ibotlog2html for #crosstool-ng

<< Previous 2014-09-08 Next >>

# 01:59:58 enunes quits : Read error: Connection reset by peer
# 02:04:56 enunes joins #crosstool-ng
# 03:12:26 enunes quits : Ping timeout: 268 seconds
# 03:18:03 sh4rm4 joins #crosstool-ng
# 03:26:16 enunes joins #crosstool-ng
# 03:44:00 GrantK joins #crosstool-ng
# 03:48:06 GrantK Hi. I'm having a heck of a time wit ct-ng -- on Debian :-/ When I build an arm-target toolchain on my Opensuse env, the generated bins are, as expected, "ELF 32-bit LSB executable, ARM" bins. WHen I follow the same procedure, using the same .config, in a Debian env, the generated bins are *named* correctly, but id as 'ELF 64-bit LSB executable, x86-64' -- the env of the host, not the target.
# 03:48:31 GrantK I'm missing something simple :-/ Been beating my head against this for hours. Any hints appreciated!
# 04:48:52 lazzer quits : Ping timeout: 240 seconds
# 04:50:06 lazzer joins #crosstool-ng
# 04:50:28 GrantK quits : Quit: GrantK
# 05:27:59 perr joins #crosstool-ng
# 05:43:53 djerome quits : Remote host closed the connection
# 07:47:20 EsbenH^afk is now known as: EsbenH
# 09:43:19 Net147 joins #crosstool-ng
# 10:39:32 sh4rm4 quits : Ping timeout: 260 seconds
# 11:06:10 perr quits : Quit: Leaving
# 13:38:16 mnt_real joins #crosstool-ng
# 14:10:07 sh4rm4 joins #crosstool-ng
# 14:10:47 perr joins #crosstool-ng
# 14:10:47 perr quits : Changing host
# 14:10:47 perr joins #crosstool-ng
# 14:11:41 sh4rm4^bnc parts #crosstool-ng
# 14:14:00 alan_o joins #crosstool-ng
# 14:41:53 Net147 quits : Quit: HydraIRC -> http://www.hydrairc.com <-
# 15:27:10 perr quits : Quit: Leaving
# 17:10:24 y_morin joins #crosstool-ng
# 17:11:30 bhundven y_morin: I'm ok with adding the optimization patch later, either the debug or warning patch will need to 'declare extra_config'
# 17:11:49 bhundven and add it to the command line
# 17:11:54 bhundven *configure line
# 17:12:15 y_morin bhundven: Howdy!
# 17:12:27 bhundven y_morin: hello! happy monday
# 17:13:03 y_morin bhundven: Yeah.. First monday back at work after a 4-week holiday. No need to say that was a looong day...
# 17:13:34 bhundven :D
# 17:13:47 bhundven I'm sure you have "real work" to catch up on, too.
# 17:13:56 y_morin bhundven: Yep!
# 17:14:31 y_morin bhundven: There is one last sample that causes some headaches: powerpc-unknown-linux-uclibc: that one does not build :-(
# 17:14:52 bhundven y_morin: I can take a loo
# 17:14:54 bhundven *look
# 17:15:02 y_morin bhundven: I think I'll just ditch it. PPC embedded is not usually size-constrained, so maybe uClibc does not matter too much...
# 17:15:19 y_morin bhundven: Oh, if you want to have a look, be my guest! ;-)
# 17:15:40 y_morin bhundven: I'll push the last cset that updates the samples (except that one).
# 17:15:47 bhundven ok
# 17:15:48 y_morin bhundven: BTW, can you reach the server, now?
# 17:15:58 bhundven y_morin: yes, server works now for me
# 17:16:06 y_morin bhundven: OK, perfect.
# 17:16:33 bhundven I still have a p1020/p2020 freescale powerpc-spe (32-bit) board... somewhere.
# 17:16:49 bhundven I think it's in storage
# 17:17:09 y_morin bhundven: Well, I'm usually doing build-time testing only on the samples.
# 17:17:24 y_morin bhundven: That one does not even build the toolchain.
# 17:18:07 bhundven yea... I sort of put the toolchain build test stuff on the back burner, till I get a job.
# 17:18:31 bhundven I have some short-term contract work, that should be starting any day now.
# 17:19:01 y_morin bhundven: Great news! :-)
# 17:19:33 RushPL quits : Ping timeout: 260 seconds
# 17:34:06 diorcety joins #crosstool-ng
# 17:40:01 bhundven smartin: I see from watching #buildroot, you are the cmake master.
# 17:40:58 bhundven smartin: I'm looking for a build configuration option that is lighter then Kconfig, but easier then autotools.
