=== lamont__ [~lamont@15.238.5.156] has joined #ubuntu-toolchain === doko [~doko___@dsl-082-082-192-254.arcor-ip.net] has joined #ubuntu-toolchain [03:06] doko : Hrm. That list doesn't really give me a terribly good indication of what to dep-wait on. [06:13] morning [06:18] jbailey: ping? [06:23] infinity: did you get the script done? [08:32] fabbione : Have you looked at doko's frozenapps.txt?... It's not particularly evident what one should be dep-waiting on. [08:33] (I suppose I could dep-wait on package renames as I hope/expect them to be, but that's kinda icky) [08:34] infinity: no i didn't [08:34] and there is no way to track what pkg needs what [08:34] so basically i have sparc stalling [08:34] and i think i did build all the libs [08:34] but we can't really check if not manually [08:37] Well, you can check his list and those should be the only packages that need to be left unbuilt. [08:38] But I can't do proper dep-waits for them without some more input. [08:38] I might just not-for-us the mess of them, and have doko tell me when to release them, though. [08:38] Since it's pretty simple to toggle a -n. [08:46] yeah [08:47] morning alll [08:48] here is the guilty boy [08:50] infinity: what do you need? [09:10] doko : An indication of what each of those frozen packages should actually build-dep on? :) [09:10] doko : Erm, dep-wait, I mean. [09:11] doko : Knowing that foo build-deps on libbar0 isn't as helpful as knowing that foo should dep-wait on libbar0c2, but i don't want to make assumptions about package renames. [09:15] infinity: so, you do need the new package names? [09:19] infinity: added a column, let mw knwo what you need as the format [09:26] doko : I'm not so picky on format, I just want to know that "when package (or package_version) is in the archive, this should be built" [09:27] doko : So, if that's "foo dep-wait libbar0c2, libbaz1c2, libquux2", or will need versioned build-deps for some, I need to know. [09:28] doko : So that needs ot take into account c102 libs being names back, libs being names to c2, and god knows what else. If it's easier for you to just ping me each time you think an app can/should be tried, we can do that instead, but automating it is nice. :) [09:34] and foo is the source name? [09:36] infinity: list updated [09:36] foo is the source package, libbar, etc are binary packages. [09:38] is this the kind of thing you need? [09:39] Beautiful. [09:39] Even better if it's all correct. :) [09:39] wait ... [09:39] I have to go grocery shopping with my girlfriend, when I get back, whatever the state of that file is, I'll hammer it all into wanna-build and remove the blacklists from the buildds. [09:40] infinity: hoorah for safeway [09:40] infinity: get some Thai for me [09:40] for this going to work, I'll have to update the list in the wiki with the new library names ... [09:40] Hooray indeed. Those bastards are holding my cow hostage, and I want it back! [09:40] infinity: mmm. it's thai night for me, tonight. [09:40] infinity, when are you back? [09:40] doko : To make sure the MOTUs upload with the right renames? [09:41] doko : 1.5-2 hours tops. [09:41] However long it takes to get some steaks and other essentials and head home. [09:41] ok. well I need to give YOU the new names well ... [09:43] doko : I'll ping you as soon as we're back. [09:44] <\sh> doko: what should we do with this ocaml gcc4 issue on i386? i didn't find anything yesterday about a solution...and I'm not this asm hacker ;) [09:50] \sh, build-dep on GCC 3.4 [09:52] <\sh> doko: that means all packages depending on ocaml in any way, have to be transistioned to gcc-3.4 [10:01] \sh why? do all the packages use this assembler code? === chmj [~d3vic3@dumbledore.hbd.com] has joined #ubuntu-toolchain [10:02] <\sh> doko: no, but if i want to compile a g++ app with g++ 4 and it's linking against ocaml it breaks at linker stage [10:03] hmm, what does break? [10:05] <\sh> shit..i'm searching..i need a konqui shortcut for buildlogs ;) [10:07] <\sh> forget it...gcc-3.4.gcc-opt: not found... [10:07] <\sh> something else [10:08] \sh that looks like a