[Xcb] [PATCH libX11] xcb_io: Fix Xlib 32-bit request number issues

[Christian Linhart]

Hi all,

I have started to work on this but it is somewhat more complex than anticipated.
I'll continue to work on this the next weekends, and I'll hopefully get a useful result then.
As soon as I'll have something that can be basis for a discussion ( even if not 100% ), I'll post it here.

Regards,

Chris

On 02/26/15 08:14, chris at demorecorder.com wrote:
> Hi Olivier and all,
>
> Thank you for the reminder about this problem. 
>
> On this weekend and/or on the following weekend I will write a different patch,
> as I have outlined some time ago. My patch will use 64-bit request numbers 
> on the communication between XCB and Xlib.
>
> I have finally time to do this.
>
> I hope that my patch will be easier to review 
> and will fix the problem in a simpler way than Jonas' patch.
>
> Then we'll hopefilly be able to get the problem fixed in a way
> that can get positive reviews.
>
> ***
>
> In any case I think that such central changes need to be reviewed
> and ACK-ed by core X11 developers.
> Evidence from testing alone is not sufficient
> because we all know that testing cannot prove correctness.
>
> X11 is used in critical applications like air traffic control,
> so we need to be very careful here.
>
> ***
>
> If you have additional testcases for the problem that Jonas's patch
> fixes, then please post them here.
>
> Regards,
>
> Chris
>
> Olivier Fourdan wrote on 25.02.2015 19:04:
>
>> Hi all,
>>
>> Sorry I'd wish I could reply the original thread but I wasn't subscribed at the
>> time... This is about https://freedesktop.org/patch/34068/
>>
>> Some more people have been hitting this issue and I understand the review of
>> the patch is complicated, but at the same time Jonas' patch have been reported
>> to fix the issue by quite a few people apparently.
>>
>> This patch hasn't been merged yet, is this because it is not correct or is it
>> just that it might be incomplete or not optimal?
>>
>> If the latter, could we consider merging Jonas' patch as a first step, possibly
>> leaving the bug opened until a more complete or better solution is found?
>> Taking an incremental approach to the problem might be better than status quo.
>>
>> Could we move forward with this, even if the patch is not perfect (assuming
>> it's not broken or causing havoc, of course)?
>>
>> Thanks,
>> Olivier
>>
>> _______________________________________________
>> Xcb mailing list
>> Xcb at lists.freedesktop.org
>> http://lists.freedesktop.org/mailman/listinfo/xcb
> _______________________________________________
> Xcb mailing list
> Xcb at lists.freedesktop.org
> http://lists.freedesktop.org/mailman/listinfo/xcb