ibotlog2html for #crosstool-ng

<< Previous 2015-11-26 Next >>

# 00:00:58 roolebo bhundven: thanks for that!
# 00:03:05 bhundven 5 different variants of custom sources:
# 00:03:21 bhundven 1. custom location is a file
# 00:03:27 bhundven 2. custom location is a directory
# 00:03:43 bhundven 3. custom location is a file
# 00:03:53 bhundven 4. custom location is a directory
# 00:04:15 bhundven hrm, I guess just four, I thought there was 5.
# 00:04:53 bhundven oh well :)
# 00:05:19 bhundven but I have to test all 4, with every component.
# 00:05:37 bhundven w00t :D
# 00:06:12 bhundven I'm thinking about dropping the global path
# 00:06:31 bhundven CT_CUSTOM_LOCATION_ROOT_DIR
# 00:06:44 bhundven it just makes the whole thing much more complicated.
# 00:07:39 bhundven the rewrite (and I mean I've totally rewritten) of CT_GetCustom is pretty ugly with the global path.
# 00:09:07 bhundven I've also been thinking about the possiblity of allowing local patches to be applied to custom sources.
# 00:11:12 bhundven but that may be later ;)
# 00:23:08 bhundven curious if anyone knows the answer to this, bash question
# 00:23:52 bhundven I have a variable: CT_GCC_CUSTOM_VERSION, and every component has a CT_{FOO}_CUSTOM_VERSION.
# 00:24:06 bhundven as well as a CT_{FOO}_VERSION
# 00:24:28 bhundven if the CUSTOM_VERSION is set, I want to set the normal VERSION to the custom version.
# 00:25:17 bhundven but, $(eval echo CT_${FOO}_VERSION)=${thecustomversion} doesn't really work that way...
# 00:25:25 bhundven er...
# 00:25:36 bhundven $(eval echo CT_${FOO}_VERSION)=${thecustomversion})
# 00:25:44 bhundven woops again
# 00:25:55 bhundven I had it right the first time
# 00:26:55 bhundven Right now, in each script, if custom location is set, then I set it manually, but it would be nice to wrap it up in the CT_GetCustom function without knowing the component name.
# 00:35:02 bhundven hrm, ok, delcare and indirection.
# 00:40:59 roolebo pardon me if i confuse terminology. why do we need custom source?
# 00:41:12 bhundven lazy people
# 00:41:26 roolebo bhundven: I think a path + set of patches in a directory is more than enough
# 00:41:33 bhundven I'm pretty sure I'm going to remove it.
# 00:41:54 bhundven so checkout this bash hackery
# 00:42:10 bhundven CT_GetCustom now has one argument
# 00:42:17 bhundven component_name
# 00:42:28 bhundven as it finds out the rest on it's own.
# 00:42:45 bhundven knowing that the component name is say: gcc
# 00:43:03 bhundven it knows there is a CT_GCC_CUSTOM_LOCATION and CT_GCC_CUSTOM_VERSION
# 00:43:17 roolebo bhundven: btw i tried to build with custom path set up. it only made symlink within non-custom binutils dir