# 17:41:36 bhundven I don't have a problem with m4... I just think it should have died with sendmail
# 17:41:52 bhundven :P
# 17:43:38 y_morin bhundven: :-)
# 17:47:51 bhundven smartin: the project I have is mostly shell scripts.
# 17:47:54 RushPL joins #crosstool-ng
# 17:48:55 y_morin bhundven: smartin is his work-time nick. He's known as smartin_ when he's at home.
# 17:49:08 bhundven ah, ok
# 17:49:23 y_morin bhundven: You may have a better chance of him to answer when he's in as smartin_
# 17:49:42 bhundven figured it was the opposite way
# 17:50:14 bhundven most people 'nick' when avail, 'nick_' when away, although, I just use /away
# 17:50:26 y_morin Yup.
# 17:50:38 y_morin bhundven: smartin is, well, different! :-]
# 17:50:50 y_morin Reference inside... :-)
# 17:51:29 bhundven ok, well. I'll wait till later for those questions. I guess for now, I'll read the cmake tutorial and try to come to my senses.
# 18:47:46 loide joins #crosstool-ng
# 19:41:46 bhundven y_morin: not sure if you saw this: https://sourceware.org/git/?p=glibc.git;a=commit;h=f083450f45f580bf33fa345ddadb897dc85d78e8
# 19:42:33 y_morin bhundven: Hm, no. that's news to me. Is it in 2.20 ?
# 19:42:39 bhundven yes
# 19:42:45 bhundven https://sourceware.org/ml/libc-alpha/2014-09/msg00088.html
# 19:43:37 y_morin bhundven: Damn. So we can't really apply your patch without changes in the glibc build script, then...
# 19:44:19 y_morin bhundven: I think we need to add LIBC_GLIBC_NPTL_FORCED and seelct it from LIBC_GLIBC_2_20_or_later.
# 19:44:28 bhundven y_morin: well, you can, but nptl will just be used no matter what
# 19:44:32 y_morin (or directly)
# 19:46:12 bhundven "The minimum Linux kernel version that this version of the GNU C Library
# 19:46:12 bhundven can be used with is 2.6.32."
# 19:47:48 y_morin bhundven: Re. NPTL: we currently force outr add-ons list to include nptl, so this should not be done now...
# 19:47:53 bhundven y_morin: I will make a new patch.
# 19:48:27 bhundven hrm?
# 19:48:39 bhundven which should not be done now?
# 19:50:49 y_morin bhundven: http://crosstool-ng.org/git/crosstool-ng/tree/scripts/build/libc/glibc-eglibc.sh-common#n484
# 19:51:08 bhundven y_morin: right, that really doesn't matter.
# 19:51:22 y_morin bhundven: I gues we do not want to do the case-esac statement for glibc >= 2.20, now.
# 19:51:37 bhundven but we do for <= 2.19
# 19:51:43 y_morin bhundven: Right./
# 19:51:49 y_morin s/\/$//
# 19:52:10 y_morin He... Poor regexp implementation in ctngbot... :-(
# 19:52:23 bhundven and for >= 2.20... you can pass --enable-add-on=TheWorldIsEnding and nothing different will happen.
# 19:52:31 bhundven nptl will still be forced to be enabled
# 19:52:35 bhundven and life will go on
# 19:53:08 y_morin bhundven: Well, it probably does not matter than we still ist nptl in there, but for sanity, I believe it is better not to.
# 20:00:01 y_morin bhundven: Regarding the release: do you think you can look at the remaining sample soon, or should I just cut the release now?
# 20:00:31 y_morin bhundven: Cutting the release now will allow to apply your recent patches, but it's sad we're missing a sample.
# 20:00:53 y_morin bhundven: OTOH, having all samples up-to-date is nice, but your patches languish in the meantime...
# 20:01:43 y_morin bhundven: As I said previously, this sample is very important to me, so I'm fine with cutting the release now. But I can wait if you want me to.
# 20:02:17 bhundven y_morin: you just committed the ppc samples
# 20:02:23 bhundven which one has a problem?
# 20:02:56 y_morin bhundven: powerpc-unknown-linux-uclibc which I did not update.
# 20:03:02 bhundven oh, ic
# 20:03:07 y_morin As you said you could have a look.
# 20:03:22 bhundven sure, give me a few minutes with it
# 20:04:35 y_morin bhundven: OK, thanks! :-)
# 20:18:04 bhundven runs build
# 20:31:32 diorcety quits : Read error: Connection reset by peer
# 20:34:34 diorcety joins #crosstool-ng
# 20:35:44 bhundven y_morin: do you want to bump binutils 2.23.x and 2.24 to be not experimental?
# 20:36:22 bhundven since, they've been out since 2013 and no one has really had any problems, except a few strange cases I can think of.
# 20:36:37 bhundven (2.23 since 2012)
# 20:37:13 y_morin bhundven: I'd prefer we stick without EXPERIMENTAL if possible (I tried to remove from as much samples as possible.)
# 20:37:27 y_morin bhundven: I'll unmark those after the release.
# 20:37:57 bhundven y_morin: sure.
# 20:39:07 diorcety quits : Read error: Connection reset by peer
# 20:49:03 bhundven y_morin: this is a funky uclibc config
# 20:49:55 bhundven y_morin: I vote that we skip this config for now, and release as you had suggested before.
# 20:50:39 y_morin bhundven: OK, will do the release now, then.
# 20:51:15 y_morin bhundven: We can tweak the config later, or remove the sample altogether if it proves unbearable...
# 20:51:45 bhundven y_morin: I think it needs to be updated, but I'd like to make sure it actually works. I just don't have time to mess with qemu today.
