jbailey | lamont__: What are you rambling about? | 12:18 |
---|---|---|
lamont__ | jbailey: doing the hppa archive-event | 12:19 |
lamont__ | hence rebootstrapping gcc-* | 12:19 |
jbailey | Has the fix for hppa-hpux gone into 4.0.1 now? | 12:19 |
lamont__ | dunno | 12:20 |
lamont__ | this is g{cc,++}-3.4 | 12:20 |
jbailey | Yeah, but for the rest of the archive, it would probably make sense to wait for that. | 12:20 |
lamont__ | that is, gcc-defaults points to 3.4 on hppa now. (1.24) | 12:20 |
jbailey | Err, just because of the function pointer bit? | 12:21 |
lamont__ | yes | 12:21 |
jbailey | Or is this to get past the everything-sucks-on-4.0 stage? | 12:21 |
lamont__ | just because pinters are completely b0rked | 12:21 |
lamont__ | pointers, even | 12:21 |
lamont__ | the everything-sucks thing is just a perk | 12:21 |
lamont__ | the pointer bug in 4.0 renders every .deb suspect | 12:22 |
jbailey | In another day or two that bug will be fixed on 4.0 | 12:22 |
jbailey | http://gcc.gnu.org/ml/gcc/2005-06/msg01192.html | 12:23 |
lamont__ | well then I'd better hurry and get through the sucky stage.. :) | 12:23 |
jbailey | Yup. =) | 12:24 |
lamont__ | or at least put the gcc-default binaries on hold. :-) | 12:24 |
jbailey | That would probably be better. =) | 12:25 |
lamont__ | lol | 12:25 |
lamont__ | binutils built. now for glibc | 12:26 |
jbailey | lamont__: Have comendered(sp?) a nice 8-way PA system? =) | 12:27 |
lamont__ | _2_ hours. sigh | 12:27 |
lamont__ | I have an A500 that I'll be commandeering tomorrow | 12:27 |
lamont__ | (likewise, sp??) | 12:27 |
lamont__ | 2-way, 8GB RAM | 12:28 |
jbailey | You seem to have it right. | 12:28 |
lamont__ | sadly, the buildd's run UP kernels, since they work better... | 12:28 |
jbailey | I thought recent PA kernels were supposed to behave better. | 12:28 |
=== jbailey shuts down spare machines in prep for the rolling blackouts. | ||
lamont__ | jbailey: they're much better, but.. :-( | 12:31 |
jbailey | Is there risk of memory corruption leading to wrong-code generation, or just hangs and crashes? | 12:31 |
lamont__ | it tends to generate random segv's, from what I've seen | 12:37 |
lamont__ | but only under semi-heavy memory-pressure | 12:38 |
lamont__ | does -1ubuntu7 (glibc) have hppa tests disabled? | 12:38 |
jbailey | lamont__: Looks like, yes. | 12:39 |
jbailey | I haven't done an -1u8 yet since we asked me to reenable it. | 12:39 |
lamont__ | kewl... will build faster. :-) | 12:39 |
lamont__ | Total 2 package(s) in state Building. | 12:40 |
lamont__ | Total 5 package(s) in state Installed. | 12:40 |
lamont__ | Total 6091 package(s) in state Needs-Build. | 12:40 |
lamont__ | :-( | 12:40 |
doko | lamont__, no need for doxygen if you build with -b -B | 12:46 |
lamont__ | doko: ah, coolness. | 12:47 |
lamont__ | but it builds fast. :-) | 12:47 |
lamont__ | s/builds/built/ | 12:47 |
=== jbailey wanders off for a little bit. | ||
lamont__ | doko: with g++-3.4, nothing should need libstdc++5 any more, correct? | 01:20 |
doko | lamont__: correct | 01:22 |
lamont__ | ok. I'll just make it hard to upload anything that Depends: libstdc++5 :-) | 01:24 |
doko | well, the stuff that needs libstdc++5, doesn't build on hppa. like OOo1 | 01:27 |
lamont__ | hehe | 01:31 |
lamont__ | and let us not forget gcc-3.3 :-) | 01:32 |
lamont__ | basically, I'll just make the automated uploader bounce anything that mentions libstdc++5, and then manually figure out what it really means | 01:32 |
lamont__ | some of that can get uploaded, the bulk just needs to be built in the right order | 01:32 |
doko | with the last batch of uploads I dropped the libstdc++5 rdepends from 880 to 530. at least on i386. | 01:33 |