buildd problem [10:08] no it's not [10:08] doko: that's a missing b-d on gcc-3.4 [10:08] or that [10:08] \sh, what are you looking for? [10:08] gcc-3.4.gcc-opt is only a wrapper for gcc-3.4 [10:09] but if gcc-3.4 is missing it reports exactly the error as it should be [10:10] <\sh> fabbione: it is a missing b-d [10:10] that's what i said :) [10:11] i see the same errors in the sparc buildd [10:11] \sh: add it ;-) [10:11] <\sh> ok..that means, all ocaml universe apps has to be touched e.g. findlib [10:11] <\sh> doko: on my way ;) [10:12] doko: it is probably simpler to add a Depends: gcc-3.4 deps to the ocaml-dev thingy [10:12] given that's a reasonable solution [10:13] fabbione: yes, sounds good, and then I can do automatic uploads for the remaining things. [10:13] <\sh> fabbione: there is no ocaml-dev [10:13] <\sh> Binary: ocaml-interp, ocaml-nox, ocaml-base, ocaml, ocaml-compiler-libs, ocaml-base-nox, ocaml-native-compilers, ocaml-source [10:14] <\sh> but anyways, some packages are b-d on ocaml-base-3.08 (as package name) or ocaml-nox-3.08 (as packagename) [10:14] afaik it's ocalm calling gcc-3.4 automatically [10:15] so it is ocalm that has to pull it in [10:15] and not the 3000 apps [10:15] <\sh> fabbione: ocaml has gcc-3.4 as b-d [10:16] so, on which package b-d all the 3000 apps? [10:16] <\sh> first of all, we have to get rid of things like this: [10:16] <\sh> Depends: libncurses5-dev, ocaml-base-nox (=${Source-Version}), ocaml-base-nox-3.08.3, ocaml-interp-3.08.3 [10:17] \sh: ocalm needs to have it as Depends: [10:17] so that gcc-3.4 is pulled in for apps to build [10:17] <\sh> because those ocaml-base-nox-3.08.3 is a provides in one of the packages resulting of source ocaml [10:17] <\sh> i think to be compatible to debian? [10:18] heh, did you have a look at the package maintainer ... [10:18] svenl: your ocaml package is b0rked ;-) [10:19] ahahha [10:19] doko: i am almost done with ppc64 kernels :) [10:19] if i feel lucky i will upload them today [10:20] heh, cool. time to upload GCC 4.0 ;-) [10:20] it slightly depends on how it builds [10:20] doko: go ahead really [10:20] doko: until all this C++ crap is sorted at buildd level, i am just trashing time in universe [10:20] \sh, what is a typically build-dep for an ocaml related package? [10:20] and i need the /lib64 fix for sparc [10:20] ahh, sure [10:22] well, it cannot be built at the moment. lib32z1-dev is missing in main [10:22] <\sh> doko: ocaml-3.08.3 [10:22] well that's a dep-wait [10:22] but why does it need that new lib? [10:23] <\sh> doko: as well for Runtime Depends [10:25] well, yes looking at the package description, fabbione is right, just add g++-3.4 as dependency to ocaml [10:25] doko: and gcc-3.4 please [10:25] fabbione: building 32bit libgcj on amd64 [10:25] ah ok [10:25] fabbione: not needed [10:26] this multiarch crap is crap :P [10:26] <\sh> doko: can u do it, cause it's main ;) [10:26] ok [10:26] <\sh> thx [10:26] ahh, it only needs gcc, not g++ ... [10:26] <\sh> what about the ocaml-3.08.3 package name? we should get rid of it [10:26] anyway, can't hurt ... [10:27] doko: speaking of "can't hurt" [10:27] <\sh> ok..lets fix some digital tv stuff [10:28] what about ncurses for ppc64 [10:28] ? [10:28] i don't need it as kernel b-d but quite a lot of users that recompiles kernels need it [10:28] fabbione: do it, I don't need it ;-) [10:29] doko: me neither.. i don't even a damn ppc at home! [10:29] :) [10:29] heh, "our priority are our users" ;) [10:29] so send me a g5 :) [10:30] and make me a ppc user [10:54] fabbione: this aint multiarch. :P === Seveas [~seveas@seveas.demon.nl] has joined #ubuntu-toolchain [11:00] unknown Debian architecture powerpc64, you must specify GNU system type, too at /usr/bin/dpkg-architecture line 214. [11:00] grrr [11:04] doko: nope. [11:04] doko: x86 sucks. [11:04] doko: works fine on all other arches but loosy x86 with gcc-4.0. [11:05] svenl: fix it ;) [11:05] There is already a bug report upstream though, so will probably be fixed in the next release. [11:05] doko: amd64 works fine, so throw away all that loosy legacy x86 hardware you have