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

--- Comment #26 from Igor Kushnir <[email protected]> ---
(In reply to Matthew Woehlke from comment #25)
> > The cache gets corrupted - this is the root cause and should certainly be 
> > fixed.
> 
> Okay, while I grant that you probably know more about KDevelop's innards
> than I do, I remain skeptical of this claim. Again, I note that the crash
> occurs on the first execution within a pristine environment. If the *root*
> problem is cache corruption, that implies that something produces a bad
> result, writes it to disk, *reads it back* in the same instance, and causes
> heap corruption as a result, yes?
Not sure if reading the cache back ever happens, but it might. Another
possibility is that the first crash corrupts the cache along the way - if it
happens at the moment when the cache stored on disk is temporarily
inconsistent. Whether or not the cache gets corrupted, I am not certain. But as
a cache corruption is *possible*, the cache should ideally be cleaned before
performing time-consuming tests. Otherwise, we may not be sure whether a true
bug or an unfixable crash due to a cache corruption has been detected.

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

Reply via email to