lamont__ | nice | 01:34 |
lamont__ | this is more just me trying to make sure that I'm not accidentally building out-of-order. | 01:34 |
lamont__ | glibc is in install target. coolness | 01:36 |
lamont__ | and then comes 24 hours to build all 3 versions of gcc :( | 01:38 |
lamont__ | doko: what's the minimum set of gcc-* to build? 3.4 and 4.0? (do I really need the current gcc-3.3, or can I let the one from hoary [3.3.5-8ubuntu2] get used?) | 01:39 |
doko | lamont__: the one from hoary should be ok, although I don't plan any further gcc-3.3 uploads now. 3.3 is currently synced from Debian and I don't plan to update that one for a while. the next 3.3 uploads will just drop support for languages like java, treelang, objc, | 01:41 |
lamont__ | ok. | 01:42 |
lamont__ | I'll let gcc-3.3 eventually catch up, then | 01:43 |
lamont__ | X.o.r.g.................................. | 01:55 |
doko | I'm innocent :-) | 01:57 |
lamont__ | you build-depend on it. That's enough. :-( | 01:57 |
=== lamont__ tries ignoring breezy's arch: all packages for the momenbt | ||
doko | lamont__, for a bootstrap, you can disable java and ignore all the X b-d's | 02:00 |
lamont__ | pb is that I want to actually build the real source. | 02:01 |
lamont__ | after all, I have hoary/main | 02:03 |
=== lamont__ cheers - gcc-3.4 is actually unpacking source to build... build-deps met | ||
lamont__ | with luck, it'll even build | 02:24 |
lamont__ | doko: would be nice if you hollered before uploading gcc-* this week - I'd like to let these builds finish and upload before I have to start new ones. | 02:38 |
=== warthylog [~warthylog@port49.ds1-van.adsl.cybercity.dk] has joined #ubuntu-toolchain | ||
=== Topic for #ubuntu-toolchain: GNU Compiler Collection, Glibc, Binutils, Linux-kernel-headers | GLIBC Todo: hppa, sparc NPTL, i386 biarch, C++ ABI change: 33/55 library packages in the archives | ||
=== Topic (#ubuntu-toolchain): set by doko at Thu May 19 00:40:28 2005 | ||
=== fabbione [~fabbione@port49.ds1-van.adsl.cybercity.dk] has joined #ubuntu-toolchain | ||
fabbione | morning | 06:13 |
lamont | Total 6 package(s) in state Installed. | 06:21 |
lamont | Total 6091 package(s) in state Needs-Build. | 06:21 |
lamont | fabbione: I think sparc is ahead. :-( | 06:21 |
fabbione | lamont: doh! | 06:21 |
fabbione | i guess you are restarting with gcc-3.4? | 06:21 |
lamont | yeah | 06:22 |
fabbione | well actually sparc has a bigger problem | 06:22 |
lamont | saddest part of it is that 33% of that 6 packages Depend: libstdc++5 | 06:22 |
fabbione | breezy libc6 is triggering a bug in gcc | 06:22 |
fabbione | and the kernel miscompile | 06:22 |
lamont | (dpkg and doxygen) - I'll need to rebuild them eventually, but there's a new dpkg in debian (albeit FTBFS) | 06:23 |
lamont | ew. poor sparc | 06:23 |
fabbione | so sparc has 3/4 of the archive.. but no kernel :) | 06:23 |
lamont | hppa has almost .1% of the archive. :-) | 06:23 |
fabbione | how are you going to cope with the g++ transition? | 06:23 |
lamont | g++-3.4 is libstdc++6 | 06:23 |
fabbione | will you do it later? or are you swutching only gcc- ? | 06:23 |
lamont | so I get the transition | 06:23 |
fabbione | ah ok | 06:24 |
lamont | the later switch from 3.4 to 4.0 is not an abi event | 06:24 |
=== lamont is most happy that 3.4 exists | ||
fabbione | yeah that's right | 06:24 |
lamont | but must get to bed | 06:24 |
fabbione | ehhe | 06:24 |
lamont | with luck, I'll have 3.4 and 4.0 by morning, and can semi-unleash the beast. | 06:24 |
fabbione | it sounds like i am the "GO TO BED" alarm clock in here ;) | 06:25 |
lamont | yeah - that's the one... :-) | 06:25 |