around and get real stuff instead :) [11:06] seriously, mark this upstream, and let some time pass, or write a follow up to the bug report. [11:10] http://pauillac.inria.fr/bin/caml-bugs/incoming?id=3637;expression=uther;user=guest [11:17] Back. [11:20] doko : Is your dep-wait stuff deinifely accurate and ready to go into wanna-build? [11:20] I did ... [11:21] 1) get the pairs of oldlib/newlib [11:21] 2) looked, which oldlib's are still in the archive [11:22] 3) looked for packages, which depend on those old libs [11:22] 4) Match source package to new libs [11:22] 5) PROFIT [11:23] Maximum number of CPUs (2-128) (NR_CPUS) [32] (NEW) [11:23] ok.. how many CPU should we support for ppc64-smp ? [11:24] i had say 32 is more than enough :) [11:25] ahh, you do want the source names instead of the new libnames? [11:29] Eek. THis looks scary. [11:29] Use of uninitialized value in scalar assignment at /usr/bin/dh_shlibdeps line 138, line 82. [11:29] (repeat about 100 times, with different line numbers) [11:29] doko : No, no. The source packages we're building. [11:29] doko : source apps, binary libs. [11:30] fabbione : How much memory do you lose for each CPU? [11:30] so why step 4) [11:30] ? [11:31] fabbione : I can't imagine people installing Debian on a pcc64 system/partition with 128 CPUs, but such systems exist (with 4096 CPUs, even) [11:31] doko : Oh, I meant "match the source of the apps with the new libs they dep-wait on" [11:33] yes, that's done in frozenapps.txt [11:33] ahh, source ... [11:34] no, that's already there ... [11:34] Excellent. Kay. I'll shove it in wanna-build in a few mins. [11:38] okay, and then, all frozen C++ apps should build? [11:42] They all should build automagically as the libraries they depend on get installed, that's the theory. [12:00] infinity: if there only was a help entry...... [12:02] For? [12:03] for the cpu thingy [12:03] now i remember why i didn't want ppc64 [12:03] just the config part is rather frigging boring [12:05] Okay, script looks goodish. [12:05] doko : Would it be bothersome to get those dep-waits comma-separated, so I can be even lazier? :) [12:06] you're eyes are still ok, aren't they? ;-) [12:08] Last time I checked... [12:09] check the file, not your eyes ;-P [12:09] Dude, you just changed that. [12:09] You must have. [12:09] :) === infinity stares. [12:09] MY COPY HAS NO COMMAS, DAMNIT! [12:09] Also, thanks. [12:09] be honest, you removed the commas ;) [12:11] Seriously, I hate you guys.. [12:12] Oh, cock. [12:13] Those source packages need versions too. [12:13] Maybe I can just fudge it. [12:14] On, no I can't. Whose silly idea was this? [12:14] I can't dep-wait stuff that's installed. Only stuff that needs building (or is building/failed, etc) [12:15] hmm, so which versions do you need? [12:15] One version higher than the current installed one, I guess. :) [12:15] And they all need to be in a state other than installed. Which is ludicrous. Unless they already are all uploaded but just not building due to the blacklist. [12:15] (THis seems unlikely) [12:16] Maybe not-for-us on all of them was a better idea. :) [12:16] Easy to toggle it on and off at will. [12:16] Just requires more human interaction, fewer crap perl scripts. [12:17] well, then just take the first column and freeze these again === infinity sighs. [12:17] It was such a GOOD crappy perl script too. === infinity kicks sand. [12:18] you did it in perl? [12:19] I stole the wanna-build interface from buildd-mail, which is in perl. [12:19] So it made sense to just put the glue around that in perl as well. [12:19] Anyhow. Suck. I'll just Not-for-us the whole bunch. [12:19] That means I can remove the blacklist from buildd.conf, right?... [12:20] (Yes, i realise if I admit to using anything other than python I can get fired from Canonical) [12:20] (I better not mention that I use SVN and CVS all the time, even in places where I could switch to arch but don't) [12:20] :) === infinity sobs. [12:35] for i in libopenhbci-plugin-ddvcard pingus glcpu qalculate fwbuilder gnomemm xbsql libprinterconf race libaqhbci libaqhbci libccaudio