GNOME.org
 

gnome2 and OSF/1 4.0


Hi,
Over the last week or so, i have been giving porting gnome2 to OSF/1 4.0 a go ... to while away the spare time while my honours project is using up much cpu time on a supercomputer...
And things seem to not be going too badly... I got quite a bit of stuff running. But theres quite a few floating point exceptions and 'this application has crashed' message boxes which tend to like showing their heads. Anyway, I was interested in getting rid of a few of these remaining bugs (or helping someone else get rid of them - I am certainly no wizard with this gnome code - only having looked at any of it for the first time in the last week), maybe contributing a few patches which are nesecery to build gnome on OSF/1 4.0 (at least in my relatively built up configuration - I am using a patched gcc 3.2 HEAD rather then system compiler, among many other things). To this end I have started rebuilding everything again (although I didnt clean out anything before I started) - this time making notes as I go. The notes are attached for anyones reference. This time I am paying more attention to the testsuites - and noticing a significant number of failures in a couple of packages. I am building from the packages in /pub/gnome/pre-gnome2/latest/sources/ - rather then cvs ... for 2 reasons, 1. my autoconf/automake is 2.53/1.6 at the moment so autogen.sh can get a bit nasty and 2. I make the bold leap of faith and hope they are more likely to compile. ;)
anyway ... Any suggestions as to what I should do next, if this is the wrong place to report this sort of thing... etc etc - would be great. 