# 00:43:26 bhundven right, and it doesn't work
# 00:44:02 bhundven so back to the hack for setting CT_GCC_VERSION to CT_GCC_CUSTOM_VERSION's value (hack)
# 00:44:17 bhundven in CT_GetCustom()
# 00:44:26 bhundven local component_name="$1"
# 00:45:04 bhundven local upper_component="$(echo ${component_name} | ${awk} '{ print toupper($0)}')"
# 00:45:38 bhundven local component_version=$(eval echo \${CT_${upper_component}_CUSTOM_VERSION})"
# 00:46:04 roolebo $(eval $(echo ...))?
# 00:46:16 roolebo though
# 00:46:33 bhundven nope
# 00:46:50 bhundven now is the trick to set CT_GCC_VERSION
# 00:47:18 bhundven local version_name="CT_${upper_component}_VERSION"
# 00:47:36 bhundven local "version_name"="${component_version}"
# 00:47:37 roolebo i'm quite behind. Still thinking of your original quesiton
# 00:48:12 bhundven tada, now CT_GCC_VERSION=${CT_GCC_CUSTOM_VERSION}
# 00:48:40 bhundven although, it's `local'
# 00:49:21 bhundven and, that doesn't work for uClibc/uClibc-ng
# 00:54:21 bhundven forgot his original question
# 07:30:26 imMute^ joins #crosstool-ng
# 07:35:36 aneyman_ joins #crosstool-ng
# 07:36:02 wycats quits : *.net *.split
# 07:36:28 bhundven quits : *.net *.split
# 07:36:29 blueness_ quits : *.net *.split
# 07:37:18 xenoxaos- joins #crosstool-ng
# 07:40:21 wbx quits : *.net *.split
# 07:40:21 Triskelios quits : *.net *.split
# 07:40:47 muep quits : *.net *.split
# 07:40:48 aneyman quits : *.net *.split
# 07:40:50 smartin__ quits : *.net *.split
# 07:40:51 imMute quits : *.net *.split
# 07:40:56 xenoxaos quits : *.net *.split
# 07:40:56 ovf quits : *.net *.split
# 07:41:04 xenoxaos- is now known as: xenoxaos
# 07:45:58 smartin_ joins #crosstool-ng
# 07:53:51 blueness joins #crosstool-ng
# 07:56:37 muep joins #crosstool-ng
# 07:58:00 diorcety joins #crosstool-ng
# 08:03:30 wbx joins #crosstool-ng
# 08:06:14 ovf joins #crosstool-ng
# 08:11:32 diorcety quits : Read error: Connection reset by peer
# 08:11:59 wycats joins #crosstool-ng
# 08:24:31 ovf quits : Changing host
# 08:24:31 ovf joins #crosstool-ng
# 08:24:49 wycats quits : Changing host
# 08:24:49 wycats joins #crosstool-ng
# 08:28:07 bhundven joins #crosstool-ng
# 08:30:13 bhundven git status so far on the _CUSTOM rewrite: +393/-258
# 08:30:32 bhundven 2 commits already in:
# 08:30:45 bhundven CT_Extract: Move check extracted up
# 08:31:00 bhundven config: Remove CUSTOM_LOCATION_ROOT_DIR
# 08:34:47 wycats_ joins #crosstool-ng
# 08:44:47 wycats quits : *.net *.split
# 08:45:05 wycats_ is now known as: wycats
# 08:45:18 bhundven man, what's up with the netsplits tonight?
# 08:56:29 diorcety joins #crosstool-ng
# 09:33:16 enunes quits : Ping timeout: 264 seconds
# 09:34:38 enunes joins #crosstool-ng
# 09:48:50 bhundven wbx: interesting warnings when building uClibc-ng with gcc master branch. Just warnings. It builds :)
# 09:49:09 bhundven aka gcc-6
# 10:08:45 bhundven But, I see something during the installing C library headers step that only goes to stdout and is not in the log
