https://bugs.kde.org/show_bug.cgi?id=495261

--- Comment #22 from Matthew Woehlke <[email protected]> ---
> Finding where exactly on disk the cache has been corrupted ...

What does that have to do with *heap* corruption? And even if the problem *is*
related to a corrupt cache (which it isn't; it's reproducible on a system
running KDevelop for the first time), isn't knowing how file corruption turns
into memory corruption still useful?

Anyway, this is academic, because (not entirely surprisingly) the valgrind run
didn't crash. Which almost certainly means the issue *is* heap corruption,
because valgrind is known to change memory access patterns (and is also
somewhat forgiving of heap corruption). Unfortunately, I didn't think to save
the logs or turn off qDebug, which means any errors got lost in the incredible
spam of other output. I can do this again (and, ahem, maybe you can remind me
how to purge the cache first?) with qDebug disabled, but it takes many, many
hours to run, so I can't do it during the day. (Well, unless I break out the
live image again, which I might do; I have a separate computer for that so it
doesn't tie down the machine I need for work. But it's more effort to set up.)

-- 
You are receiving this mail because:
You are watching all bug changes.

Reply via email to