Regards,
Gareth




_________________________________________________________________
MSN Photos is the easiest way to share and print your photos: http://photos.msn.com/support/worldwide.aspx 
#0  gtk_range_calc_layout (range=0x140053700, adjustment_value=0)
   at gtkrange.c:2031
#1  0x30000943df0 in gtk_range_size_allocate (widget=0x140053700,
   allocation=0x11fffd0a0) at gtkrange.c:755
#2  0x30002027130 in g_cclosure_marshal_VOID__BOXED
(closure=0x140002d80,
   return_value=0x11fffd0a0, n_param_values=14,
param_values=0x11fffd340,
   invocation_hint=0x0, marshal_data=0x30000943d90) at gmarshal.c:566
#3  0x3000200cab0 in g_type_class_meta_marshal (closure=0x140002d80,
   return_value=0x0, n_param_values=2, param_values=0x11fffd340,
   invocation_hint=0x11fffd200, marshal_data=0xf8) at gclosure.c:514
#4  0x3000200c5e8 in g_closure_invoke (closure=0x140002d80,
return_value=0x0,
   n_param_values=2, param_values=0x11fffd340,
invocation_hint=0x11fffd200)
   at gclosure.c:437
#5  0x30002025344 in signal_emit_unlocked_R (node=0x140024a40, detail=0,
   instance=0x140053700, emission_return=0x0,
instance_and_params=0x11fffd340)
   at gsignal.c:2271
#6  0x30002024890 in g_signal_emit_valist (instance=0x140053700,
signal_id=1,
   detail=0, var_args={__base = 0x11fffd5d0, __offset = 24}) at
gsignal.c:2100
#7  0x30000963270 in gtk_signal_emit (object=0x140053700, signal_id=10)
   at gtksignal.c:355
#8  0x30000a2b270 in gtk_widget_size_allocate (widget=0x140053700,
   allocation=0x1) at gtkwidget.c:2323
#9  0x30000982e10 in gtk_table_size_allocate_pass2 (table=0x140053700)
---Type <return> to continue, or q <return> to quit---
   at gtktable.c:1645
#10 0x30002027130 in g_cclosure_marshal_VOID__BOXED
(closure=0x140002d80,
   return_value=0x11fffd0a0, n_param_values=14,
param_values=0x11fffd910,
   invocation_hint=0x0, marshal_data=0x30000981000) at gmarshal.c:566
#11 0x3000200cab0 in g_type_class_meta_marshal (closure=0x140002d80,
   return_value=0x0, n_param_values=2, param_values=0x11fffd910,
   invocation_hint=0x11fffd7d0, marshal_data=0xf8) at gclosure.c:514
#12 0x3000200c5e8 in g_closure_invoke (closure=0x140002d80,
return_value=0x0,
   n_param_values=2, param_values=0x11fffd910,
invocation_hint=0x11fffd7d0)
   at gclosure.c:437
#13 0x30002025344 in signal_emit_unlocked_R (node=0x140024a40, detail=0,
   instance=0x14002e480, emission_return=0x0,
instance_and_params=0x11fffd910)
   at gsignal.c:2271
#14 0x30002024890 in g_signal_emit_valist (instance=0x14002e480,
signal_id=1,
   detail=0, var_args={__base = 0x11fffdba0, __offset = 24}) at
gsignal.c:2100
#15 0x30000963270 in gtk_signal_emit (object=0x14002e480, signal_id=10)
   at gtksignal.c:355
#16 0x30000a2b270 in gtk_widget_size_allocate (widget=0x14002e480,
   allocation=0x1) at gtkwidget.c:2323
#17 0x30000a21748 in gtk_vbox_size_allocate (widget=0x14002e160,
   allocation=0x11fffe230) at gtkvbox.c:257
#18 0x30002027130 in g_cclosure_marshal_VOID__BOXED
(closure=0x140002d80,
   return_value=0x11fffd0a0, n_param_values=14,
param_values=0x11fffdf30,
---Type <return> to continue, or q <return> to quit---
   invocation_hint=0x0, marshal_data=0x30000a21550) at gmarshal.c:566
#19 0x3000200cab0 in g_type_class_meta_marshal (closure=0x140002d80,
   return_value=0x0, n_param_values=2, param_values=0x11fffdf30,
   invocation_hint=0x11fffddf0, marshal_data=0xf8) at gclosure.c:514
#20 0x3000200c5e8 in g_closure_invoke (closure=0x140002d80,
return_value=0x0,
   n_param_values=2, param_values=0x11fffdf30,
invocation_hint=0x11fffddf0)
   at gclosure.c:437
#21 0x30002025344 in signal_emit_unlocked_R (node=0x140024a40, detail=0,
   instance=0x14002e160, emission_return=0x0,
instance_and_params=0x11fffdf30)
   at gsignal.c:2271
#22 0x30002024890 in g_signal_emit_valist (instance=0x14002e160,
signal_id=1,
   detail=0, var_args={__base = 0x11fffe1c0, __offset = 24}) at
gsignal.c:2100
#23 0x30000963270 in gtk_signal_emit (object=0x14002e160, signal_id=10)
   at gtksignal.c:355
#24 0x30000a2b270 in gtk_widget_size_allocate (widget=0x14002e160,
   allocation=0x1) at gtkwidget.c:2323
#25 0x30000a3e204 in gtk_window_size_allocate (widget=0x14002e200,
   allocation=0x11fffe7f0) at gtkwindow.c:3402
#26 0x30002027130 in g_cclosure_marshal_VOID__BOXED
(closure=0x140002d80,
   return_value=0x11fffd0a0, n_param_values=14,
param_values=0x11fffe4f0,
   invocation_hint=0x0, marshal_data=0x30000a3e0f0) at gmarshal.c:566
#27 0x3000200cab0 in g_type_class_meta_marshal (closure=0x140002d80,
   return_value=0x0, n_param_values=2, param_values=0x11fffe4f0,
---Type <return> to continue, or q <return> to quit---
   invocation_hint=0x11fffe3b0, marshal_data=0xf8) at gclosure.c:514
#28 0x3000200c5e8 in g_closure_invoke (closure=0x140002d80,
return_value=0x0,
   n_param_values=2, param_values=0x11fffe4f0,
invocation_hint=0x11fffe3b0)
   at gclosure.c:437
#29 0x30002025344 in signal_emit_unlocked_R (node=0x140024a40, detail=0,
   instance=0x14002e200, emission_return=0x0,
instance_and_params=0x11fffe4f0)
   at gsignal.c:2271
#30 0x30002024890 in g_signal_emit_valist (instance=0x14002e200,
signal_id=1,
   detail=0, var_args={__base = 0x11fffe780, __offset = 24}) at
gsignal.c:2100
#31 0x30000963270 in gtk_signal_emit (object=0x14002e200, signal_id=10)
   at gtksignal.c:355
#32 0x30000a2b270 in gtk_widget_size_allocate (widget=0x14002e200,
   allocation=0x1) at gtkwidget.c:2323
#33 0x30000a3d760 in gtk_window_show (widget=0x14002e200) at
gtkwindow.c:3059
#34 0x3000202617c in g_cclosure_marshal_VOID__VOID (closure=0x1400036e0,
   return_value=0x11fffd0a0, n_param_values=14, param_values=0x1,
   invocation_hint=0x0, marshal_data=0x30000a3d5e0) at gmarshal.c:77
#35 0x3000200cab0 in g_type_class_meta_marshal (closure=0x1400036e0,
   return_value=0x0, n_param_values=1, param_values=0x11fffeb00,
   invocation_hint=0x11fffe9c0, marshal_data=0xb0) at gclosure.c:514
#36 0x3000200c5e8 in g_closure_invoke (closure=0x1400036e0,
return_value=0x0,
   n_param_values=1, param_values=0x11fffeb00,
invocation_hint=0x11fffe9c0)
   at gclosure.c:437
---Type <return> to continue, or q <return> to quit---
#37 0x30002025344 in signal_emit_unlocked_R (node=0x14001b4a0, detail=0,
   instance=0x14002e200, emission_return=0x0,
instance_and_params=0x11fffeb00)
   at gsignal.c:2271
#38 0x30002024890 in g_signal_emit_valist (instance=0x14002e200,
signal_id=3,
   detail=0, var_args={__base = 0x11fffed90, __offset = 16}) at
gsignal.c:2100
#39 0x30000963270 in gtk_signal_emit (object=0x14002e200, signal_id=3)
   at gtksignal.c:355
#40 0x30000a29808 in gtk_widget_show (widget=0x14002e200) at
gtkwidget.c:1637
#41 0x12000373c in main (argc=1, argv=0x11fffee68) at
testselection.c:486