# 10:09:01 bhundven wbx: I'm trying to do the build again so I can isolate where it is happening
# 10:09:34 bhundven [EXTRA] Applying configuration
# 10:09:34 bhundven [19:45] / yes: standard output: Broken pipe
# 10:09:34 bhundven yes: write error
# 10:50:05 ragedragon joins #crosstool-ng
# 11:27:02 Net147 joins #crosstool-ng
# 11:35:59 ragedragon quits : Ping timeout: 250 seconds
# 13:11:11 bhundven roolebo: https://github.com/crosstool-ng/crosstool-ng/pull/288
# 13:11:20 bhundven roolebo: test that and let me know how it goes.
# 13:11:46 bhundven roolebo: read the comment for the CT_GetCustom to know what is expected of custom files
# 13:11:58 bhundven I still need to do the massive documentation update/rewrite...
# 13:12:12 bhundven If things work well, I'll update the kconfig help entries too.
# 13:14:50 bhundven quits : Quit: leaving
# 13:16:41 bhundven joins #crosstool-ng
# 13:21:10 bhundven quits : Ping timeout: 260 seconds
# 13:22:57 bhundven joins #crosstool-ng
# 13:22:57 bhundven quits : Changing host
# 13:22:57 bhundven joins #crosstool-ng
# 13:57:32 roolebo bhundven: checking out that. on the latest binutils build fails in glibc right after cc pass 2
# 14:24:49 ragedragon joins #crosstool-ng
# 14:33:13 ragedragon quits : Ping timeout: 250 seconds
# 16:32:39 wycats_ joins #crosstool-ng
# 16:35:36 wycats quits : Read error: Connection reset by peer
# 16:35:40 enunes quits : Ping timeout: 240 seconds
# 16:38:43 wycats_ is now known as: wycats
# 16:39:56 enunes joins #crosstool-ng
# 16:58:04 y_morin joins #crosstool-ng
# 17:11:46 bhundven roolebo: strange. I tested 4 differnet samples
# 17:11:56 bhundven roolebo: can you pastebin your log?
# 17:13:04 bhundven y_morin: hello!
# 17:13:30 y_morin bhundven: Hey! Sorry, I missed you yesterday...
# 17:13:37 bhundven no problem :)
# 17:13:57 bhundven was wondering if you could checkout #288 as well?
# 17:14:42 y_morin bhundven: Yes, sure. In a moment (what was the other you wanted me to look at, btw?)
# 17:15:18 bhundven #282
# 17:15:32 roolebo bhundven: I don't have access to laptop right now. That's the issue I hit https://sourceware.org/ml/crossgcc/2015-07/msg00023.html
# 17:15:32 y_morin OK. On the TODO list...
# 17:15:48 bhundven idk why #288 isn't getting hit by travis-ci... strange
# 17:16:37 roolebo bhundven: yup. I checked out #288 but haven't built it. Haven't finished review of the changes
# 17:21:38 bhundven roolebo: I'm surprised that "fix" isn't in the binutils-2_25-branch
# 17:30:18 roolebo bhundven: you mean the fix up issue?
# 17:30:25 bhundven yup
# 17:32:16 roolebo bhundven: I agree. We can propose to add other OS X related patches prior to 2.26 release
# 17:32:48 roolebo It's been a year since the last binutils release
# 17:33:11 bhundven I'm just going to backport that one.
# 17:38:19 roolebo bhundven: fixup patch can be ported, but there're two more crosstool-ng patches that start from 3xx to fix darwin build. They're not in binutils master
# 17:39:08 bhundven roolebo: ?
# 17:40:02 bhundven roolebo: I'm confused, because I already added those to ct-ng master
# 17:41:29 bhundven roolebo: see #289
# 17:41:40 bhundven https://github.com/crosstool-ng/crosstool-ng/pull/289
# 17:42:18 bhundven as the 'brits would say: And bob's your uncle...
# 17:43:14 roolebo bhundven: I tested latest snapshot of binutils today, right. I thought it will compile just fine on OS X but had to apply all crosstool's patches to proceed with binutils build. I was just thinking whether it'd make sense to submit them upstream.
# 17:43:58 bhundven roolebo: I say try as you might, half of those patches are in debian and redhat builds...
# 17:44:22 bhundven I'm guessing they tried and were denied
# 17:45:06 roolebo bhundven: I see. I didn't know that
# 17:45:19 roolebo bhundven: thx for #289
# 17:45:26 bhundven yea, I'm merging it now
# 17:46:30 bhundven done
# 17:47:37 bhundven roolebo: that's why after we get #288 reviewed and merged, I need to figure out how to allow local patches to apply to custom sources
# 17:50:10 y_morin bhundven: Basically, in Buildroot we took the position that, if a user is providing custom local sources, it is his responsibility to apply the necesssary patches.
# 17:50:28 bhundven y_morin: Yup, I'm saying the same thing
# 17:50:34 y_morin bhundven: There is virtually no way to know whether the bundled patches would still apply to a custom tree.
# 17:50:46 bhundven but allowing the local patch side to apply, since we only maintain the bundled side.
# 17:50:58 y_morin bhundven: Sorry, you said: how to allow local patches to apply to custom sources
# 17:51:05 bhundven :D
# 17:51:19 y_morin bhundven: Hmm.. I must be misunderstanding things, then...
# 17:51:47 y_morin bhundven: To me, what you wrote means you'd have ct-ng apply our patches on an external source.
# 17:52:00 y_morin bhundven: Hmmm... 'local patches'...
# 17:52:05 bhundven Bill Gatliff says: "THINK!"... Bryan Hundven says: "Read!" ;D
# 17:52:12 bhundven lol
# 17:52:17 y_morin :-)
# 17:52:41 y_morin bhundven: But does that stil lmake sense, either?
# 17:53:12 y_morin bhundven: Really, if someone says "hey, these are the source I want to use" surely we can expect him to provide the full sources, already patched.
# 17:53:56 bhundven or, maybe I'm tracking a development branch and preping ct-ng for a release. Say gcc-6
# 17:54:14 bhundven and I'm tracking my patches in a separate git repo to be for bundled.
# 17:54:38 bhundven don't want to taint upstream branch
# 17:54:47 bhundven makes 'git pull' hard.
# 17:55:22 y_morin bhundven: Well, you branch off upstream branch, you call it 'ct-ng-fixes' and you commit on that branch.
