112635 - [Xlib] XXlibRgb revamp ...

Status: RESOLVED FIXED

(Core :: XUL, enhancement, P1)

Description

[Roland Mainz]

I'll file an all-in-one patch to fix the following bugs/RFEs:
- bug 112634 [Xlib] Support NS4.x "-visual" option to force a specific visual
- bug 83355  [xlib] Xlib port doesn't use MIT-SHM extension
- bug 79100  Xprint does not print grayscale

Comment 1

[Roland Mainz]

Accepting bug ...

Comment 2

[Roland Mainz]

More items to fix:
- bug 112688 [Xlib] StaticColor8 not working in Mozilla

Comment 3

[Roland Mainz]

Attachment: Patch for 2001-11-26-08-trunk

All-in-one patch for all issues ...

Comment 4

[Roland Mainz]

Requesting r=/sr= ...

Comment 5

[Roland Mainz]

Attachment: Better patch for 2001-11-26-08-trunk

Better patch, I fixed a crash which occured during browser buster stress
testing ...

Comment 6

[Donnie Cranford]

I will be reviewing this patch on Solaris 8 Sparc

Comment 7

[Donnie Cranford]

gisburn

The patch went through fine but did not compile correctly

make[4]: Entering directory `/export/home/mozilla/gfx/src/xlibrgb'
xlibrgb.c
Building deps for xlibrgb.c
/opt/SUNWspro/bin/cc -o xlibrgb.o -c -DOSTYPE=\"SunOS5\" -DOSARCH=\"SunOS\"
-DOJI -D_IMPL_NS_GFXONXP -DUSE_MOZILLA_TYPES   -I../../../dist/include/xlibrgb
-I../../../dist/include -I/export/home/mozilla/dist/include/nspr     
-I/usr/local/include/gtk-1.2 -I/usr/local/include/glib-1.2
-I/usr/local/lib/glib/include -I/usr/openwin/include -I/usr/openwin/include  
-KPIC -I/usr/openwin/include -mt  -DNDEBUG -DTRIMMED -O -I/usr/openwin/include
-DBROKEN_QSORT=1 -DNSCAP_DISABLE_DEBUG_PTR_TYPES=1 -DD_INO=d_ino
-DSTDC_HEADERS=1 -DHAVE_ST_BLKSIZE=1 -DHAVE_INT16_T=1 -DHAVE_INT32_T=1
-DHAVE_INT64_T=1 -DHAVE_UINT=1 -DHAVE_UINT_T=1 -DHAVE_UINT16_T=1
-DSTDC_HEADERS=1 -DHAVE_DIRENT_H=1 -DHAVE_SYS_BYTEORDER_H=1 -DHAVE_MEMORY_H=1
-DHAVE_UNISTD_H=1 -DHAVE_NL_TYPES_H=1 -DHAVE_X11_XKBLIB_H=1
-DHAVE_SYS_STATVFS_H=1 -DHAVE_SYS_STATFS_H=1 -DHAVE_SYS_VFS_H=1
-DHAVE_SYS_MOUNT_H=1 -DHAVE_LIBM=1 -DHAVE_LIBDL=1 -DHAVE_LIBSOCKET=1
-D_REENTRANT=1 -DHAVE_RANDOM=1 -DHAVE_STRERROR=1 -DHAVE_LCHOWN=1 -DHAVE_FCHMOD=1
-DHAVE_SNPRINTF=1 -DHAVE_LOCALTIME_R=1 -DHAVE_STATVFS=1 -DHAVE_MEMMOVE=1
-DHAVE_RINT=1 -DHAVE_NL_LANGINFO=1 -DHAVE_STRTOK_R=1 -DHAVE_IOS_BINARY=1
-DHAVE_CPP_EXPLICIT=1 -DHAVE_CPP_SPECIALIZATION=1
-DHAVE_CPP_MODERN_SPECIALIZE_TEMPLATE_SYNTAX=1
-DHAVE_CPP_PARTIAL_SPECIALIZATION=1 -DHAVE_CPP_ACCESS_CHANGING_USING=1
-DHAVE_CPP_AMBIGUITY_RESOLVING_USING=1 -DHAVE_CPP_NAMESPACE_STD=1
-DHAVE_CPP_UNAMBIGUOUS_STD_NOTEQUAL=1 -DHAVE_CPP_NEW_CASTS=1
-DHAVE_CPP_DYNAMIC_CAST_TO_VOID_PTR=1 -DNEED_CPP_UNUSED_IMPLEMENTATIONS=1
-DHAVE_I18N_LC_MESSAGES=1 -DMOZ_DEFAULT_TOOLKIT=\"gtk\" -DMOZ_WIDGET_GTK=1
-DMOZ_ENABLE_XREMOTE=1 -DMOZ_X11=1 -DIBMBIDI=1 -DACCESSIBILITY=1 -DMOZ_LOGGING=1
-DMOZ_USER_DIR=\".mozilla\" -DMOZ_XUL=1 -DINCLUDE_XUL=1 -DNS_MT_SUPPORTED=1
-DUSE_IMG2=1 -DMOZ_DLL_SUFFIX=\".so\" -DXP_UNIX=1 -DUNIX_ASYNC_DNS=1
-DJS_THREADSAFE=1 -DNS_PRINT_PREVIEW=1  -DMOZILLA_CLIENT xlibrgb.c
"xlibrgb.c", line 282: prototype mismatch: 10 args passed, 8 expected
cc: acomp failed for xlibrgb.c
gmake[4]: *** [xlibrgb.o] Error 2
gmake[4]: Leaving directory `/export/home/mozilla/gfx/src/xlibrgb'
gmake[3]: *** [libs] Error 2
gmake[3]: Leaving directory `/export/home/mozilla/gfx/src'
gmake[2]: *** [libs] Error 2
gmake[2]: Leaving directory `/export/home/mozilla/gfx'
gmake[1]: *** [tier_9] Error 2
gmake[1]: Leaving directory `/export/home/mozilla'
gmake: *** [default] Error 2

Comment 8

[Roland Mainz]

Attachment: New patch for 2001-11-26-08-trunk

I fixed dcran's build problems and added a "safeguard" that this will never
happen again ...