fabbione | ahah | 06:25 |
fabbione | good night lamont | 06:25 |
lamont | the other thing I'd like to get is a list of all the packages that I need to build before I can build xorg, so that I can queue those up | 06:25 |
fabbione | i think it will just go DepWait now | 06:26 |
fabbione | so it shouldn't be too difficult | 06:26 |
fabbione | but take into account daniels is going to upload/if he didn't already | 06:26 |
fabbione | some extra crack soon | 06:26 |
fabbione | (i can't check.. my server crashed and secondary MX flushing) | 06:26 |
lamont | fabbione: no, the issue is that things Conflict with older packages from xorg, so it's an ugly failure, not a pretty auto-depwaited thing | 06:31 |
fabbione | oh | 06:31 |
fabbione | crap | 06:31 |
lamont | fabbione: but I got the list from daniels, so life is better. | 06:33 |
fabbione | ehhe | 06:33 |
=== Seveas [~seveas@seveas.demon.nl] has joined #ubuntu-toolchain | ||
=== chmj [~chmj@196.36.161.235] has joined #ubuntu-toolchain | ||
=== elmo [~james@83-216-141-215.jamest298.adsl.metronet.co.uk] has joined #ubuntu-toolchain | ||
=== Riddell [jr@jriddell.kde] has joined #ubuntu-toolchain | ||
=== daniels [~daniels@amnesiac.heapspace.net] has joined #ubuntu-toolchain | ||
=== ajmitch [~ajmitch@port163-214.ubs.maxnet.co.nz] has joined #ubuntu-toolchain | ||
=== Mithrandir [~tfheen@vawad.err.no] has joined #ubuntu-toolchain | ||
=== svenl [~luther@AStrasbourg-251-1-13-61.w82-126.abo.wanadoo.fr] has joined #ubuntu-toolchain | ||
=== \sh [~shermann@server3.servereyes.de] has joined #ubuntu-toolchain | ||
elmo | * glibc_2.3.5-1ubuntu7 builds: | 10:54 |
elmo | [...] | 10:54 |
elmo | but no longer builds: | 10:54 |
elmo | o 2.3.5-1ubuntu4: libc6-sparcv9 | 10:54 |
elmo | fabbione: deliberate? | 10:54 |
fabbione | elmo: yeps | 10:54 |
fabbione | we are killing sparc32 support | 10:55 |
fabbione | sparc32 is going the same way as m68k | 10:55 |
fabbione | doorstep use basically | 10:55 |
fabbione | elmo: did you have any time to setup logs@ ? | 10:56 |
elmo | meh, no, let me try and do that now | 10:57 |
fabbione | elmo: ok :) | 10:57 |
=== doko_ [~doko___@dsl-084-059-073-068.arcor-ip.net] has joined #ubuntu-toolchain | ||
=== _infinity [~adconrad@loki.0c3.net] has joined #ubuntu-toolchain | ||
doko_ | lamont-away: gcc-3.4_3.4.4-3ubuntu1 will FTBFS on hppa. | 11:55 |
=== Seveas [~seveas@ksl403-uva-141.wireless.uva.nl] has joined #ubuntu-toolchain | ||
=== doko [~doko___@dsl-084-059-073-068.arcor-ip.net] has joined #ubuntu-toolchain | ||
=== chmj [~chmj@196.36.161.235] has joined #ubuntu-toolchain | ||
=== lamont-away [~lamont@mix.mmjgroup.com] has joined #ubuntu-toolchain | ||
=== jbailey [~jbailey@CPE00501836c657-CM014260028338.cpe.net.cable.rogers.com] has joined #ubuntu-toolchain | ||
jbailey | Oy, Fabio. =) | 01:34 |
fabbione | hey jbailey | 01:36 |
jbailey | Do I owe you anything still, or should I dive back into my task list. | 01:36 |
jbailey | ? | 01:36 |
fabbione | jbailey: a fix for glibc so i can compile the kernel? ;) | 01:37 |
fabbione | i was told that also gentoo is hitting the same problem as we do | 01:37 |
Mithrandir | jbailey: did you get your fix tested yesterday? | 01:37 |
fabbione | nah go ahead with your list :) | 01:37 |
fabbione | there is nothing in my list for you today :P | 01:37 |
jbailey | Mithrandir: I didn't - there was some cool things in parliament yesterday and so I spent my evening glued to news sites rather than working. | 01:38 |
jbailey | (And a chunk of the day...) | 01:38 |
fabbione | anyway i am off for the next hour/today | 01:38 |
fabbione | i had enough | 01:38 |
fabbione | catch you later guys | 01:38 |
