After an evaluation, GNOME has moved from Bugzilla to GitLab. Learn more about GitLab.
No new issues can be reported in GNOME Bugzilla anymore.
To report an issue in a GNOME project, go to GNOME GitLab.
Do not go to GNOME Gitlab for: Bluefish, Doxygen, GnuCash, GStreamer, java-gnome, LDTP, NetworkManager, Tomboy.
Bug 596654 - pango based apps take *forever* (~15 minutes) to launch on OSX
pango based apps take *forever* (~15 minutes) to launch on OSX
Status: RESOLVED DUPLICATE of bug 593240
Product: pango
Classification: Platform
Component: general
1.26.x
Other Mac OS
: Normal critical
: ---
Assigned To: pango-maint
pango-maint
Depends on:
Blocks:
 
 
Reported: 2009-09-28 19:26 UTC by Jeremy Huddleston
Modified: 2009-09-28 23:17 UTC
See Also:
GNOME target: ---
GNOME version: ---


Attachments
stdout/err with the requested debugging CPPFLAG (45.52 KB, text/plain)
2009-09-28 21:23 UTC, Jeremy Huddleston
Details

Description Jeremy Huddleston 2009-09-28 19:26:03 UTC
This is a regression over 1.24.

gtk-demo and xchat take about 10 minutes to display the connections window (gtk-demo has the same issue).  Attached is a 30-second sample about 4 minutes into the launch (gtk-demo's sample looks the same):

It's fairly deep, and here are the big tickets:

Sort by top of stack, same collapsed (when >= 5):
       hb_ot_layout_substitute_lookup        6768
       hb_ot_layout_position_lookup        6704
       PosLookupSubTable::apply(_hb_ot_layout_context_t*, _hb_buffer_t*, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int) const        5438
       SubstLookupSubTable::apply(_hb_ot_layout_context_t*, _hb_buffer_t*, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int) const        5191

Reverting to pango-1.24.5 corrects this problem.
Comment 1 Behdad Esfahbod 2009-09-28 19:28:29 UTC
Can you compile with CPPFLAGS=-DHB_DEBUG=100 and attach the log?
Comment 2 Jeremy Huddleston 2009-09-28 21:23:02 UTC
Created attachment 144202 [details]
stdout/err with the requested debugging CPPFLAG

Interestingly, when I compile with that CPPFLAG, I don't hit the bug.  gtk-demo starts up fast.

I recompiled again without the flag just to make sure it wasn't a fluke or compiler bug and it's back to bad performance.
Comment 3 Behdad Esfahbod 2009-09-28 21:27:19 UTC
Ok, this is definitely something PPC specific.  Is this a PPC?  Sounds like a dupe of bug 593240.  What compiler?

*** This bug has been marked as a duplicate of bug 593240 ***
Comment 4 Jeremy Huddleston 2009-09-28 23:13:41 UTC
Nope, this has nothing to do with PPC.  This is gcc-4.2.1:

i686-apple-darwin10-gcc-4.2.1 (GCC) 4.2.1 (Apple Inc. build 5651)
Copyright (C) 2007 Free Software Foundation, Inc.
This is free software; see the source for copying conditions.  There is NO
warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.

This is happening with both i386 and x86_64 code.
Comment 5 Behdad Esfahbod 2009-09-28 23:17:17 UTC
Comment on bug 593240 please.

*** This bug has been marked as a duplicate of bug 593240 ***