[PATCH] [RFC] socket activation code for xserver

[Łukasz Stelmach]

It was <2013-11-28 czw 15:59>, when Hans de Goede wrote:
> Hi,
>
> On 11/28/2013 09:54 AM, Łukasz Stelmach wrote:
>> It was <2013-11-27 śro 11:37>, when Hans de Goede wrote:
>>> Hi Łukasz,
>>>
>>> So I was taking a closer look at the libxtrans patch today,
>>> and I noticed that it does not build. It adds a check for
>>> trans->flags&TRANS_RECEIVED inside the main loop in
>>> MakeAllCOTSServerListeners. But TRANS_RECEIVED does not get
>>> defined anywhere.
>>
>> It definitely is
>>
>> http://article.gmane.org/gmane.comp.freedesktop.xorg.devel/37694
>> http://thread.gmane.org/gmane.comp.freedesktop.xorg.devel/36092/focus=37694
>>
>>>
>>> I've added a define for it to Xtransint.h, which fixes the
>>> build issue. But other then fixing the BuildIssue this
>>> makes little sense, since the flag is never being set.
>>
>> It is, in the patch above.
>>>
>>> So an alternative approach to fixing the build issue, would
>>> to just remove the addition of the check from your patch.
>>>
>>> Which has left me wondering why did you add this check in
>>> the first place? What was it supposed to do, and if we
>>> drop the check, do we need some other code to achieve
>>> the same result?
>>>
>>> While looking into this, I've polished the patch a but
>>> up, addressing some of the review comments from its
>>> earlier posting, and improving error reporting. I've
>>> attached my cleaned-up version.
>>
>> Could you please make sure you are working with the v4 version of the
>> patches as available here:
>>
>> http://thread.gmane.org/gmane.comp.freedesktop.xorg.devel/36092/focus=37693
>
> Ah, Peter had given me a link to the older v3.2 patches. Using v4 helps.
>
> I've applied v4 to my local tree, then added some cleanups addressing the
> review comments plus better error-reporting I already did to v3 to:
> the "Enable systemd socket activation" libxtrans patch. And with that
> I've run some simple tests successfully.
>
> I'm going to re-post my version of the patches (only the "Enable systemd
> socket activation" is changed) to the list, and if there are no objections
> I'll push it to the official xorg repos.
>
> Note I'm not including: "Do not define ErrorF in xstrans.c, libxtrans handles it."
> as that no longer applies (trivial to fix, but I wonder if it is still valid),
> and it seems to be an unrelated cleanup patch.

Yes, the only relationship is that I've spotted the problem while
developing the rest of the code. Indeed the code I removed with the
patch is now surrounded by an ifdef that probably fixes the problem.

-- 
Łukasz Stelmach
Samsung R&D Institute Poland
Samsung Electronics
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 489 bytes
Desc: not available
URL: <http://lists.x.org/archives/xorg-devel/attachments/20131129/297eacd3/attachment-0001.pgp>