Re: webkit-gtk fails to build

[Chavdar Ivanov <ci4ic4%gmail.com@localhost>]

This was the problem indeed. I found a random package having
'introspection' as an option and tried to build it without DISPLAY set
(in this case, gtksourceview4). It failed in exactly the same way:
....
  GISCAN   GtkSource-4.gir

** (process:9416): CRITICAL **: 10:52:26.148: Failed to get connection
to xfconfd: Cannot autolaunch D-Bus without X11 $DISPLAY

** (process:16605): CRITICAL **: 10:52:31.273: Failed to get
connection to xfconfd: Cannot autolaunch D-Bus without X11 $DISPLAY
  GICOMP   GtkSource-4.gir

** (g-ir-compiler:9024): CRITICAL **: 10:52:32.890: Failed to get
connection to xfconfd: Cannot autolaunch D-Bus without X11 $DISPLAY
gmake[4]: *** [/usr/pkg/share/gobject-introspection-1.0/Makefile.introspection:172:
GtkSource-4.typelib] Trace/BPT trap (core dumped)
gmake[4]: Leaving directory
'/usr/pkgsrc/x11/gtksourceview4/work/gtksourceview-4.0.2/gtksourceview'
gmake[3]: *** [Makefile:1651: all-recursive] Error 1
gmake[3]: Leaving directory
'/usr/pkgsrc/x11/gtksourceview4/work/gtksourceview-4.0.2/gtksourceview'
gmake[2]: *** [Makefile:982: all] Error 2
gmake[2]: Leaving directory
'/usr/pkgsrc/x11/gtksourceview4/work/gtksourceview-4.0.2/gtksourceview'
gmake[1]: *** [Makefile:578: all-recursive] Error 1
gmake[1]: Leaving directory
'/usr/pkgsrc/x11/gtksourceview4/work/gtksourceview-4.0.2'
gmake: *** [Makefile:485: all] Error 2
*** Error code 2

Stop.
make[1]: stopped in /usr/pkgsrc/x11/gtksourceview4
*** Error code 1
.....

It builds (and packages after taking off two files from PLIST, there
is no option-specific PLIST here).