# 17:55:28 bhundven yea, true
# 17:55:39 bhundven I was just writing that
# 17:55:44 y_morin bhundven: When it builds, you run: git format-patch -o /path/toct-ng/patches upstream..ct-ng-fixes
# 17:56:08 bhundven get out of my head y_morin :P lol (jk)
# 17:56:15 y_morin bhundven: It is in my local patches to apply to custom sourcesopinion a much worhtwhile solution.
# 17:56:18 y_morin Hey! ;-)
# 17:56:44 y_morin bhundven: Besides, not patching external sources will make your life much easier, trust me! ;-)
# 17:56:51 bhundven it's true
# 17:57:23 bhundven the goal of the patches directory is that hopefully they would be in the next version, and we could drop that version when it gets old.
# 17:57:34 y_morin bhundven: Because, otherwise, what I understand is you want to support the case where someone is using a non-official (say, gcc) source tree, with custom local patches.
# 17:57:50 bhundven but there are some patches like the check path length in binutils that should be upstream but for some reason or another, isn't.
# 17:58:09 y_morin bhundven: Which means you would have to maintain thos e patches, probably out of a git tree somewhere else
# 17:58:36 y_morin bhundven: And so you'd be duplicating efforts...
# 17:58:39 bhundven is already maintaining a patch tree..
# 17:58:49 y_morin bhundven: Well, you carry it in the ct-ng-fixes branch...
# 17:59:05 y_morin bhundven: Patch-trees are evil! ;-]
# 18:00:34 bhundven y_morin: well, #282 and #288 conflict, so which ever one gets reviewed first the other needs to be rebased.
# 18:01:14 y_morin :-)
# 18:01:23 bhundven then after that, it's cleanup for the rest of the components in the same manner.
# 18:01:25 y_morin OK, so I know which I'll be reviewing first, then! ;-)
# 18:01:34 bhundven bleh
# 18:01:40 bhundven 288? :D
# 18:02:02 bhundven really, either way, it's just as bad
# 18:02:08 y_morin bhundven: But I won;t be able to look at it until tomorrow morning (UTC+1), because tonight I'm debuggng python code here...
# 18:02:16 bhundven fun
# 18:02:28 y_morin Yep, and I don;t know much about python... :-/
# 18:03:17 bhundven well, stop talking to me :D Learn python, it's worth the time.
# 18:03:41 bhundven is learning golang
# 18:05:47 y_morin bhundven: The Go ecosystem is just deeply broken.
# 18:06:03 y_morin bhundven: Building the golang-1.5 compiler needs... a golang compiler already...
# 18:06:24 bhundven y_morin: same with c, or ada, or fortran
# 18:06:32 y_morin (1.5 is written in go,while up to 1.4 it was in C)
# 18:06:43 y_morin bhundven: Well, a C compiler *is* standard.
# 18:06:50 y_morin bhundven: Ada is just weird, I agree.
# 18:06:56 y_morin bhundven: But really, Ada? ;-]
# 18:07:06 bhundven shrugs
# 18:07:20 bhundven what, I can't use javascript to bootstrap gcc c compiler?
# 18:07:26 bhundven :P
# 18:07:26 y_morin bhundven: Are you sure you need a fortran compiler to build fortran?
# 18:07:31 y_morin :-)
# 18:07:44 bhundven y_morin: yes, you need gfortran to build gfortran
# 18:07:53 bhundven same with gnat (ada)
# 18:08:00 bhundven same with gcj
# 18:08:03 bhundven or gccgo
# 18:08:12 y_morin bhundven: I knew about gnat. Did not know about gfortran... Sigh... :-(
# 18:08:12 bhundven worse even
# 18:08:21 y_morin Hmm...
# 18:08:24 bhundven it works best with the same version
# 18:08:42 bhundven I've seen problems building gnat 5.2 with gnat 4.9
# 18:08:51 y_morin Well, at least you don't need python to buld the python interpreter... :-]
# 18:09:01 y_morin hides...
# 18:09:03 bhundven hehe
# 18:10:12 bhundven and actually, I need to add that to configure.ac, is that if gnat is not on the build host, CT_BUILD_HAS_GNAT is not set, and the CC_LANGUAGE_ADA is not an option.