octave-forge ginac trophy gnunet gnucash python-omniorb2 libsdl-ruby valknut clanbomber acovea lablgtkmathview rsplib libaqbanking fwbuilder bakery-gnomeui2.0 qalculate php4-interbase xmule libccscript debtags fwbuilder rezound libbonobouimm1.3 xfe dbbalancer ginac libaqhbci-qt-tools doodle monopd debtags-edit bayo [12:36] echo "$i"; for j in i386 amd64 powerpc ia64; do wanna-build -b $j/build-db -n $i; done ;done [12:36] At least it's a more familiar language. :) [12:40] Neat, there were duplicates in that list. [12:43] fabbione : Do that, but make the final argument "$i"_1 (so you get a package version), and remove the duplicates from the list (or they'll toggle NFU, and then back again in the loop) [12:44] doko : Okay, can I remove the blacklist from the config files completely, then? [12:44] doko : Those were all the apps we care about not building? [12:45] infinity: yes, modulo the bugs that I probably made [12:46] Mmkay. The floodgates will open soon, then. Making config file changes on all the hosts. [12:58] infinity: sorry.. i wasn't following... [12:58] infinity: what should i do? [12:59] /usr/bin/make -j15 EXTRAVERSION=-1-powerpc64-smp ARCH=ppc64 \ [12:59] BTW :) [01:02] fabbione : Well, you only have one buildd, right? [01:02] yes [01:02] fabbione : You can either not-for-us everything in the list above (which is what I did), or just put that list in no_auto_build. [01:02] (I not-for-us'd it cause I have 12 buildds to deal with..) [01:03] The dep-wait thing won't actually work, since all but 2 of those packages are in state "Installed".. Whoever wasn't thinking of that earlier wasn't thinking clearly. [01:04] i already have the apps in no_auto_build [01:05] Right, is it just that set? [01:05] (it's a lot smaller than the old set) [01:05] yeps [01:05] Kay, then you have nothing more to do except wait for people to whine at us each time one needs to be removed. :/ [01:05] but i have still the full list banned becuase we don't have a way to check if all the libs have been built [01:05] at least not automatically [01:05] We can't do much better than that, unless someone wants to upload updated versions of all of them right now. [01:05] Then we can dep-wait them all. [01:06] infinity: I'll uploading things, after I know, what is currently in the queue and builds now [01:07] doko : Right, there were about 30 in the queue when I restarted buildd on the first machine. [01:11] interesting... [01:11] HAS_BIARCH := $(call cc-option-yn, -m64) [01:11] this one fails miserably on ppc64 [01:13] well i need food now [01:16] I need rest and real life now, I think. [01:17] As soon as I'm done restarting all the buildds. (sure, buildd re-reads its config, but it doesn't remove things from no_auto on re-read... clever, eh?... Feh) [01:24] fabbione: do you know about the correct fix to the xbase-clients postinst? [01:25] fabbione: pong [01:26] fabbione: You usually start at midnight my time, which is why when you show up I kow it's time to go to bed. =) (But I went to bed early last night *g*) [01:33] <\sh> doko: looks like now we have a new issue with ocaml ;) [01:36] which one? [01:38] re [01:38] doko: no, i didn't even look at it. [01:39] doko: daniels promised a fix asap [01:39] jbailey: ehhe no problem dude [01:39] elmo: ping? [01:41] doko, jbailey: at this point in time, is there any difference in build ppc64 in breezy-ppc64 or breezy chroot on davis? [01:42] if breezy is up-to-date, no [01:42] ok it is [01:42] fabbione: I am not aware of the states of the david choots. [01:42] (sorry, the only one I use often is Concordia) [01:42] Version: 3.4.4-0ubuntu3 [01:42] is this enough to build ppc64? [01:43] fabbione: Look for libc6-dev-ppc64 [01:44] mehhhh [01:44] I depends on a new enough gcc-3.4, and will also mean that the ppc64 bits have been loaded. [01:46] if neither thom or elmo are around i am stocked [01:46] breezy-ppc64 doesn't have kernel-build deps [01:46] and breezy is not updated [01:46] jbailey: we need to talk about that lustre thing [01:46] <\sh> doko: labltk [01:47] do you have time now? [01:47] fabbione: Yup. [01:47] ok [01:47] let's