All done with --prefix=$HOME and no root access
gtk-doc 0.9
-- all works good.
-- no testsuite.
-- but I dont have jade - shouldnt it complain about that?  Lack of jade
means index.sgml cant be made - meaning many of the below 'just fine'
make installs are actually dieing part way, because they dont ignore
errors in doc install. (but I manually bypass that problem, and if make
install works otherwise I have labeled it as fine) (jade doesnt seem to
like g++ 3.2 either :P)

glib 2.0.1
-- compiles perfectly including --with-threads
-- testsuite pass all 29 tests.

pango 1.0.1
-- compiles perfectly - no extra configure flags.
-- testsuite passed all 1 test.

atk 1.0.1
-- compiles perfectly - no extra configure flags.
-- no testsuite tests.

gtk+ 2.0.1
-- fails straight compile - no XkbKeySymEntry in osf/1 4.0d and 4.0f
can possibly avoid using that specific function. But I dont know how.
-- compiles perfectly with --disable-xkb
-- no testsuite tests.
-- make install causes make (3.79.1) to core dump in
install-gtkincludeHEADERS - manual installation and removing associated
make line allowed continuation.
-- testselection - core dump floating point error - gtkrange.c line 2031
according to my not very reliable instalation of gdb. adding 0 and 1 ...
something very odd... - hmmm - probably not the real line -g -O2 combo
problem.
see backtrace.txt anyway.
-- all other manual tests seem pretty much fine.

libIDL 0.7.4
-- depends on glib? - but not shown on gnome2.0 dependency diagram.
-- compiles perfectly - no extra confgiure flags.
-- no testsuite tests.

linc 0.1.20
-- unsure as to if --with-openssl does anything or not, but it is set.
-- added #if __osf__ #undef AF_INET6 #endif - to linc-protocols.c
-- copied 'NI_MAXHOST' related stuff from linc-protocols to
linc-server.c
-- added #if __osf__ #define socklen_t size_t #endif to
include/linc/linc-protocols.h
-- then compiles perfectly
-- fails testsuite - test-blocking fails at g_assert (buffer[i] ==
idx++) when buffer == 1024 - buffer[i] is 32 and idx is 16.
-- closer analysis shows buffer is recieving 16 out of every 32 sent,
concated, and the rest blank at the end.
-- consider possibly a 64bit problem? yes it is, test code assumes 4byte
longs.
-- removing that assert causes the test to pass.