jbailey | g'n Fabio. | 01:38 |
Mithrandir | this vax is sloooooooooow | 01:39 |
Mithrandir | jbailey: ok, tell me if you want me to power up my box. | 01:39 |
jbailey | Mithrandir: In about 30 minutes okay? | 01:40 |
jbailey | +Is | 01:40 |
Mithrandir | jbailey: sure, I just have to extend my hand to press the power button. :-) | 01:40 |
Mithrandir | it's booting now | 01:40 |
jbailey | Thanks. | 01:41 |
doko | jbailey: did you have a chance for the ecj-bootstrap upload? | 01:48 |
jbailey | doko: No, wasabi's box crashed before I fetched it, and he said that he had an ecj.1 for me in his other package. | 01:53 |
jbailey | Also, the ant he had posted was corrupted. | 01:53 |
jbailey | java-common was fine, though, so I did upload that. | 01:54 |
jbailey | I pinged him in #ubuntu-java a few minutes ago, I'm hoping to get this all done while he's still at home and has access to the box. | 01:54 |
doko | ok, fine | 02:16 |
=== Seveas [~seveas@seveas.demon.nl] has joined #ubuntu-toolchain | ||
=== jbailey [~jbailey@CPE00501836c657-CM014260028338.cpe.net.cable.rogers.com] has joined #ubuntu-toolchain | ||
lamont-away | ../../src/gcc/cppdefault.c:75: error: `LOCAL_INCLUDE_DIR' undeclared here (not in a function) | 04:24 |
lamont-away | WTH? | 04:24 |
lamont | -DLOCAL_INCLUDE_DIR=\"/usr/local/include\" | 04:24 |
=== lamont scratches his head | ||
jbailey | lamont: Run it through with gcc -E -dD into a file, and I can take a look for you. | 04:26 |
lamont | gcc/hppa64-linux-gnu/3.4.5/../../../..`echo /usr | sed -e 's|^/usr||' -e 's|/[^/] *|/..|g'`/include/c++/3.4.5\" -DGPLUSPLUS_TOOL_INCLUDE_DIR=\"/usr/lib/gcc/hppa64-linux-gnu/3.4.5/../../../..`echo /usr | sed -e 's|^/usr||' -e 's|/[^/] *|/..|g'`/include/c++/3.4.5/hppa64-linux-gnu\" -DGPLUSPLUS_BACKWARD_INCLUDE_DIR=\"/usr/lib/gcc/hppa64-linux-gnu/3.4.5/../../../..`echo /usr | sed -e 's|^/usr||' -e 's|/[^/] *|/..|g'`/include/c++/3.4.5/backward\" -DLOC | 04:27 |
lamont | AL_INCLUDE_DIR=\"/usr/local/include\" -DCROSS_INCLUDE_DIR=\"/usr/lib/gcc/hppa64-linux-gnu/3.4.5/../../../../hppa64-linux-gnu/sys-include\" -DTOOL_INCLUDE_DIR=\"/usr/lib/gcc/hppa64-linux-gnu/3.4.5/../../../../hppa64-linux-gnu/include\" -DTARGET_MACHINE=\"hppa64-linux-gnu\" \ | 04:27 |
lamont | -c ../../src/gcc/cppdefault.c -o cppdefault.o | 04:27 |
lamont | ../../src/gcc/cppdefault.c:75: error: `LOCAL_INCLUDE_DIR' undeclared here (not in a function) | 04:27 |
lamont | build directory is purged - I'll rebuild it once gcc-4.0 finishes dying | 04:28 |
jbailey | If you have to rebuild anyway, can you do it on j5k? | 04:28 |
jbailey | That way I can poke into your build env. | 04:29 |
lamont | yeah - but then I have to create said build environment | 04:29 |
lamont | or do you already have a hoary chroot there? | 04:29 |
jbailey | I had a breezy one I think. | 04:31 |
jbailey | I suspect for this build, though, it wouldn't make that big of a different. Dunno for sure. | 04:32 |
lamont | given that we tossed the entire breezy hppa world..... | 04:35 |
lamont | may just be easier to create access for you on this machine... wanna email me an ssh key? | 04:35 |
fabbione | hey lamont | 04:35 |
lamont | morning fabbione | 04:39 |
lamont | ah, libcairo. sigh | 04:40 |
doko | <doko> lamont-away: gcc-3.4_3.4.4-3ubuntu1 will FTBFS on hppa. | 05:04 |
doko | lamont: ^^^ | 05:05 |
lamont | doko: GAH! | 05:05 |
lamont | I need a buildable gcc-3.4.... | 05:05 |
elmo | doko: dude, I think you need to add some conflicts to gcc | 05:05 |
elmo | almost all my buildds got fucked by partial upgrades of gcc and not g++ | 05:06 |
doko | lamont: chinstrap:~doko/multiarch-include.dpatch | 05:07 |
