Built-from-CVS Phoenix isn't responsive
|
|
Thread rating:  |
Chuck Musser - 27 Dec 2003 03:03 GMT I've just pulled the source for Phoenix and built it, but it doesn't quite work. When launched, a browser window appears but I can't make the window active and use it. Clicking on it has no effect--it just sits there. I can't select it with Alt-Tab or Expose either.
My build environment:
-- Mac G5 with 10.3.2 -- Latest developer tools from Apple (Xcode 1.1, gcc-3.3) --Fink 6.2, ORBit and Shared Menus Framework installed per the build instructions
My .mozconfig file:
# sh # Build configuration script # # See http://www.mozilla.org/build/unix.html for build instructions. #
# Options for 'configure' (same as command-line options). ac_add_options --disable-mailnews ac_add_options --disable-composer ac_add_options --disable-ldap ac_add_options --enable-crypto export MOZ_PHOENIX=1 mk_add_options MOZ_PHOENIX=1
When I launch, this is what appears in the console window:
chuck@tricky:~/src/mozilla/dist/bin$ ./MozillaFirebird
Type Manifest File: /Users/chuck/src/mozilla/dist/bin/components/xpti.dat ###!!! ASSERTION: : 'Error', file xptiInterfaceInfoManager.cpp, line 955 Break: at file xptiInterfaceInfoManager.cpp, line 955 +++ JavaScript debugging hooks installed. nsNativeComponentLoader: autoregistering begins. nsNativeComponentLoader: autoregistering succeeded ###!!! ASSERTION: : 'Error', file xptiInterfaceInfoManager.cpp, line 955 Break: at file xptiInterfaceInfoManager.cpp, line 955 nNCL: registering deferred (0) WARNING: Error parsing GRE default preferences. Is this an old-style embedding app?, file nsPrefService.cpp, line 762 WEBSHELL+ = 1 ###!!! ASSERTION: : 'Error', file xptiInterfaceInfoManager.cpp, line 955 Break: at file xptiInterfaceInfoManager.cpp, line 955 For application/x-java-vm found plugin Java Applet.plugin ./MozillaFirebird-bin: can't map file: /Library/Internet Plug-Ins/Java Applet.plugin ((os/kern) invalid argument) [loaded plugin /Library/Internet Plug-Ins/Java Applet.plugin] WARNING: NS_ENSURE_TRUE(NS_SUCCEEDED(rv)) failed, file nsChromeRegistry.cpp, line 3185 WARNING: NS_ENSURE_TRUE(NS_SUCCEEDED(rv)) failed, file nsChromeRegistry.cpp, line 3185 WEBSHELL+ = 2 Note: verifyreflow is disabled WARNING: Failed to get a global Object Owner, file nsJSEnvironment.cpp, line 156 WARNING: Failed to get a global Object Owner, file nsJSEnvironment.cpp, line 156 WARNING: Failed to get a global Object Owner, file nsJSEnvironment.cpp, line 156 XP Popups: This is a nag to indicate that an inconsistent hack is being done on the Mac for popups. WARNING: trying to find screen for sizeless window, using primary monitor, file nsScreenManagerMac.cpp, line 120 Note: styleverifytree is disabled Note: frameverifytree is disabled XP Popups: This is a nag to indicate that an inconsistent hack is being done on the Mac for popups. WARNING: trying to find screen for sizeless window, using primary monitor, file nsScreenManagerMac.cpp, line 120 XP Popups: This is a nag to indicate that an inconsistent hack is being done on the Mac for popups. WARNING: trying to find screen for sizeless window, using primary monitor, file nsScreenManagerMac.cpp, line 120 XP Popups: This is a nag to indicate that an inconsistent hack is being done on the Mac for popups. WARNING: trying to find screen for sizeless window, using primary monitor, file nsScreenManagerMac.cpp, line 120 XP Popups: This is a nag to indicate that an inconsistent hack is being done on the Mac for popups. WARNING: trying to find screen for sizeless window, using primary monitor, file nsScreenManagerMac.cpp, line 120 XP Popups: This is a nag to indicate that an inconsistent hack is being done on the Mac for popups. WARNING: trying to find screen for sizeless window, using primary monitor, file nsScreenManagerMac.cpp, line 120 WEBSHELL+ = 3 CSS Error (chrome://global/skin/progressmeter.css :80.8): Unknown property 'align'. Declaration dropped. WARNING: getting z level of unregistered window, file nsWindowMediator.cpp, line 636 WARNING: getting z level of unregistered window, file nsWindowMediator.cpp, line 636 ###!!! ASSERTION: Cannot support translucent elements with doublebuffering disabled: 'aRCSurface', file