316845 - Build error in gfx/cairo/libpixman/src/fbpict.c: symbol `_cairo_pixman_composite' is already defined

Status: VERIFIED FIXED

(Core :: Graphics, defect)

Description

[Hideo Oshima]

gcc -o fbpict.o -c  -DMOZILLA_INTERNAL_API -DOSTYPE=\"Linux2.6\" -DOSARCH=\"Linux\" -DBUILD_ID=0000000000 -I.  -I../../../../dist/include/libpixman -I../../../../dist/include -I../../../../dist/include/nspr    -I../../../../dist/sdk/include -I/usr/X11R6/include   -fPIC -I/usr/X11R6/include -Wall -W -Wno-unused -Wpointer-arith -Wcast-align -Wno-long-long -pedantic -pthread -pipe  -DNDEBUG -DTRIMMED -O2  -I/usr/X11R6/include -include ../../../../mozilla-config.h -DMOZILLA_CLIENT -Wp,-MD,.deps/fbpict.pp fbpict.c
fbpict.c: In function '_cairo_pixman_composite':
fbpict.c:1421: warning: enumeration value 'PIXMAN_OPERATOR_CLEAR' not handled in switch
fbpict.c:1421: warning: enumeration value 'PIXMAN_OPERATOR_DST' not handled in switch
fbpict.c:1421: warning: enumeration value 'PIXMAN_OPERATOR_OVER_REVERSE' not handled in switch
fbpict.c:1421: warning: enumeration value 'PIXMAN_OPERATOR_IN' not handled in switch
fbpict.c:1421: warning: enumeration value 'PIXMAN_OPERATOR_IN_REVERSE' not handled in switch
fbpict.c:1421: warning: enumeration value 'PIXMAN_OPERATOR_OUT' not handled in switch
fbpict.c:1421: warning: enumeration value 'PIXMAN_OPERATOR_OUT_REVERSE' not handled in switch
fbpict.c:1421: warning: enumeration value 'PIXMAN_OPERATOR_ATOP' not handled in switch
fbpict.c:1421: warning: enumeration value 'PIXMAN_OPERATOR_ATOP_REVERSE' not handled in switch
fbpict.c:1421: warning: enumeration value 'PIXMAN_OPERATOR_XOR' not handled in switch
fbpict.c:1421: warning: enumeration value 'PIXMAN_OPERATOR_SATURATE' not handled in switch
fbpict.c:1870: warning: comparison between signed and unsigned
fbpict.c:1876: warning: comparison between signed and unsigned
fbpict.c:1885: warning: comparison between signed and unsigned
fbpict.c:1891: warning: comparison between signed and unsigned
{standard input}: Assembler messages:
{standard input}:5717: Error: symbol `_cairo_pixman_composite' is already defined

.mozconfig is follows
mk_add_options MOZ_CO_PROJECT=browser
# Options for 'configure' (same as command-line options).
ac_add_options --enable-application=browser
ac_add_options --enable-default-toolkit=gtk2
ac_add_options --disable-debug
ac_add_options --disable-tests
ac_add_options --enable-xft
ac_add_options --disable-freetype2
ac_add_options --enable-pango
ac_add_options --enable-svg
ac_add_options --enable-cairo
ac_add_options --enable-canvas
ac_add_options --enable-static
ac_add_options --disable-shared
ac_add_options --enable-optimize='-O2'

Debian GNU/Linux Sid, gcc 4.0.2-4

Comment 1

[Manish Singh]

Attachment: Patch to fix the issue

This fixes breakage seen with newer binutils. This is already in upstream cairo CVS.

Comment 2

[Hideo Oshima]

I can build with the patch(attachment 203761 [details] [diff] [review]).
Thanks for your help.

Can anyoune check in the patch, or must I wait new cairo source 
into the mozilla source code tree?

Comment 3

[Calum Mackay]

I'm also seeing:

gcc -o icformat.o -c  -DMOZILLA_INTERNAL_API -DOSTYPE=\"Linux2.6\" -DOSARCH=\"Li
nux\" -DBUILD_ID=2005112105 -I.  -I../../../../dist/include/libpixman -I../../..
/../dist/include -I../../../../dist/include/nspr    -I../../../../dist/sdk/inclu
de    -fPIC  -Wall -W -Wno-unused -Wpointer-arith -Wcast-align -Wno-long-long -p
edantic -pthread -pipe  -DNDEBUG -DTRIMMED -O2 -march=i686   -include ../../../.
./mozilla-config.h -DMOZILLA_CLIENT -Wp,-MD,.deps/icformat.pp icformat.c
{standard input}: Assembler messages:
{standard input}:160: Error: symbol `_cairo_pixman_color_to_pixel' is already de
fined
make[6]: *** [iccolor.o] Error 1
make[6]: *** Waiting for unfinished jobs....
icformat.c: In function ‘_cairo_pixman_format_init’:
icformat.c:106: warning: signed and unsigned type in conditional expression
icformat.c:112: warning: signed and unsigned type in conditional expression
icformat.c:116: warning: signed and unsigned type in conditional expression
icformat.c:119: warning: signed and unsigned type in conditional expression
icformat.c:125: warning: signed and unsigned type in conditional expression
icformat.c:131: warning: signed and unsigned type in conditional expression
icformat.c:135: warning: signed and unsigned type in conditional expression
icformat.c:138: warning: signed and unsigned type in conditional expression
icformat.c:145: warning: signed and unsigned type in conditional expression
{standard input}: Assembler messages:
{standard input}:5681: Error: symbol `_cairo_pixman_composite' is already define
d
make[6]: *** [fbpict.o] Error 1
make[5]: *** [libs] Error 2
make[5]: Leaving directory `/usr/local/src/mozilla/cvs/mozilla/gfx/cairo'


related?

