T2 IRC Log: 2007-04-08

This is the log as captured by an IRC bot in the channel. The statements are those of the individual people and might not neccessarily reflect the policy and legal rules as set forth by the T2 SDE Project.

« prev | next »

--- Log opened Sun Apr 08 00:00:07 2007
00:13 < AStorm> oh, btw, gcc4.2.0 fixed the version_specific_runtime_libs logic
00:14 < AStorm> but --with-gxx-include-dir is much nicer anyway :>
00:16 < AStorm> and the $(version) would be kept in the gcc.desc or .conf anyway
00:24 < AStorm> drat, I have to fix the uclibc patch
00:24 < AStorm> or at least look through it
00:35 < AStorm> they've fixed dynamic linking with uclibc..
00:35 < AStorm> but I don't see the size reductor :|
00:39 < AStorm> at least they've debugged the support :>
00:43 < AStorm> heh, you can even choose the dynamic linker on the fly :P
00:43 < AStorm> (but the defaults are sane, set in specs)
00:46 < AStorm> Great even, they macroised everything
00:56 < AStorm> Hmm, the patch was a tiny bit wrong too :P
00:56 < AStorm> missing -
00:56 < AStorm> in boehm-gc configure patch
00:57 < AStorm> No, it's correct, blah.
01:12 < AStorm> Guess I'll do all the hard work of porting gcc to 4.2 for you :P
01:17 -!- idealm [n=ideal@c58-107-18-60.belrs2.nsw.optusnet.com.au] has joined #t2
01:21 -!- emte_ [n=emte@d64-180-45-14.bchsia.telus.net] has quit [Read error: 110 (Connection timed out)]
01:21 -!- emte_ [n=emte@d64-180-45-14.bchsia.telus.net] has joined #t2
01:33 < AStorm> I wonder what is the second part of the fastmath patch for
01:33 < AStorm> It's not needed.
01:33 < AStorm> The other part is in 4.2.0
02:04 < AStorm> Another build try...
02:38 < AStorm> Ok, now it's building
02:39 < AStorm> Uhoh
02:40 < AStorm> checking for library containing strerror... configure: error: Link tests are not
02:40 < AStorm> allowed after GCC_NO_EXECUTABLES.
02:40 < AStorm> I need a newer, less broken snapshot :>
02:41 < AStorm> Actually, it's some target which shouldn't be built probably
02:47 < AStorm> WTF?!?
02:47 < AStorm> /t2-trunk/build/default-7.0-trunk-multipen-x86-64-cross/TOOLCHAIN/tools.cross/us
02:47 < AStorm> r/x86_64-t2-linux-uclibc/bin/ld: crt1.o: No such file: No such file or directory
02:47 < AStorm> collect2: ld returned 1 exit status
02:48 < AStorm> I thought the added uclibc scan would take care of all these
02:50 < AStorm> I guess I was wrong and I have to hunt for some more places, where t-linux-uclibc rule can be skipped over
02:51 < AStorm> libgomp fails
02:54 < AStorm> libiberty fails its configure too
02:54 < AStorm> (due to NO_EXECUTABLES check)
02:55 < AStorm> strerror test isn't skipped - will have to paper over it
02:56 < AStorm> but that... later today
04:01 -!- idealm [n=ideal@c58-107-18-60.belrs2.nsw.optusnet.com.au] has quit [Remote closed the connection]
05:35 -!- emte_ [n=emte@d64-180-45-14.bchsia.telus.net] has quit [Read error: 60 (Operation timed out)]
05:45 -!- emte_ [n=emte@d64-180-45-14.bchsia.telus.net] has joined #t2
08:53 -!- emte__ [n=emte@d64-180-45-14.bchsia.telus.net] has joined #t2
08:54 -!- emte_ [n=emte@d64-180-45-14.bchsia.telus.net] has quit [Read error: 110 (Connection timed out)]
09:08 -!- idealm [n=ideal@c58-107-18-60.belrs2.nsw.optusnet.com.au] has joined #t2
09:19 < sepp> moin :)
09:24 -!- emte__ [n=emte@d64-180-45-14.bchsia.telus.net] has quit [Read error: 110 (Connection timed out)]
11:26 < rxr> re
11:26 < rxr> moin moin sepp
11:32 < sepp> hi rxr :)
11:40 < CIA-18> rene * r23462 /trunk/package/editors/nano/nano.desc: * updated nano (2.0.3 -> 2.0.4)