nsViewManager.cpp, line 1317 Break: at file nsViewManager.cpp, line 1317 ###!!! ASSERTION: nested lock attempt: '0', file nsDrawingSurfaceMac.cpp, line 158 Break: at file nsDrawingSurfaceMac.cpp, line 158 ###!!! ASSERTION: Cannot support translucent elements with doublebuffering disabled: 'aRCSurface', file nsViewManager.cpp, line 1317 Break: at file nsViewManager.cpp, line 1317 ###!!! ASSERTION: nested lock attempt: '0', file nsDrawingSurfaceMac.cpp, line 158 Break: at file nsDrawingSurfaceMac.cpp, line 158 ###!!! ASSERTION: Cannot support translucent elements with doublebuffering disabled: 'aRCSurface', file nsViewManager.cpp, line 1317 Break: at file nsViewManager.cpp, line 1317 ###!!! ASSERTION: nested lock attempt: '0', file nsDrawingSurfaceMac.cpp, line 158 Break: at file nsDrawingSurfaceMac.cpp, line 158 ###!!! ASSERTION: Cannot support translucent elements with doublebuffering disabled: 'aRCSurface', file nsViewManager.cpp, line 1317 Break: at file nsViewManager.cpp, line 1317 ###!!! ASSERTION: nested lock attempt: '0', file nsDrawingSurfaceMac.cpp, line 158 Break: at file nsDrawingSurfaceMac.cpp, line 158 ###!!! ASSERTION: Cannot support translucent elements with doublebuffering disabled: 'aRCSurface', file nsViewManager.cpp, line 1317 Break: at file nsViewManager.cpp, line 1317 ###!!! ASSERTION: nested lock attempt: '0', file nsDrawingSurfaceMac.cpp, line 158 Break: at file nsDrawingSurfaceMac.cpp, line 158 ###!!! ASSERTION: Cannot support translucent elements with doublebuffering disabled: 'aRCSurface', file nsViewManager.cpp, line 1317 Break: at file nsViewManager.cpp, line 1317 ###!!! ASSERTION: nested lock attempt: '0', file nsDrawingSurfaceMac.cpp, line 158 Break: at file nsDrawingSurfaceMac.cpp, line 158
Any ideas what's going on?
Chuck
Riscky - 28 Dec 2003 21:35 GMT Well a few things... first the build instructions are for suite and not Mozilla Firebird... and you don't need the Shared Menus Framework build...
you might want to check out of the CVS 'mozilla/browser/config/mozconfig' and add it you your config file with '. $topsrcdir/browser/config/mozconfig'
Chuck Musser spoke::
> I've just pulled the source for Phoenix and built it, but it doesn't > quite work. When launched, a browser window appears but I can't make the [quoted text clipped - 31 lines] > > Chuck Chuck Musser - 30 Dec 2003 06:57 GMT Still no luck, although it's good to know about that mozconfig file. Thanks for the tip.
Here's what I did:
Attempt #1: source the referenced mozconfig, as suggested. Rebuild with "make -w -f client.mk" Same startup behavior.
Attempt #2: Threw away the Shared Menus Framework, did a "make clean", then a "make -w -f client.mk" Still the same behavior.
Bummer. Any other ideas?
Chuck
> Well a few things... first the build instructions are for suite and not > Mozilla Firebird... and you don't need the Shared Menus Framework build... > > you might want to check out of the CVS > 'mozilla/browser/config/mozconfig' and add it you your config file with > '. $topsrcdir/browser/config/mozconfig' Riscky - 30 Dec 2003 20:12 GMT Is your tree up-to-date?
If you would like I would be willing to post my build script and mozconfig file for you.
Chuck Musser spoke::
> Still no luck, although it's good to know about that mozconfig file. > Thanks for the tip. [quoted text clipped - 18 lines] >> 'mozilla/browser/config/mozconfig' and add it you your config file >> with '. $topsrcdir/browser/config/mozconfig' Chuck Musser - 31 Dec 2003 01:57 GMT > Is your tree up-to-date? I think so. When I do a "make -w -f client.mk", it seems to do a CVS update. Maybe I need to pull the tree from scratch. Haven't tried that yet.
> If you would like I would be willing to post my build script and > mozconfig file for you. That'd be great. Much appreciated.
Chuck
Jerry Talkington - 31 Dec 2003 02:06 GMT > >Is your tree up-to-date? > > > I think so. When I do a "make -w -f client.mk", it seems to do a CVS > update. Maybe I need to pull the tree from scratch. Haven't tried that yet. That's the best way to do it, just pull mozilla/client.mk mozilla/browser/config/mozconfig
then dump your .mozconfig (which should source mozilla/browser/config/mozconfig,) then run "make -w -f client.mk".
