[Bug 53112] New: winegstreamer build error

[WineHQ Bugzilla]

https://bugs.winehq.org/show_bug.cgi?id=53112

Bug ID: 53112
           Summary: winegstreamer build error
           Product: Wine
           Version: 7.10
          Hardware: x86-64
                OS: Linux
            Status: UNCONFIRMED
          Severity: normal
          Priority: P2
         Component: winegstreamer
          Assignee: [e-mail]
          Reporter: [e-mail]
      Distribution: ---

/root/code/dlls/winegstreamer/wg_transform.c:84:12: error: 
void value not ignored as it ought to be
     return gst_video_info_align(info, align);

return type is void, not bool.

The function is declared as follows in gstreamer-1.0：
/usr/include/gstreamer-1.0/gst/video/video-info.h:
void gst_video_info_align (GstVideoInfo * info, GstVideoAlignment * align);

[Bug 53112] winegstreamer build error

[WineHQ Bugzilla]

https://bugs.winehq.org/show_bug.cgi?id=53112

--- Comment #1 from Rafał Mużyło  ---
Your gstreamer version is *ridiculously* out of date.

The exact change was made during fixing
https://bugzilla.gnome.org/show_bug.cgi?id=774588 and made it to gstreamer 1.12
over 5 years ago.

[Bug 53112] winegstreamer build error

[WineHQ Bugzilla]

https://bugs.winehq.org/show_bug.cgi?id=53112

Zeb Figura  changed:

What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |[e-mail]

--- Comment #2 from Zeb Figura  ---
(In reply to Rafał Mużyło from comment #1)
> Your gstreamer version is *ridiculously* out of date.
>
> The exact change was made during fixing
> https://bugzilla.gnome.org/show_bug.cgi?id=774588 and made it to gstreamer
> 1.12 over 5 years ago.

There's no need for statements like that. Many stability-focused distributions
ship similarly old library versions, and we have no difficulty supporting that
in this case.

[Bug 53112] winegstreamer build error

[WineHQ Bugzilla]

https://bugs.winehq.org/show_bug.cgi?id=53112

--- Comment #3 from Rafał Mużyło  ---
(In reply to Zeb Figura from comment #2)
> There's no need for statements like that. Many stability-focused
> distributions ship similarly old library versions, and we have no difficulty
> supporting that in this case.

I'd say YMMV.

That is on one hand, I see this argument a bit off target - given how many
changes during that time wine has seen...

Also, it's just reasonable to update the deps while you're updating the leafs,
as long as as they're ABI compatible.

Also also, in this particular context that function was a bit dangerous before
the change - in wine context (no control over data fed into the app) that would
be inviting trouble.

[Bug 53112] winegstreamer build error

[WineHQ Bugzilla]

https://bugs.winehq.org/show_bug.cgi?id=53112

--- Comment #4 from Zeb Figura  ---
(In reply to Rafał Mużyło from comment #3)
> (In reply to Zeb Figura from comment #2)
> > There's no need for statements like that. Many stability-focused
> > distributions ship similarly old library versions, and we have no difficulty
> > supporting that in this case.
>
> I'd say YMMV.
>
> That is on one hand, I see this argument a bit off target - given how many
> changes during that time wine has seen...
>
> Also, it's just reasonable to update the deps while you're updating the
> leafs, as long as as they're ABI compatible.

No part of this makes any sense. We don't have any pressing need to require
newer GStreamer, therefore we can and should accommodate older GStreamer.
> Also also, in this particular context that function was a bit dangerous
> before the change - in wine context (no control over data fed into the app)
> that would be inviting trouble.

Please stop making statements like this without a complete understanding of the
relevant code, it is far more harmful than helpful.

[Bug 53112] winegstreamer build error

[WineHQ Bugzilla]

https://bugs.winehq.org/show_bug.cgi?id=53112

--- Comment #5 from Rafał Mużyło  ---
,,,

Let me clarify: "that *gstreamer* function was a bit dangerous".

After all, the referenced bug was about unsigned integer overflow - while the
values required are quite high, gstreamer team still consider the risk of
hitting it high enough to make the change.

Well, 'dangerous' might be a bit strong word, as most likely that would only
cause visual glitches...but the key part of that statement is "most likely".

And yes, I see this is more of a wine policy decision than a technical one (at
least the way I define those terms). I just consider such policy wrong...but as
I have no say in it :shrug:.

[Bug 53112] winegstreamer build error

[WineHQ Bugzilla]

https://bugs.winehq.org/show_bug.cgi?id=53112

Robert Wilhelm  changed:

What    |Removed                     |Added
----------------------------------------------------------------------------
      Fixed by SHA1|                            |9f3eb4b7f04c0233d1e62d55c29
                   |                            |9362f2de4ca52
         Resolution|---                         |FIXED
                 CC|                            |[e-mail]
             Status|UNCONFIRMED                 |RESOLVED

--- Comment #6 from Robert Wilhelm  ---
Thanks to Rémi,
compilation with some older GStreamer versions which return void from
gst_video_info_align is fixed now in current git.
See 9f3eb4b7f04c0233d1e62d55c299362f2de4ca52.

[Bug 53112] winegstreamer build error

[WineHQ Bugzilla]

https://bugs.winehq.org/show_bug.cgi?id=53112

Alexandre Julliard  changed:

What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|RESOLVED                    |CLOSED

--- Comment #7 from Alexandre Julliard  ---
Closing bugs fixed in 7.11.