Comment 9

[Donnie Cranford]

After having Roland's new patch all compiled fine and moz is running like a champ.

Solaris 8 10/01

r=dcran

Comment 10

[Roland Mainz]

Requesting sr= ...

Comment 11

[pocemit]

<pocemit> In file included from
          ../../../../mozilla/widget/src/xlibxtbin/xlibxtbin.cpp:23:
<pocemit> ../../../dist/include/xlibrgb/xlibrgb.h:101: parse error before
          `template'


line 101:
<pocemit> xxlib_rgb_create_handle (const char *name,
<pocemit>                          Display *display, Screen *screen,
<pocemit>                          XVisualInfo *template, long template_mask);

We don't want to call a variable "template" since thats a c++ reserved word...

s/template/xtemplate

tim

Comment 12

[Roland Mainz]

Comment on attachment 59877 [details] [diff] [review]
New patch for 2001-11-26-08-trunk

pocemit found some issues on LSB machines ...

Comment 13

[Roland Mainz]

Attachment: New patch for 2001-11-30-08-trunk

Updated patch per pocemit's comments

Comment 14

[pocemit]

Attachment: the demonstration

still some MSB/LSB issues...
Images are broken if SHM is used. if SHM is disabled on command line, images
are OK. (only some images are broken with SHM).

Comment 15

[Roland Mainz]

Comment on attachment 60451 [details] [diff] [review]
New patch for 2001-11-30-08-trunk

I need to build the Zilla on a LSB machine.

I'll file a patch with MIT-SHM turned-off for now... I need the other bugs
fixed and kill this one next week (e.g. after 0.9.7 freeze) ...

Comment 16

[Roland Mainz]

Attachment: New patch for 2001-11-30-08-trunk

Same as attachment 60451 [details] [diff] [review] but with MIT-SHM support turned off for now. I#ll
defer that until I have a test configuration ...

Comment 17

[Roland Mainz]

Attachment: New patch for 2001-11-30-08-trunk

|#ifdef|'ed MIT-SHM support out - not all platforms have support for this. I
filed bug 113743 ("Need configure-test for X11/extensions/XShm.h") to get a
configure.in test - we can reenable this code once we have a test for it ...

Comment 18

[pocemit]

works for me with the latest patch, of course without SHM.
Thanks gisburn.
BTW, what about renaming xxlib_rgb back into xlib_rgb_?

tim

Comment 19

[Roland Mainz]

pocemit write:
> works for me with the latest patch, of course without SHM.
> Thanks gisburn.

Sorry about that, I need to reproduce the color problems on little-endian
machines first - but I assume the patch to fix that will be much smaller than
this monster.
If we get sr= for this one fast I can try to get MIT-SHM working in 0.9.7 ...
hopefully ...

> BTW, what about renaming xxlib_rgb back into xlib_rgb_?

That would add another huge pile of changes to the patch (which is already
incredible big). Not now... later. It's on my ToDo list ...

Comment 20

[jag (Peter Annema)]

pocemit, have you done a code level review of this?

Comment 21

[pocemit]

A large part of the shm code was from a bug I opened earlier for SHM support in
xlib port. Yes, I looked over the last posted patch and it looks OK. It will
probably not apply cleanly by now though.

Comment 22

[Roland Mainz]

Yup... the patch has rotten a lot (well... it's an month old patch).
I'll file an updated patch in a few hours ...

Comment 23

[Roland Mainz]

Suprise:
-- snip --
% gpatch -p0 --dry-run <attachment.cgi\?id\=60592\&action\=view 
patching file mozilla/gfx/src/xlib/Makefile.in
Hunk #1 succeeded at 112 (offset 7 lines).
patching file mozilla/gfx/src/xlib/XIE.c
patching file mozilla/gfx/src/xlibrgb/xlibrgb.c
patching file mozilla/gfx/src/xlibrgb/xlibrgb.h
patching file mozilla/gfx/src/xprint/nsXPrintContext.cpp
Hunk #2 succeeded at 175 (offset 5 lines).
patching file mozilla/gfx/src/xprint/xprintutil.h
patching file mozilla/widget/src/xlib/nsAppShell.cpp
-- snip --

The patch still works!

jag ?

Comment 24

[jag (Peter Annema)]

Too much xlib/xprint fu for me. blizzard, care to give this a shot?

Comment 25

[Marc Attinasi]

Comment on attachment 60592 [details] [diff] [review]
New patch for 2001-11-30-08-trunk

rs=attinasi - rubber stamping more or less, this is not my area of experteise,
but the code looks good and I see no obvious errors. I'm running on the
assumption that this is tested to work and not cause leaks...

Comment 26

[Roland Mainz]

Fix checked in
(http://bonsai.mozilla.org/cvsquery.cgi?module=MozillaTinderboxAll&branch=HEAD&cvsroot=/cvsroot&date=explicit&mindate=1011137220&maxdate=1011137640&who=timeless%25mac.com),
marking bug as FIXED.