> >If you would like I would be willing to post my build script and > >mozconfig file for you. > That'd be great. Much appreciated. You can also grab the script that I use to build the nightlies: http://tinyurl.com/lnok
 Signature GPG public key: http://pgp.mit.edu:11371/pks/lookup?op=get&search=0x9D5B8762
Chuck Musser - 31 Dec 2003 21:55 GMT > You can also grab the script that I use to build the nightlies: > http://tinyurl.com/lnok I tried your .mozconfig here, but the build failed with a different error. Nonetheless, thanks for the advice. Here's the error:
ld: /usr/lib/crt1.o can't create relocation entry for prebinding (address of section (__DATA,__la_symbol_ptr) more than 24-bits away from first segment, use -noprebind)
In addition there are lots of these messages sprinkled throughout the build. These are warning, not fatal errors:
ld: warning prebinding not disabled because (__LINKEDIT segment (address = 0xc8000 size = 0xf4c30) of @executable_path/libsoftokn3.dylib overlaps with __LINKEDIT segment (address = 0x2f000 size = 0xd1000) of @executable_path/libxpcom_compat.dylib
I tried a couple different things, adding the --disable-prebinding flag being the most significant. No luck. Sigh, this is getting a bit arcane for me. I should probably repeat that I'm on a G5 running 10.3.2; not sure if that's what's complicating things.
Perhaps I need to take a different tack. My goal was to isolate bugzilla bug #228479, a crash that happens only on my Mac and not on any of several Linux boxes I have lying around. The idea was to run MozillaFirebird with the --debug flag and use GDB to help me pinpoint the error. Trouble is, I need a Mozilla with debugging symbols and the nightlies don't seem to have them.
Can anyone point to a recent build (has to be later than Firebird 0.7) that has the debugging symbols?
Chuck
Jerry Talkington - 31 Dec 2003 22:16 GMT > >You can also grab the script that I use to build the nightlies: > >http://tinyurl.com/lnok [quoted text clipped - 13 lines] > with __LINKEDIT segment (address = 0x2f000 size = 0xd1000) of > @executable_path/libxpcom_compat.dylib I actually just changed the script today so that it will fix these. It only seems to be a problem when not building it optimized. Making sure that LD_PREBIND_ALLOW_OVERLAP is not set seems to do the trick.
> I tried a couple different things, adding the --disable-prebinding flag > being the most significant. No luck. Sigh, this is getting a bit arcane > for me. I should probably repeat that I'm on a G5 running 10.3.2; not > sure if that's what's complicating things. I just update my site with the new build script and new .mozconfigs (an new one for debugs, since -O2 and -g3 don't go well together.)
> Perhaps I need to take a different tack. My goal was to isolate bugzilla > bug #228479, a crash that happens only on my Mac and not on any of [quoted text clipped - 5 lines] > Can anyone point to a recent build (has to be later than Firebird 0.7) > that has the debugging symbols? There really isn't an official place for these on the mozilla FTP site, and I just started building them for myself after the nightlies so I'd have a debug build. Since I started building them specifically to knock out some of the critical bugs, I'll take a look at #228479.
 Signature GPG public key: http://pgp.mit.edu:11371/pks/lookup?op=get&search=0x9D5B8762
Chuck Musser - 02 Jan 2004 20:55 GMT > I actually just changed the script today so that it will fix these. It
> only seems to be a problem when not building it optimized. Making sure > that LD_PREBIND_ALLOW_OVERLAP is not set seems to do the trick. OK, I finally built a running and debuggable Firebird. A couple things contributed to this glorious achievement:
1.) I used your mozconfig.debug, which ensured (via --enable-optimize="-g3" that debugging symbols were included.
2.) I took note of the above hint concerning LD_PREBIND_ALLOW_OVERLAP and discovered something about Fink. The version I use (Fink 0.6.2) includes a shell initialization script that sets this variable to "1". If you source this init script in your own shell initialization (and Fink's instructions tell you to do so) the variable will be set and the compile will fail. Once I unset the var, and rebuilt, things worked fine.
As a side note, "mozmake" doesn't explicitly unset this variable for debug builds. You might want to do that to robustify things in environments where Fink (or something else) sneaks it into the mix.
Thanks for all the help,
Chuck
Jerry Talkington - 03 Jan 2004 00:02 GMT > > I actually just changed the script today so that it will fix these. It > >only seems to be a problem when not building it optimized. Making sure [quoted text clipped - 12 lines] > Fink's instructions tell you to do so) the variable will be set and the > compile will fail. Once I unset the var, and rebuilt, things worked fine. Done. It will go up (with a couple of changes that I made to make the script work on Linux) tonight.
 Signature GPG public key: http://pgp.mit.edu:11371/pks/lookup?op=get&search=0x9D5B8762
Robert Mohr - 31 Dec 2003 02:34 GMT >> Is your tree up-to-date? >> > I think so. When I do a "make -w -f client.mk", it seems to do a CVS > update. Maybe I need to pull the tree from scratch. Haven't tried that yet. 'make -f client.mk' ('-w' doesn't have much effect anyway) will do something analagous to 'cvs update' and then build. 'make -f client.mk checkout' is preferred over 'cvs update', but does essentially the same thing. 'make -f client.mk build' builds what you have downloaded without looking for updates.
I haven't even gotten something from source to compile though :(.
 Signature Robert Mohr mohr.42@osu.edu mohrr@cis.ohio-state.edu
<http://www.mozilla.org/wishlist-faq.html> <http://extensionroom.mozdev.org/> <http://texturizer.net/firebird/> <http://www.mozilla.org/community-etiquette.html>
|
|
|