On Wed, 23 Jan 2019 at 10:17, Chavdar Ivanov <ci4ic4%gmail.com@localhost> wrote:
>
> Well, as expected, it breaks. I couldn't find anything obvious in the
> copious log file. However, I ran the same test for xfce4-thunar, which
> fails the same way when there is no DISPLAY set; it failed with the
> default options, but succeeded when I took off 'introspection'. So it
> seems the same should happen with webkit-gtk, which also has this
> option, whatever it means (obviously, I don't know that and don't know
> how and if the functionality would be changed if this option is
> unset). I'll try this later.
>
> Chavdar
>
>
>
> On Tue, 22 Jan 2019 at 22:25, Chavdar Ivanov <ci4ic4%gmail.com@localhost> wrote:
> >
> > webkit-gtk built fine on both machines when DISPLAY was set, midori
> > works OK, although the WebGtk process seems to be aufully CPU
> > intensitive. I'll try to build once more without DISPLAY set to find
> > where it breaks.
> >
> > On Tue, 22 Jan 2019 at 20:40, Chavdar Ivanov <ci4ic4%gmail.com@localhost> wrote:
> > >
> > > It hasn't quite finished yet, but went past the 92% it was bailing out
> > > before. The only difference is the presence of the DISPLAY variable.
> > > Around 43% is the generation of the JavaScript .gir file which I
> > > noticed on one of the builds fails with the message that DISPLAY is
> > > not present, as mentioned earlier. So it appears it is the same
> > > problem, perhaps it has been succeeding for others are they were
> > > running  the build under an X server.
> > >
> > > Later.
> > >
> > > On Tue, 22 Jan 2019 at 16:40, Chavdar Ivanov <ci4ic4%gmail.com@localhost> wrote:
> > > >
> > > > In the second occasion, the process was killed as there was no more
> > > > swap available. I increased the memory of the VM substantially and
> > > > also configured both of them to use ccache as suggested, placing the
> > > > ccache directory in a md device. Later.
> > > >
> > > > On Tue, 22 Jan 2019 at 16:14, Chavdar Ivanov <ci4ic4%gmail.com@localhost> wrote:
> > > > >
> > > > > No luck. On the first box, with ccache setup, I got the same error as
> > > > > on my other message about the lack of DISPLAY:
> > > > > ....
> > > > > [ 43%] Linking CXX shared library ../../lib/libjavascriptcoregtk-4.0.so
> > > > > [ 44%] Built target JavaScriptCore
> > > > > Scanning dependencies of target JavaScriptCore-4-gir
> > > > > [ 44%] Generating ../../JavaScriptCore-4.0.gir
> > > > >
> > > > > ** (process:14188): CRITICAL **: 15:56:52.864: Failed to get
> > > > > connection to xfconfd: Cannot autolaunch D-Bus without X11 $DISPLAY
> > > > >
> > > > > ** (process:14203): CRITICAL **: 15:56:53.755: Failed to get
> > > > > connection to xfconfd: Cannot autolaunch D-Bus without X11 $DISPLAY
> > > > > [ 44%] Generating ../../JavaScriptCore-4.0.typelib
> > > > >
> > > > > ** (g-ir-compiler:17622): CRITICAL **: 15:56:53.986: Failed to get
> > > > > connection to xfconfd: Cannot autolaunch D-Bus without X11 $DISPLAY
> > > > > gmake[2]: *** [Source/JavaScriptCore/CMakeFiles/JavaScriptCore-4-gir.dir/build.make:61:
> > > > > JavaScriptCore-4.0.typelib] Trace/BPT trap (core dumped)
> > > > > gmake[1]: *** [CMakeFiles/Makefile2:587:
> > > > > Source/JavaScriptCore/CMakeFiles/JavaScriptCore-4-gir.dir/all] Error 2
> > > > > gmake: *** [Makefile:152: all] Error 2
> > > > > *** Error code 2
> > > > >
> > > > > Stop.
> > > > > make[1]: stopped in /usr/pkgsrc/www/webkit-gtk
> > > > > *** Error code 1
> > > > >
> > > > > Stop.
> > > > > make: stopped in /usr/pkgsrc/www/webkit-gtk....
> > > > >
> > > > > So the next try will be with a valid DISPLAY, I guess.
> > > > >
> > > > > On the second machine, where I started from scratch, I got:
> > > > > ----
> > > > > [ 59%] Building CXX object
> > > > > Source/WebCore/CMakeFiles/WebCore.dir/__/__/DerivedSources/WebCore/unified-sources/UnifiedSource20.cpp.o
> > > > > g++: internal compiler error: Killed (program cc1plus received signal 9)
> > > > > Please submit a full bug report,
> > > > > with preprocessed source if appropriate.
> > > > > See <http://www.NetBSD.org/support/send-pr.html> for instructions.
> > > > > gmake[2]: *** [Source/WebCore/CMakeFiles/WebCore.dir/build.make:3786:
> > > > > Source/WebCore/CMakeFiles/WebCore.dir/__/__/DerivedSources/WebCore/unified-sources/UnifiedSource20.cpp.o]
> > > > > Error 4
> > > > > gmake[1]: *** [CMakeFiles/Makefile2:888:
> > > > > Source/WebCore/CMakeFiles/WebCore.dir/all] Error 2
> > > > > gmake: *** [Makefile:152: all] Error 2
> > > > > *** Error code 2
> > > > >
> > > > > Stop.
> > > > > make[1]: stopped in /usr/pkgsrc/www/webkit-gtk
> > > > > *** Error code 1
> > > > >
> > > > > Stop.
> > > > >
> > > > > which is perhaps an exhaused resource - it is just 2CPU 2GB VirtualBox
> > > > > guest. I'll boost it and try again.
> > > > >
> > > > > As I said, I am not particularly bohered by the lack of the package,
> > > > > but by the points one could get to whilst using pkgsrc...
> > > > >
> > > > >
> > > > > On Tue, 22 Jan 2019 at 14:57, Chavdar Ivanov <ci4ic4%gmail.com@localhost> wrote:
> > > > > >
> > > > > > Thanks for the tip! I'll try it, haven't so far. I am actually trying
> > > > > > to build it on a copletely different machine, starting with an empty
> > > > > > /usr/pkg, on -current as well (albeit a bit older, from early
> > > > > > January).
> > > > > >
> > > > > > On Tue, 22 Jan 2019 at 14:53, Greg Troxel <gdt%lexort.com@localhost> wrote:
> > > > > > >
> > > > > > > Chavdar Ivanov <ci4ic4%gmail.com@localhost> writes:
> > > > > > >
> > > > > > > >> This can often be a symptom of building where the path is a symlink.
> > > > > > > >>
> > > > > > > >> make clean, and then
> > > > > > > >>
> > > > > > > >> cd `pwd -P`
> > > > > > > >
> > > > > > > > I had it tried twice before, always cleaning. Also:
> > > > > > > > ...
> > > > > > > > # pwd
> > > > > > > > /usr/pkgsrc/www/webkit-gtk
> > > > > > > > # pwd -P
> > > > > > > > /usr/pkgsrc/www/webkit-gtk
> > > > > > > > ...
> > > > > > > >
> > > > > > > > (pkgsrc is on /).
> > > > > > >
> > > > > > > It sounds like that is not the issue.
> > > > > > >
> > > > > > > > I cleaned PATH and am trying again. There were no suspicious links,
> > > > > > > > except /usr/src and /usr/xsrc, but these should not have been used; I
> > > > > > > > deleted them anyway.
> > > > > > >
> > > > > > > So proabbly something else
> > > > > > >
> > > > > > > > Trouble is, it happens at 92% of the execution, and it takes hours to
> > > > > > > > get there...
> > > > > > >
> > > > > > > ccache!
> > > > > > >
> > > > > > > Seriously, I use ccache for pkgsrc, and it helps a lot, especially for
> > > > > > > debugging.
> > > > > > >
> > > > > > > CCACHE_DIR?=            ${HOME}/.ccache
> > > > > > > PKGSRC_COMPILER=        ccache gcc
> > > > > > >
> > > > > > > (if you are using gcc)
> > > > > > >
> > > > > > > and do 'ccache -M5G' which is a reasonable size.
> > > > > > >
> > > > > > > It's fine for ~/.ccache to be a symlink off someplace.
> > > > > > >
> > > > > > >
> > > > > > > So all in all somebody will have to trace what's happening.
> > > > > >
> > > > > >
> > > > > >
> > > > > > --
> > > > > > ----
> > > > >
> > > > >
> > > > >
> > > > > --
> > > > > ----
> > > >
> > > >
> > > >
> > > > --
> > > > ----
> > >
> > >
> > >
> > > --
> > > ----
> >
> >
> >
> > --
> > ----
>
>
>
> --
> ----



-- 
----