,7 @@
     PanTreeStore * tree = PAN_TREE_STORE(model);
     g_return_val_if_fail (tree, false);
     g_return_val_if_fail (iter, false);
-    g_return_val_if_fail (parent->stamp == tree->stamp, false);
+    g_return_val_if_fail (!parent || (parent->stamp==tree->stamp), false);

Thanks for reporting this.

>> Second one is more of a feature request than anything else. It appears
>> that pan doesn't do too well with articles that don't have charset
>> marked in the headers.
>
> Charles and I were actually discussing that off-list yesterday. This
> definitely is a regression vs. 0.14.x, so I guess Charles will settle on
> a solution at some point.

Not that I wouldn't turn down contributed code, of course.