hook up jdub too [01:47] For you, my dear, I always have time. [01:47] lovely :) [01:48] should we other leave the channel for a few minutes? [01:48] hahahah [01:49] doko: Voyeur. [01:50] doko: You would just read the logs anyway. [01:50] sure, wouldn't you? [01:50] Oh sure. =) [01:51] Well, the first few paragraphs anyway. I like my erotica to be well written. === chmj clears the screen [02:24] hey ppc64 guys.. you need to modify dpkg to understand powerpc64 [02:24] dpkg-architecture -apowerpc64 -qDEB_BUILD_ARCH [02:24] unknown Debian architecture powerpc64, you must specify GNU system type, too at /usr/bin/dpkg-architecture line 214. [02:25] Why are you building ppc64 packages? [02:26] I thought you were working on biarch 64-bit-on-ppc32 stuff? [02:26] fabbione, it's ppc64 [02:26] fabbione: Does it do it that way for sparc64 too? [02:28] jbailey checking [02:29] jbailey: yes it does... [02:29] bah what a fucking mess [02:29] doko, ping [02:29] Ugh, that's insane. [02:29] I don't think there should ever be a _powerpc64.deb - nothing would know what to do with it. [02:30] jbailey: dpkg-arch needs to tell me that powerpc64 is built as powerpc [02:30] nothing more than that [02:30] Oh, dpkg-architecture, right. === jbailey smokes another one. [02:30] but i think the overall crap can be worked around from kernel-package [02:31] keybuk's not online. hmm. [02:31] I can just update the package, but I'd hate for the change to be lost if he thinks everything is in some VCS somewhere. [02:33] i might as well doing something wrong [02:36] infinity: why is libgnomemm2.0 2.0.1-2ubuntu2 not in the archives? [02:36] Because it's shy? [02:36] Or did you want a better reason? === infinity goes to look. [02:37] built an hour or so before [02:39] doko : It's uploaded by all 4 buildds, if it's not made it past katie, blame elmo. [02:39] doko : katie's been acting up all day :/ [02:41] infinity, thanks for looking [02:41] I've bugged elmo on your behalf. [02:41] He can rescue the builds, but with any luck he can actually find and fix the problem instead. :) [02:42] And I think this is now WAY past the end of my work day. [02:42] looks like he already did start the party :) [02:42] 1am? [02:43] 11pm, but I was supposed to quit at 6ish. [02:43] Mmm, so that's what divorce smells like. [02:44] *lol* [02:44] I read that as s/bugged/buggered/ [02:45] I wish.. [02:45] Erm. [02:45] ehehhe [02:45] Out loud voice. [02:45] La la la. [02:45] Well it would've explained the divorce... [02:46] See, and the next sentence was just too rude for a public channel. Shame. [02:51] infinity: how is Xfree86 going on m68k??? [02:51] About to sign and upload. [02:51] cool [02:56] And up she goes. [03:23] Mithrandir: I meant to be in this channel, sorry, ping. =) [03:23] So many "#ub..." tabs, all alike. === lamont__ [~lamont@15.238.6.116] has joined #ubuntu-toolchain [04:59] are there any known problems with dbus ? [05:24] jbailey: pong [05:25] Mithrandir: I'd like to start assembling some of the multiarch stuff, starting with getting amd64 building an i386 glibc and putting the files in the right places. [05:25] If there's a file in /usr/include and a file in /usr/include/i386-linux, the one in -i386-linux takes prioirty under your setup, right? [05:26] I don't remember. [05:26] look at the patch :-) [05:26] bah. ;) [05:31] Mithrandir: Do you know off hand if it's been applied yet? Wondering if the gcc-3.4 I have downloaded would have it in already or if I need to fetch it from you. [05:32] jbailey: ask doko. I don't think so. [05:32] doko: *poke* [05:32] jbailey: but then, asking me when I've been fed champagne might not be wise. :-) [05:33] Ah, is there an occasion? [05:34] Yeah, elmo and doko's birthdays. We're all celebrating. [05:34] I've been drinking for hours. [05:34] What's wrong with YOU? Funspoiler. [05:34] infinity: You're pretending to be Australian now. [05:34] infinity: You've been drinking for *months* [05:35] It's sort of like being in Vancouver. After you've been there you can always tell a vancouverite by the smell of pot oozing from their pores. [05:35] It takes years