# 20:52:07 bhundven s/think it/think the uclibc config/
# 20:53:05 y_morin bhundven: Sure, no problem. Thanks for looking! :-)
# 20:53:23 bhundven y_morin: it's definitely a problem with the uclibc config, though.
# 20:53:26 y_morin bhundven: BTW what was the issue you got? Mine was about GCC_NO_EXECUTABLE in libstdcxx
# 20:53:40 y_morin libstdc++
# 20:53:41 bhundven mine was with crtn.o
# 20:53:54 y_morin bhundven: OK, let's ditch that sample for now.
# 20:53:56 bhundven so I must have gotten further
# 20:54:11 bhundven I did modify the ct-ng config
# 20:54:25 bhundven but I didn't touch the uclibc config
# 20:54:35 bhundven so I'm guessing the uclibc build used to be 3-pass
# 20:54:37 y_morin bhundven: I'll test-build the current sample, and if it does not even build as-is, I'll mark it broken, and do the release.
# 20:54:40 bhundven and that sample is from then
# 20:55:41 bhundven the biggest thing that needs to be updated is the gcc version, binutils (2.22), and the companion libs
# 20:55:57 bhundven for that specific sample
# 20:56:14 bhundven also, run oldconfig before you try it
# 20:56:23 y_morin bhundven: Why?
# 20:56:39 y_morin bhundven: because of: override: reassigning to symbol ARCH_32
# 20:56:41 y_morin ?
# 20:56:44 bhundven yup
# 20:57:04 y_morin bhundven: Harmless, it's just kconfig being whinny for nothing.
# 20:57:16 y_morin bhundven: But I already did use oldconfig anyway! :-)
# 21:10:01 y_morin Damn... MPFR: automatic de-ANSI-fication support has been removed
# 21:10:10 y_morin OK, ditching this sample for now...
# 21:10:12 bhundven :)
# 21:10:13 bhundven lol
# 21:10:21 bhundven yea, that's why I updated the config
# 21:10:26 bhundven cause I got that
# 21:10:49 y_morin Well, as I said: ditch! :-)
# 21:11:27 bhundven "That sample didn't work... CUT THE RELEASE!" :D
# 21:14:15 y_morin bhundven: Basically what I wrote in the commit log! :-)
# 21:15:00 y_morin bhundven: http://code.bulix.org/8k3zgt-86927
# 21:15:19 bhundven :D
# 21:19:06 y_morin bhundven: What about the versioning scheme? Should I continue with the current one, or should I switch to something that is more sensible ?
# 21:21:31 bhundven y_morin: the only thing that doesn't make sense about the current version scheme is the release number
# 21:21:40 bhundven Major.Minor.Release
# 21:21:58 bhundven it's been a very long time since 1.19.0 was out
# 21:22:01 bhundven and no releases since
# 21:22:05 y_morin bhundven: Yup.
# 21:22:21 y_morin bhundven: What about YYYY.MM as Buildroot does, then?
# 21:22:24 bhundven if you don't plan to do releases often, then the release number makes no sense.
# 21:23:16 bhundven I think we should stick with, what... 1.20.0, and discuss this one more, in a non-hasty-last-moment-change.
# 21:23:38 y_morin bhundven: OK.
# 21:24:31 y_morin bhundven: I know I haven't cut a release for about a year, indeed. But there was almost nothing to make a new release, until recently.
# 21:24:50 y_morin (or I'm completely oblivious of hwat went in?)
# 21:25:06 y_morin Anyway, 1.20.0 it will be.
# 21:25:09 bhundven y_morin: the other side of that coin is that you have also been really busy with other stuff.
# 21:25:30 y_morin bhundven: Unless I ditch the .0 and just keep a two-digit version?
# 21:25:33 y_morin bhundven: True.
# 22:09:09 bhundven I realy need a slimmer desktop
# 22:09:18 bhundven debian with gnome3... :sigh:
# 22:09:45 bhundven http://dpaste.com/04HBNY5
# 22:10:13 bhundven Now, I know how to read that, and that 1968 is cached
# 22:10:22 bhundven *1968M
# 22:10:34 bhundven but still...
# 22:11:35 y_morin bhundven: "cached" -> "caché" in french, which means "hidden"!
# 22:12:43 bhundven y_morin: more or less, it is hidden, but is calculated in to "Used"
# 22:12:58 bhundven A while back I was running just openbox
# 22:13:13 bhundven and from a fresh boot up, I I was only using 250M
# 22:13:24 y_morin bhundven: I just use a full-screen terminal (no decoration, no nothing) with just tmux inside.
# 22:14:05 bhundven I use tmux as well.
# 22:47:17 y_morin quits : Quit: Nighty Night!
# 23:31:33 blarghlarghl quits : Ping timeout: 246 seconds
# 23:39:04 blarghlarghl joins #crosstool-ng