xlibs

Bug #507062
Comment #109

Comment 109 for bug 507062

Revision history for this message

In freedesktop.org Bugzilla #27552, Arthur Thomson (speystar) wrote on 2012-02-23: Re: [Bug 507062] Re: synaptic assert failure: synaptic: ../../src/xcb_io.c:385: _XAllocID: Assertion `ret != inval_id' failed.

#109

Thanks Adam...

On a slightly different but still a little connected topic... just how
resistent to viruses/virii is LINUX namely UBUNTU 11.04???

I have reverted to this version from 11.10 as I prefer all the
widgets... Am using the KDE desktop although it does run a little slower
than the GDE DT...

Are there are hints etc that you may be able to offer about keeping all
the dangerous things out???

regards

arthur

On 02/23/2012 05:16 PM, Adam Thompson wrote:
> Sad to say, folks, that I've been seeing this bug VERY intermittently on multiple generations of hardware, although I think I can safely make these assertions:
> -only since GTK became widespread (this definitely did not occur back in the Athena, Motif, OpenWindows, etc. days)
> **specifically, I don't think I've ever seen this bug appear when I wasn't running GTK+-linked or GTK2-linked software**
> -only since GL extensions became widespread (including software, e.g. MESA)
> -mostly but not exclusively on Linux systems (at least once under Solaris shortly after GNOME became the default desktop)
> -some (hardware) systems tended to produce this error more than others
> -certain applications tend to cause it more often than others (notably Synaptic, per the Ubuntu reports)
> -I *think* higher compiler optimization causes it to happen more often, not sure
> -not exclusively on multiprocessor systems: I've definitely seen this error on a Pentium III 1.0GHz (ULV mobile) and a Sparc IIi (~400MHz?).
>
> It's almost certainly a race condition, since nothing else would produce this sort of intermittent behaviour across 10+ years and multiple hardware platforms.
> It may not be a regression - it's likely we're dealing with a design flaw that has existed since<some piece of code> was written, but is being triggered more and more often because of the proliferation of multiprocessor platforms with multithreaded libs and multithread-generating-compilers.
>
> Unfortunately, this all means that finding it will be nearly impossible.
> So much for the "all bugs are shallow" theory :-).
>
> Or it could just be something low-level in GTK, which would explain why
> it's been cropping up more and more.
>

Thanks Adam...

On a slightly different but still a little connected topic... just how 
resistent to viruses/virii is LINUX namely UBUNTU 11.04???

I have reverted to this version from 11.10 as I prefer all the 
widgets... Am using the KDE desktop although it does run a little slower 
than the GDE DT...

Are there are hints etc that you may be able to offer about keeping all 
the dangerous things out???

regards

arthur

On 02/23/2012 05:16 PM, Adam Thompson wrote:
> Sad to say, folks, that I've been seeing this bug VERY intermittently on multiple generations of hardware, although I think I can safely make these assertions:
> -only since GTK became widespread (this definitely did not occur back in the Athena, Motif, OpenWindows, etc. days)
>   **specifically, I don't think I've ever seen this bug appear when I wasn't running GTK+-linked or GTK2-linked software**
> -only since GL extensions became widespread (including software, e.g. MESA)
> -mostly but not exclusively on Linux systems (at least once under Solaris shortly after GNOME became the default desktop)
> -some (hardware) systems tended to produce this error more than others
> -certain applications tend to cause it more often than others (notably Synaptic, per the Ubuntu reports)
> -I *think* higher compiler optimization causes it to happen more often, not sure
> -not exclusively on multiprocessor systems: I've definitely seen this error on a Pentium III 1.0GHz (ULV mobile) and a Sparc IIi (~400MHz?).
>
> It's almost certainly a race condition, since nothing else would produce this sort of intermittent behaviour across 10+ years and multiple hardware platforms.
> It may not be a regression - it's likely we're dealing with a design flaw that has existed since<some piece of code>  was written, but is being triggered more and more often because of the proliferation of multiprocessor platforms with multithreaded libs  and multithread-generating-compilers.
>
> Unfortunately, this all means that finding it will be nearly impossible.
> So much for the "all bugs are shallow" theory :-).
>
> Or it could just be something low-level in GTK, which would explain why
> it's been cropping up more and more.
>