Comment 4

[Vladimir Vukicevic [:vlad] [:vladv] (needinfo me, slow to respond)]

This will be fixed on the trunk shortly; I'm waiting to sync up a few windows/linux fixes, and then I'll merge the current status of the cairo work to the trunk.

Comment 5

[Hideo Oshima]

Calum, I think your problem is the same as this.
Try the patch to fix.

Comment 6

[bugzilla]

I am seeing this build error on the 1.8 Branch when compiling on linux.

Comment 7

[Hideo Oshima]

Today I checkout all source codes again
and build Firefox on the same environment as comment 0.
No error happened without the patch.
Is this fixed?

Comment 8

[bugzilla]

(In reply to comment #7)
> Today I checkout all source codes again
> and build Firefox on the same environment as comment 0.
> No error happened without the patch.
> Is this fixed?
> 

This is fixed on the trunk but not the 1.8 branch.  The version on this bug should probably be changed from Trunk to 1.8 Branch by someone with sufficient bugzilla privileges.

Comment 9

[Stuart Parmenter]

*** Bug 323420 has been marked as a duplicate of this bug. ***

Comment 10

[Christian :Biesinger (don't email me, ping me on IRC)]

*** Bug 327720 has been marked as a duplicate of this bug. ***

Comment 11

[Steffen Wilberg]

This should probably block 1.8.1 since it breaks Firefox with SVG using gcc4.

Vlad, what are your plans for the 1.8 branch? Will you sync cairo from the trunk soon, or should we check in this patch in the meantime?

Comment 12

[Vladimir Vukicevic [:vlad] [:vladv] (needinfo me, slow to respond)]

There are no current plans to touch the version of cairo on the 1.8.1 branch; it's probably worth taking the patch.

Comment 13

[Vladimir Vukicevic [:vlad] [:vladv] (needinfo me, slow to respond)]

Comment on attachment 203761 [details] [diff] [review]
Patch to fix the issue

I can't a+ this, but we'll need this  to build.

Comment 14

[Masatoshi Kimura [:emk]]

(In reply to comment #13)
You just have to be a module owner to approve 1.8.1 branch.
See http://developer.mozilla.org/devnews/index.php/2006/01/30/tree-rules-for-the-18-branch/
Hence you (or pavlov) can approve this without trouble with drivers.
> I can't a+ this, but we'll need this  to build.
Feel free to a+ if you need :-)

BTW, isn't a superreview required to checkin to the cairo and thebes?

Comment 15

[Vladimir Vukicevic [:vlad] [:vladv] (needinfo me, slow to respond)]

Comment on attachment 203761 [details] [diff] [review]
Patch to fix the issue

a=me for 1.8.1

Comment 16

[Akkana Peck]

How about the 1.8.0 branch? I just hit it there.

Comment 17

[Zibi Braniecki [:zbraniecki][:gandalf]]

trunk doesn't need a=. I'll check it in today

Comment 18

[Stuart Parmenter]

(In reply to comment #14)
> BTW, isn't a superreview required to checkin to the cairo and thebes?
> 

we're on a r+sr system (from me or vlad) similar to layout and such.  if the r= isn't one of us please get a sr=.

Comment 19

[Zibi Braniecki [:zbraniecki][:gandalf]]

checked in
Checking in pixman-remap.h;
/cvsroot/mozilla/gfx/cairo/libpixman/src/pixman-remap.h,v  <--  pixman-remap.h
new revision: 1.1.6.3; previous revision: 1.1.6.2
done

Comment 20

[Christian :Biesinger (don't email me, ping me on IRC)]

(In reply to comment #17)
> trunk doesn't need a=. I'll check it in today

trunk also doesn't need this patch.

Comment 21

[Zibi Braniecki [:zbraniecki][:gandalf]]

(In reply to comment #20)
> (In reply to comment #17)
> > trunk doesn't need a=. I'll check it in today
> 
> trunk also doesn't need this patch.

Yes, of course. That's one of the reasons for which I didn't check it there ;)

Comment 22

[Darin Fisher]

vlad, pav: how badly do we want this on the 1.8.0.x branch?

Comment 23

[Vladimir Vukicevic [:vlad] [:vladv] (needinfo me, slow to respond)]

(In reply to comment #22)
> vlad, pav: how badly do we want this on the 1.8.0.x branch?

It doesn't affect anything but gcc4 compilation, so there's no reason not to.

Comment 24

[Jay Patel [:jay]]

Comment on attachment 203761 [details] [diff] [review]
Patch to fix the issue

a=jay for drivers on 1.8.0 branch.

Comment 25

[Zibi Braniecki [:zbraniecki][:gandalf]]

checked in to 1.8.0 branch

Comment 26

[Chris Cooper [:coop] (he/him)]

Verified by code inspection.