to go away. [05:35] (It's a theory Angie and I have as to why we only seem to meet people from Vancouver in Toronto) [05:35] jbailey: end of studies thing, aka "exmatriculation". [05:35] ;) [05:35] which means I've drunk close to a bottle of champagne. [05:36] jbailey; but what about the pot smokers in toronto? [05:36] desrt: It's not the same. People here aren't proud of their pot smoking the way folks in Vancouver are. [05:37] desrt: Here they seem to be more drug addicts. It's sort of like finding someone who drinks as much as an Australian outside of Australia. [05:37] hmm. so the smell-oozing-from-pores thing goes along with pride [05:37] heh. [05:37] it's funny, though... everyone has a different metric for who drinks a lot [05:37] germans, australians, canadians, ... [05:37] Vancouverites don't smoke pot, they just live in a constant cloud of burning week. [05:38] It's a whole different ballgame. [05:38] infinity: Right. =) [05:38] d/week/weed/ [05:38] jbailey: do australians drink? [05:38] s/d/s/ [05:38] Mithrandir : Australians pretend not to be alcoholics, but they seem to drink far more than anyone else I've ever met. [05:38] Mithrandir : The whole country is in denial. [05:38] Mithrandir: No, it's like how Adam described it for Vancouverites. They bathe in it. [05:38] infinity: I didn't see much alcohol in sydney. [05:39] We drank on the job at my last job. [05:39] Frequently. [05:39] And copiously. [05:39] TO drunkenness. [05:39] And we dealt with clients in that state. [05:39] And that was "normal". [05:39] infinity: yes? [05:39] The clients were also in that state, probably. [05:39] is that special? [05:40] Mithrandir : Well, it is to us anal-retentive north americans. :) [05:40] we did the same, hot dogs and beer, then back to coding at Opera. [05:40] Mithrandir : Maybe Australians are just scandinavians of the south. [05:40] infinity: this sounds perfectly normal to me. :-) [05:40] infinity: you coming to debconf? [05:40] Mostly, it was the complete lack of social activities that didn't involve drinking that got to me. [05:41] And still does on occasion. But then I get drunk and forget about it. [05:41] And no, not coming to debconf. Can't afford it. :/ [05:41] bah, you should have proposed a toast^Wtalk and gotten sponsorship [05:42] Meh. Next year. [05:42] Mexico city. [05:42] Oh. [05:42] The year after, then. [05:42] Mexico City and I have a longstanding agreement to disagree. [05:42] Scottland, I think? [05:42] what's mexico city like? [05:42] yeah, edinburgh [05:42] Mithrandir : The world's nastiest city. Period. [05:42] how so? [05:42] Mithrandir: You're less likely to get mugged than in Sao Paolo... [05:43] Over 30 million people, and nowhere near enough of them are employed in the cleaning business. [05:43] jbailey: I guess we'll just have Fortress Debconf, then? [05:43] Mithrandir: It worked in Porto Alegre... [05:43] jbailey: mpe [05:43] (I think it's cause all the cleaners are working illegally in the US, but I'm not sure) [05:44] Of course, it would've been nice if the guards had been more cautious or more explicit with their guns. [05:44] even jonas smedgaard didn't manage to get mugged. [05:44] which guards? [05:44] If Jonas got hit hard enough to forget about Debian, that might be nice. [05:44] But I didn't say that out loud. [05:45] I would be happy if he just forgot about bitching on debian-edu about _everything_ === lamont__ [~lamont@15.238.6.116] has joined #ubuntu-toolchain [06:01] hey lamont [06:19] infinity: edinburgh is the world's nastiest city? [06:22] Riddell : No, Mexico City is. [06:22] phew :) [06:23] what's happening in Edinburgh then? [06:30] Riddell: debconf7 [06:30] Mithrandir: where is that decided? [06:31] in porto alegre [06:31] iirc [06:31] google knows nothing of it [06:31] google doesn't know _everything_ [06:32] ah, found one reference http://listas.softwarelivre.org/pipermail/debconf4/2004-June/000651.html [09:55] fabbione: 'sup? [10:58] elmo; how's biarch coming? [10:59] (or going, i suppose) [11:01] desrt: I've no idea [11:02] <\sh> hmmm...g77 is not available for the 4.0 toolset?