doko | elmo: how does it fail? | 05:07 |
elmo | doko: c++ compiler can not create excutable, or massive include file fuckage | 05:07 |
elmo | see e.g. the apt build log for ssparc | 05:08 |
doko | where can I get the log? | 05:08 |
doko | ahh, Debian? | 05:08 |
elmo | yes, sorry, wildly and abusively OT | 05:08 |
doko | hmm, buildd only has a log of a sucessful build | 05:09 |
elmo | blink? | 05:10 |
doko | 0.6.38 (sparc) (latest build at Jun 29 05:45: maybe-successful) | 05:10 |
elmo | http://buildd.debian.org/build.php?&pkg=apt&ver=0.6.38&arch=sparc&file=log | 05:11 |
elmo | follow the link of version num | 05:11 |
elmo | http://buildd.debian.org/fetch.php?&pkg=apt&ver=0.6.38&arch=sparc&stamp=1119821334&file=log&as=raw | 05:11 |
doko | ahh, yes, that's the changed internal gcc_lib_dir (/usr/lib/gcc-lib/...) | 05:15 |
doko | so, basically, I would have to add conflicts in cpp-3.3 to all previous versions of gcc-3.3, gobjc-3.3, g77-3.3, g++-3.3, gnat-3.3, treelang-3,3, gpc-3.3 | 05:16 |
doko | interesting. | 05:17 |
elmo | or Breaks: if we had that | 05:17 |
doko | hmm, I did promise lamont not to upload gcc-3.3 again ;-) | 05:19 |
elmo | oh, don't do it again in a hurry, or I suspect ryan will fly over to kick your ass | 05:20 |
elmo | but I think we should fix it at some point | 05:20 |
elmo | most buildds are or will be fixed anyway | 05:20 |
lamont | doko: feel free to upload 3.3 | 05:32 |
elmo | feel free not to | 05:33 |
elmo | :P | 05:34 |
lamont | hehe | 05:34 |
lamont | doko: if the 3.4 build works for me on hppa, should I upload the source? | 05:34 |
doko | lamont: no, elmo surely wants the dont-break-the-stupid-buildd patch there as well ;) | 05:40 |
lamont | yeah | 05:40 |
lamont | doko: I just need to drop that into debian/patches? or add it's name to something as well | 05:41 |
lamont | ? | 05:41 |
doko | just drop it in, it's a corrected patch. | 05:42 |
lamont | replaces the one there... got it. | 05:42 |
lamont | doko: so there will be a -3ubuntu2 or such sometime soonish? | 05:44 |
doko | lamont: yes | 05:46 |
=== Seveas [~seveas@seveas.demon.nl] has joined #ubuntu-toolchain | ||
=== lamont [~lamont@15.238.5.143] has joined #ubuntu-toolchain | ||
=== chuck [~chuck@CPE0006258ec6c2-CM000a73655d0e.cpe.net.cable.rogers.com] has joined #ubuntu-toolchain | ||
=== chuck_ [~chuck@CPE0006258ec6c2-CM000a73655d0e.cpe.net.cable.rogers.com] has joined #ubuntu-toolchain | ||
=== lamont grumbles at doko for uploading ardour -1build1 when -1 was FTBFS everywhere | ||
doko | lamont: tell me, where I can get this information (besides scanning the build logs) ... | 08:32 |
lamont | archive.ubuntu.com/ubuntu/pool/universe/a/ardour/ | 08:35 |
lamont | :-) | 08:35 |
lamont | OTOH, -1build1 is far better than -1ubuntu1 :-) | 08:35 |
doko | lamont: when do you want a gcc-3.4 -3ubuntu2 upload? | 08:43 |
lamont | anytime is fine | 08:44 |
lamont | I | 08:44 |
lamont | I'll let -3ubuntu1hppa1 finish building, and use that locally to let things run. | 08:44 |
doko | ok, finishing a test build here. | 08:45 |
lamont | given hppa's gcc-defaults.... do I need to build gcc-4.0 before I open the floodgates? (No hppa debs from gcc-4.0 in my archive) | 08:45 |
lamont | hrm... actually, gcc-4.0 uses gcc-3.4 to build, correct? | 08:46 |
doko | no, do you have an libgcc2 and libstdc++6? | 08:46 |
lamont | no libgcc2 | 08:46 |
doko | you need that one. it's built from the gcc-4.0 sources ... | 08:47 |
lamont | libstdc++6_4.0.0-8ubuntu3_hppa.deb | 08:47 |
lamont | that work for you? | 08:47 |
doko | yes, I think so. | 08:47 |
lamont | I'll run with that then | 08:57 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!