11:56 < CIA-18> rene * r23463 /trunk/package/audio/ (2 files in 2 dirs):
11:56 < CIA-18> * updated audacious (1.3.1 -> 1.3.2)
11:56 < CIA-18> * updated audacious-plugins (1.3.1 -> 1.3.2)
11:57 < CIA-18> rene * r23464 /trunk/package/develop/hte/hte.desc: * updated hte (2.0.1 -> 2.0.2)
11:57 < CIA-18> rene * r23465 /trunk/package/network/ndiswrapper/ndiswrapper.desc: * updated ndiswrapper (1.40 -> 1.41)
11:59 < CIA-18> rene * r23466 /trunk/package/network/aoetools/aoetools.desc: * updated aoetools (14 -> 15)
11:59 < CIA-18> rene * r23467 /trunk/package/network/pound/pound.desc: * updated pound (2.2.7 -> 2.2.8)
11:59 < CIA-18> rene * r23468 /trunk/package/www/mozplugger/mozplugger.desc: * updated mozplugger (1.7.3 -> 1.8.1)
12:12 < CIA-18> rene * r23469 /trunk/package/kde/kmymoney/ (kmymoney.conf kmymoney.desc uic.patch): * updated kmymoney (0.8.1 -> 0.8.5)
12:23 < CIA-18> rene * r23470 /trunk/package/audio/nas/ (X11R7.patch nas.desc): * updated nas (1.8 -> 1.9)
12:43 -!- sepp_ [n=sepp@p85.212.69.212.tisdip.tiscali.de] has joined #t2
12:52 -!- sepp [n=sepp@p213.54.7.213.tisdip.tiscali.de] has quit [Read error: 60 (Operation timed out)]
12:57 [Users #t2]
12:57 [@ChanServ] [ fke ] [ R4gnar0k] [ sparc-kly] [ TobiX ]
12:57 [ AStorm ] [ idealm] [ rxr ] [ Stelz ] [ valentin]
12:57 [ CIA-18 ] [ LMJ ] [ sepp_ ] [ th ]
12:57 -!- Irssi: #t2: Total of 14 nicks [1 ops, 0 halfops, 0 voices, 13 normal]
13:00 -!- idealm [n=ideal@c58-107-18-60.belrs2.nsw.optusnet.com.au] has quit [Remote closed the connection]
13:02 -!- idealm [n=ideal@c58-107-18-60.belrs2.nsw.optusnet.com.au] has joined #t2
13:23 -!- idealm [n=ideal@c58-107-18-60.belrs2.nsw.optusnet.com.au] has quit ["Leaving"]
13:24 -!- idealm [n=ideal@c58-107-18-60.belrs2.nsw.optusnet.com.au] has joined #t2
13:28 < AStorm> Hey guys.
13:28 < AStorm> Could one of you help me with that one patch against that NO_EXECUTABLE bug?
13:28 < sepp_> hey AStorm
13:28 < AStorm> That's probably the last thing missing.
13:30 < sepp_> no_clue :/
13:30 < AStorm> Hmm.
13:30 < rxr> AStorm: patch url or quote?
13:30 < AStorm> Ok, I'll give you the quote
13:31 < AStorm> 2 errors, actually
13:33 < AStorm> I've already ported the uclibc patches
13:39 < AStorm> http://pastebin.ca/429707
13:40 < AStorm> That "C compiler cannot create executables"
13:40 < AStorm> I'll paste the log for that
13:42 < AStorm> I'm probably missing some patches
13:46 < AStorm> config.log:
13:46 < AStorm> http://pastebin.ca/429708
13:46 < AStorm> crt1.o is missing
13:46 < AStorm> Some uclibc incompatibility I have to fix :P
13:47 < AStorm> the NO_EXECUTABLES is the build system bug :P
13:48 < AStorm> If you want the ported patches, I can sent these to you
13:52 < rxr> re
13:53 < rxr> do you have patches for the above issues ?
13:53 < rxr> just send them, yes.
14:00 < AStorm> For these, not
14:00 < AStorm> other uclibc were ported ok
14:00 < AStorm> wait a while, I'll google a bit more
14:01 < AStorm> Oh, busybox has them ]:->
14:02 < AStorm> and I spent 3h forward-porting the patches.... damn!
14:03 < AStorm> Let's check these...
14:06 < rxr> did you do something special?
14:06 < rxr> as usually uclibc and dietlibc fluff builds in T2 just fine
14:07 < AStorm> http://buildroot.uclibc.org/cgi-bin/viewcvs.cgi/trunk/buildroot/toolchain/gcc/4.2/#dirlist
14:08 < AStorm> I mean 4.2.0
14:08 < AStorm> 4.2.0 prerelease 20070316
14:08 < AStorm> Soon to be out :>
14:10 < AStorm> (less than 6 regressions to fix :> )
14:11 < AStorm> It'll just be better than 4.1.2 :P
14:17 < AStorm> try 2 :>
14:17 < AStorm> with proper buildroot patches
14:25 < AStorm> ok, it's... building!
