Bug 163591 – [GTK] Cannot build GTK+ with CUPS 2.* versions

• Status: RESOLVED
• Resolution: FIXED
• Component: WebKitGTK

[Manuel Rego Casasnovas]

[GTK] Cannot build GTK+ with CUPS 2.* versions

[Manuel Rego Casasnovas]

GTK+ is wrongly checking about the avaible CUPS version, so if you have CUPS 2.* the configure step fails.

The issue has already been fixed in GTK+ upstream:
https://git.gnome.org/browse/gtk+/commit/id=631f6b536485829a0bd00532f5826ad302b4951b
But the change is not part of the 3.16.* versions.

[Manuel Rego Casasnovas]

Created attachment 291930 [details]
Patch

[Carlos Garcia Campos]

Comment on attachment 291930 [details]
Patch

Have you checked that printing tests still work with cups disabled? I would add a patch for gtk+ with the upstream fix, at least until we bump the GTK+ version in jhbuild.

[Manuel Rego Casasnovas]

Ok, I can add the patch if you prefer it. Thanks for the quick reviews!

(In reply to comment #1)
> The issue has already been fixed in GTK+ upstream:
> https://git.gnome.org/browse/gtk+/commit/
> id=631f6b536485829a0bd00532f5826ad302b4951b
Sorry this link was wrong, the right one is:
https://git.gnome.org/browse/gtk+/commit/?id=631f6b536485829a0bd00532f5826ad302b4951b

[Michael Catanzaro]

Comment on attachment 291930 [details]
Patch

I agree with Carlos, this doesn't look like a good idea.

[Manuel Rego Casasnovas]

Created attachment 292102 [details]
Patch

This is the version using the upstream patch to fix the build with CUPS. However it doesn't work if glib is not in version 2.50. So I've provided a different patch in bug #163689 to bump the glib version that should be landed before this one.

[WebKit Commit Bot]

Comment on attachment 292102 [details]
Patch

Clearing flags on attachment: 292102

Committed r207609: <http://trac.webkit.org/changeset/207609>

[WebKit Commit Bot]

All reviewed patches have been landed.  Closing bug.

[WebKit Commit Bot]

Re-opened since this is blocked by bug 163776

[Manuel Rego Casasnovas]

(In reply to comment #3)
> Have you checked that printing tests still work with cups disabled?

FWIW, I've verified that TestWebKitAPI/WebKit2Gtk/TestPrinting fails (actually timeouts) if you pass --disable-cups to GTK+ build.

[Xabier Rodríguez Calvar]

Created attachment 293007 [details]
Patch to workaround the build

[Joanmarie Diggs]

Comment on attachment 292102 [details]
Patch

This patch should once again be good.

[ChangSeok Oh]

What blocks this to land? If bumping glib and its consequence making bots red is the blocker, how about applying the change in the configure.ac to the configure directly? We don't need to bump up the glib so we don't need to worry  about bots red.

[Michael Catanzaro]

Hi!

(In reply to comment #13)
> What blocks this to land?

First patch depends on bug #163689, which looks like it would probably be pretty easy to fix (see my last comment in that bug), just nobody has tried to.

Second patch uses --disable-cups, which is not OK.

[ChangSeok Oh]

Created attachment 296395 [details]
Patch

[Michael Catanzaro]

Comment on attachment 296395 [details]
Patch

OK

[ChangSeok Oh]

Comment on attachment 296395 [details]
Patch

I don't know why the gtk-wk2 bot hangs on the oragne. Anyhow, this should not hurt anything.

[WebKit Commit Bot]

Comment on attachment 296395 [details]
Patch

Clearing flags on attachment: 296395

Committed r209482: <http://trac.webkit.org/changeset/209482>

[WebKit Commit Bot]

All reviewed patches have been landed.  Closing bug.