libxml2 2.4.19
-- --with-threads --with-history
-- need to add -mieee to compilation of xpath.c
-- testsuite works, python tests show libxml2mod.so with unresolved
vsnprintf - need to add -ldb to Makefile for python for silly osf.
-- python testsuite now has tstxpath.py 8 byte memory leak, xpathext.py
8 byte memory leak - xpathret.py 4 byte memory leak.
-- unsure if they are actually considered errors? since they dont
exit(1).

libglade 1.99.10
-- no additional configure flags.
-- compiles fine straight off.
-- passes both tests.
-- make install needed a tiny hand because i am not root and my xml
catalog is local not system. I do have ROOTCATALOG env variable set.

libxslt 1.0.15
-- no additional configure flags.
-- compiles fine straight off.
-- make check ... test-2.5-1.xsl line 6 and line 2 complains.
bug-60.csl - line 6 - unknown xsl:foo-of
fpe? - means i got
/bin/ksh: 'number - probably pid' Floating exeception(coredump)
Fatal error, no 'testname'.res
datetime.2.xml - fpe?
gday.1.xml - fpe?
gday.2.xml - fpe?
gmonth.1.xml - fpe?
gmonth.2.xml - fpe?
gmonthday.1.xml -fpe?
gnomthday.2.xml - fpe?
gyear.1.xml - fpe?
gyear.2.xml - fpe?
gyearmonth.1.xml - fpe?
gyearmonth.2.xml - fpe?
time.1.xml - fpe?
time.1.xml - fpe?
-- python tests - 'could not guess RTLD_GLOBAL and RTLD_NOW for OSF1' -
but still return OK.
-- make install - gtkdoc-fixxref called with module libxslt rather then
libxslt-1.0.15 - otherwise seems okay.

ORBit2 2.3.107
-- no additional configure switches.
-- compiles straight off.
-- all tests pass cleanly.
-- make install - looks fine.


GConf 1.1.9
-- no additional configure switches
-- compiles straight off
-- no auto testsuite
-- (not checked the examples)
-- make install looks fine.

libart_lgpl 2.3.8
-- no additional configure switches
-- make straight off
-- no auto testsuite
-- manual tests run (havent checked output in a viewer)
-- make install looks fine.

libgnomecanvas 1.114.0
-- no additional configure switches
-- make fails in demos dueto libraries not being passed to libtool link
stage. - rather odd.
-- added LIBGNOMECANVAS_LIBS - make in demos completes.
-- testsuite not run yet (testing over telnet not possible :P)
-- make install looks fine.

gnome-common 1.2.4
-- everything just fine.
-- no testsuite.

bonobo-activation 0.9.6
-- normal configure
-- compiles fine
-- passes testsuite
-- make install okay.

libbonobo 1.113.0
-- normal configure
-- compiles fine
-- testsuite passes
-- echo sample works.
-- make install fine

gnome-mime-data 1.0.6
-- normal configure
-- compiles fine
-- no testsuite
-- make install fine.

gnome-vfs 1.9.11
-- normal configure
-- compile - dies in test because -lpopt missing for link of
test-async-directory same with test-directory, test-shell, test-symlinks
and test-xfer
-- when -lpopt added - make completes fine.
-- testsuite  - 'no rule to make makeheaders.pl' - file not in
distribution, nothing to make it.
-- make test in test dir - lots of 'test failed: open cancel n: job
never went away'
-- make install just fine.

libgnomeprint 1.112.0
-- standard configure
-- compiles fine
-- no auto testsuite
-- testprint fpe's
-- make install is fine.

also at least compiled - but not documented my progress of - given
number of testsuite failures already, thought it would be good to iron
them out first before I bothered listing all the bugs i have found in
the rest.
audiofile
esound
libgnome
libbonoboui
libgnomeui
gail
at-spi
libgnomeprintui
libgail-gnome
eog
scrollkeeper (this one was a bit messy)
gtkhtml2
yelp
libzvt
libwnk
gnome-core
gnome-control-center
gnome-games (no iagno - freecell crashes on startup, majong on first
move)
gnome-media2
librep
rep-gtk
sawfish
gal
gnumeric
[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]