14:26 < AStorm> Blah, the strerror is still broken :P
14:28 < AStorm> I'll mutate the config.cache for the time
14:28 < AStorm> this should be reported to gcc devs... or maybe it's on their bugzilla
14:30 < AStorm> PR28949
14:30 < rxr> *click*
14:31 < AStorm> http://gcc.gnu.org/PR28949
14:31 < AStorm> :-)
14:31 < AStorm> I'll paper over it with config.cache mod
14:34 < AStorm> how would I get the objdir off gcc?
14:34 < AStorm> Aha, I see.
14:35 < AStorm> and the $CTARGET?
14:37 < AStorm> got it: ${arch_target}
14:39 < rxr> yeah, I'm working on the upcomming XML based handbook for T2
14:39 < rxr> do not watch the IRC windows all the time
14:39 < rxr> off for "lunch" in some minutes, anway
14:39 < rxr> thw new handbook looks sooooo much beter than the former LaTeX based one, especially the HTML output is awesome ...
14:42 < AStorm> :-)
14:42 < AStorm> I usually use Asciidoc for documentation
14:42 < AStorm> looks nice even in notepad :P
14:42 < AStorm> convertible to nice html, docbook, latex and xml
14:45 < AStorm> ok, my hack seems to have worked
14:47 < rxr> yeah - but this is a 160 page book that goes into print this summer ...
14:47 < rxr> 160 pages, now - about the same to be written the next 2 months ...
14:47 < AStorm> :-)
14:47 < AStorm> Niiiiice.
14:48 < AStorm> multiple-compiler stuff is easy to fix when you have that ABI/TARGET separation
14:49 < AStorm> I'll experiment with eprefix modification :>
14:49 < AStorm> (once the base system builds :> )
14:50 < AStorm> Damn.
14:50 < AStorm> More like these :>
14:50 < AStorm> asprintf...
14:51 < AStorm> And it takes like 5 minutes for a rehash (stupid autoconf)
14:51 < AStorm> All those link tests have to be patched off, anyway
14:51 < AStorm> in the cross0 stage.
14:51 < AStorm> Will do.
14:52 < rxr> nice to see you that motivated :-)
14:52 < rxr> cu - I'm off, cu then
14:53 < AStorm> I just have a broken os :P
14:53 < AStorm> have to get something nice up and running
14:53 < AStorm> and it has to be suitable to build that pendrive system
14:53 < AStorm> and I'll have to get 4.2.0 running anyway
14:55 < AStorm> Hmm, probably all checkfuncs have to be switched off...
14:56 < AStorm> I wonder if all are available in uclibc
14:56 < AStorm> (they are in glibc)
14:57 < AStorm> Hmm, I'll just hack the test senseless - not to fail
15:00 < CIA-18> rene * r23471 /trunk/package/gnome2/ (6 files in 2 dirs): * renamed gaim pidgin
15:02 < CIA-18> rene * r23472 /trunk/package/gnome2/pidgin/pidgin.desc: * further gaim^W pidgin rebranding
15:03 < AStorm> ?
15:03 < AStorm> pidgin?
15:06 < AStorm> Stupid AOL.
15:10 < AStorm> yet another try...
15:22 < Stelz> yay
15:38 -!- idealm [n=ideal@c58-107-18-60.belrs2.nsw.optusnet.com.au] has quit [Remote closed the connection]
16:08 < AStorm> hmm...
16:08 < AStorm> after patching the issue I'm having a large problem...
16:09 < AStorm> it seems to configure ok, yet still the configure fails.
16:10 < AStorm> after "executing default-1 commands"
16:10 < AStorm> Or maybe... blah
16:10 < AStorm> # is not the comment mark in autoconf
16:10 < AStorm> stupid m4
16:11 < AStorm> Blah, I can't patch :P
16:20 -!- Stelz [i=stelz@unaffiliated/stelz] has quit ["I wanna live my life like thunder!"]
16:34 < AStorm> I still need help with that patch
16:34 < AStorm> Somehow correct patch is rejected.
16:35 < AStorm> Ah, I know. Blah again :>
16:41 < AStorm> Maybe this time it's right.
16:41 < AStorm> My own fork of T2 should be called T-inf :P
16:45 < AStorm> Blah.
16:45 < AStorm> Didn't fix it, forgot to hack .ac :P
16:57 -!- Stelz [i=stelz@unaffiliated/stelz] has joined #t2
16:57 -!- Stelz [i=stelz@unaffiliated/stelz] has quit [Remote closed the connection]
16:57 -!- Stelz [i=stelz@unaffiliated/stelz] has joined #t2
17:05 -!- Stelz [i=stelz@unaffiliated/stelz] has quit ["I wanna live my life like thunder!"]
17:07 -!- Stelz [i=stelz@unaffiliated/stelz] has joined #t2
17:16 < sepp_> can i change the gtk file selector back to unix behaviour?
17:17 < sepp_> global
17:20 < sepp_> i mean uppercase dirs/files 1st ...
18:05 -!- fke_ [n=fke@p3e9e3778.dip0.t-ipconnect.de] has joined #t2
18:14 -!- fke [n=fke@p3E9E015D.dip0.t-ipconnect.de] has quit [Read error: 110 (Connection timed out)]
18:23 < AStorm> sepp_: not w/o patching
18:27 < AStorm> Damn, I can't find where to disable these tests :/
18:28 < AStorm> The best would be to just skip building libiberty altogether on stage0 cross
18:28 < AStorm> :>
18:28 < AStorm> it's not needed anyway
18:31 < AStorm> simple, just added it to skipdirs
18:31 < AStorm> Maybe that will work.
18:33 < AStorm> gcc guys should add --disable-targets=
18:41 < AStorm> damn
18:42 < AStorm> it tries linking in some ld check too
18:42 < AStorm> against -Wl,--gc-sections. This one I'll override using config.cache
18:43 < AStorm> Please go to their bugzilla and pledge to fix all that :P
18:44 < AStorm> (as people using cross builds extensively)
18:47 < AStorm> maybe they'll fix that before the release then
18:54 < AStorm> Uh, somehow the skipdirs weren't honored, the error is in target-libstdc++-v3
18:55 < AStorm> Ah, you didn't fix configure.in :>
19:08 < AStorm> Hmm, I still need to fix libgomp failure
19:08 < AStorm> any clues?
19:08 < AStorm> for now, i'll disable it
19:19 < AStorm> Yeah! It finally built!
19:19 < AStorm> Now, let's wait for stage1
19:24 < AStorm> When that one passes, I'll send you the change
19:24 < AStorm> for later testing
19:33 -!- mqueiros [n=mqueiros@c-217-70-65-169.bragatel.pt] has joined #t2
19:37 < AStorm> Good, fixed debug patch also works with uclibc :-)
19:40 < AStorm> Uhoh...
19:41 < AStorm> it seems something is spinning...
19:41 < AStorm> x86_64-t2-linux-uclibc-objcopy --add-gnu-debuglink=/t2-trunk/build/default-7.0-trunk-multipen-x86-64-cross//opt/debug/usr/lib/libc.a.debug usr/lib/libc.a
19:41 < AStorm> this thing, against static uclibc...
19:41 -!- mqueiros [n=mqueiros@c-217-70-65-169.bragatel.pt] has quit ["Ex-Chat"]
19:42 < AStorm> Sounds like a binutils bug :P
19:42 < AStorm> Well, it is eating memory...
19:43 < AStorm> Well, it did finish, but something else broke
19:43 < AStorm> CC utils/readelf
19:43 < AStorm> /t2-trunk/build/default-7.0-trunk-multipen-x86-64-cross/TOOLCHAIN/tools.cross/cr
19:43 < AStorm> osscc/../lib/gcc/x86_64-t2-linux-uclibc/4.2.0/../../../../x86_64-t2-linux-uclibc
19:43 < AStorm> /bin/ld: Scrt1.o: No such file: No such file or directory
19:43 < AStorm> Hmm...
19:46 < AStorm> I guess -static doesn't work correctly with these binutils
19:46 < AStorm> maybe some spec has to be changed
19:48 < AStorm> yep
19:48 < AStorm> *startfile:
19:48 < AStorm> %{!shared: %{pg|p|profile:gcrt1.o%s;pie:Scrt1.o%s;:crt1.o%s}} crti.o%s %{stat
19:48 < AStorm> ic:crtbeginT.o%s;shared|pie:crtbeginS.o%s;:crtbegin.o%s}
19:48 < AStorm> Ahh... something doesn't like PIE :P
19:48 < AStorm> Probably uclibc :P
19:50 < AStorm> But anyway, I think uclibc doesn't have any of these
19:50 < AStorm> At least when statically built
20:07 < AStorm> yep, uclibc is incompatible with PIE
20:07 < AStorm> add a quirk, please
20:08 < AStorm> Now on perl...
20:09 -!- mqueiros [n=mqueiros@c-217-70-65-169.bragatel.pt] has joined #t2
20:16 < AStorm> perl failed, let's see
20:16 < AStorm> /bin/ld: cannot find -ldl
20:16 < AStorm> :>
20:16 < AStorm> Why the hell would it want that?
20:17 < AStorm> It's statically linked... ah
20:17 < AStorm> this pile of junk is using system gcc for configure tests :P
20:17 < AStorm> (and system libc)
20:17 < AStorm> Please fix.
21:26 < rxr> short rehi
21:26 < rxr> (via UMTS)
21:26 < rxr> sepp - maybe the sorting depends on your locale ?
21:27 < sepp_> i dont know, somehow it changed but i can not google why :/
21:27 < rxr> locale often alter sorting of non alphanum letters
21:27 < rxr> so - handbook update committed, /me turning UMTS modem off - cu then
21:29 < sepp_> ohh, that is it, with C it sorts the-right-way ...
21:29 < sepp_> that is stupid
21:30 < sepp_> with en_US a bit silly
21:30 < sepp_> thanks :)
21:30 < sepp_> i looked if there is soemthing in gtkrc
21:31 < sepp_> cu
21:33 < sepp_> that is some dumb behaviour, i changed it to en_ because some other crap protesting all the time about C
21:34 < AStorm> sepp_: any idea on that "cannot find -ldl" with uclibc static?
21:34 < AStorm> in perl
21:35 < sepp_> no, i never used uclibc, why would it link -ldl in with a static one?
21:35 < AStorm> Weird, yes.
21:35 < AStorm> dynamic loader should be in the libc
21:35 < AStorm> perl should detect that and not link dl in
21:36 < sepp_> maybe just remove that from the gcc wrapper
21:37 < AStorm> Hmm?
21:38 < AStorm> It's there? :>
21:38 < AStorm> Well, uclibc 0.9.9 has libdl.a
21:38 < AStorm> (at least in PLD)
21:39 < AStorm> I wonder if that svn revision is newer than that
21:40 < sepp_> it is just some stupid makefile with -ldl even with a static libc?
21:41 < AStorm> Probably yes
21:41 < AStorm> it's perl - anything goes there :P
21:47 < AStorm> Ok, now on to hunting the patch
21:47 < AStorm> I wonder when to apply it and how
21:48 < AStorm> Yep, it's hardcoded in the makefiles.
21:54 < AStorm> Well...
21:55 < AStorm> How would I apply a patch/sed only on uclibc build when that uclibc is static?
21:55 < AStorm> (not just on SDECFG_STATIC - that can be changed)
21:58 < sepp_> you can name it .diff and apply it in the .conf with some condition
21:58 < sepp_> there are some .diffs already ...
21:59 < AStorm> Yes, I'm wondering how "some condition" should look
22:00 < sepp_> dont know :)
22:00 < AStorm> if && ; then...
22:00 < AStorm> the first one is simple SDECFG_LIBC check
22:00 < AStorm> but the second one...
22:03 < AStorm> Temporarily, I'm checking against SDECFG_LIBC and SDECFG_STATIC
22:04 < AStorm> There _will_ be a lot of patches once I get a working system - be warned
22:08 < AStorm> Uhoh
22:08 < AStorm> static uclibc just doesn't have a dynamic loader
22:08 < AStorm> I guess I want SDECFG_STATIC reset for it
22:08 < AStorm> At least no --enable-static
22:10 < AStorm> Rebuilding it in stage1 would be done using Emerge, right?
22:12 < AStorm> Unsetting HAVE_SHARED is unwise
22:12 < AStorm> I wonder why anyone would want that.
22:12 -!- sepp [n=sepp@p85.212.42.70.tisdip.tiscali.de] has joined #t2
22:14 -!- emte [n=emte@d64-180-45-14.bchsia.telus.net] has joined #t2
22:15 < AStorm> Well, that HAVE_SHARED should only be disabled on some other option, maybe SDECFG_UCLIBC_MINIMAL
22:16 < AStorm> Because having a static system is different than having no dynamic loader :>
22:16 < AStorm> Or, maybe perl should be built w/o DynaLoader
22:16 < AStorm> I'm not sure.
22:18 < AStorm> rxr: how would I disable SDECFG_STATIC only on uclibc?
22:30 -!- sepp_ [n=sepp@p85.212.69.212.tisdip.tiscali.de] has quit [Read error: 110 (Connection timed out)]
22:30 < AStorm> Ok, I should just use Build-Pkg, right?
22:54 < sepp> for what?
--- Log closed Mon Apr 09 00:00:04 2007