# 18:10:48 bhundven as well as the one for mac os x, that checks in autoconf if the build host can compile static, then libstdc++ and static toolchain can't be enabled.
# 18:11:10 bhundven I see the last one so often
# 18:11:29 bhundven "Why can't I build static on mac?"
# 18:11:30 diorcety quits : Ping timeout: 240 seconds
# 18:13:01 bhundven I digress
# 18:13:40 bhundven need to chug coffee, take a shower and prep for my daughter's arrival. (not in any particular order)
# 18:16:24 y_morin bhundven: I'd still keep the "daughter arrival" item last! ;-)
# 18:16:52 y_morin OMG! Raspberry Pi 0 : $5...
# 18:16:53 bhundven I think the first two are dependencies of the last, so... you're probably right :D
# 18:17:12 y_morin Same CPU as RPI-1, faster, same RAM, smaller form-factor...
# 18:17:23 y_morin https://www.raspberrypi.org/blog/raspberry-pi-zero/
# 18:17:37 y_morin Damn... When will they reach $0 ?
# 18:18:49 bhundven I still need to get a rasp2
# 18:18:56 bhundven I have a b+
# 18:19:15 y_morin bhundven: I have all of them, but the initial model A.
# 18:20:18 bhundven I got one of these http://www.intel.com/content/www/us/en/embedded/products/galileo/galileo-overview.html
# 18:20:32 bhundven has a yocto build
# 18:21:00 y_morin bhundven: We have a defconfig for it in Buildroot now! :-)
# 18:21:09 y_morin bhundven: And it was submitted by an Intel guy! :-)
# 18:21:13 bhundven freescale deprecated my p1020/p2020rdb board, but I still test on it.
# 18:21:30 bhundven e500mc
# 18:21:55 bhundven y_morin: nice!
# 18:22:06 bhundven I need to add my sample for it
# 18:23:28 bhundven y_morin: hey, who does buildroot graphics for the website?
# 18:23:54 bhundven y_morin: I've been asking around for ages to find someone to help with ct-ng logo
# 18:24:25 bhundven I got help with a website rewrite: http://crosstool-ng.github.io/
# 18:24:37 bhundven but it's way out of date and no gfx
# 18:24:39 y_morin bhundven: Last one was done by Angelo: http://git.buildroot.org/buildroot/commit/docs/website?id=97ab48658d574ded87f8af799795ba7ed76f6ba4
# 18:24:56 bhundven looks nice
# 18:25:04 bhundven just noticed :)
# 18:25:27 y_morin OK, off to dinner... Bye!
# 18:25:31 bhundven l84
# 18:25:40 bhundven s/4/r/
# 18:25:43 y_morin :-)
# 18:28:53 bhundven y_morin: hrm, I don't see the galileo in buildroot
# 18:29:45 y_morin bhundven: It's in the 'next' branch.
# 18:29:51 bhundven ah
# 18:30:23 bhundven cool!
# 18:30:46 bhundven gets his board out :D
# 18:31:33 bhundven I use a buspirate to get serial on that one :D I don't have an ftdi cable.
# 18:36:27 bhundven ah, that github kernel tree is based on: http://git.yoctoproject.org/cgit/cgit.cgi/meta-intel-quark/tree/recipes-kernel/linux/linux-yocto-quark_3.8.bb
# 18:59:01 bhundven_ joins #crosstool-ng
# 19:07:05 enunes quits : *.net *.split
# 19:07:07 bhundven quits : *.net *.split
# 19:07:11 wbx quits : *.net *.split
# 19:08:06 muep quits : *.net *.split
# 19:08:06 blueness quits : *.net *.split
# 19:08:27 feepbot quits : *.net *.split
# 19:08:30 wycats quits : *.net *.split
# 19:08:51 ovf quits : *.net *.split
# 19:09:01 xenoxaos quits : *.net *.split
# 19:09:24 devster31 quits : *.net *.split
# 19:09:25 ChanServ quits : *.net *.split
# 19:09:29 Net147 quits : *.net *.split
# 19:09:34 smartin_ quits : *.net *.split
# 19:09:43 philenotfound quits : *.net *.split
# 19:09:44 stefanct quits : *.net *.split
# 19:10:12 aneyman_ quits : *.net *.split
# 19:10:15 imMute^ quits : *.net *.split
# 19:10:28 roolebo quits : *.net *.split
# 19:10:36 y_morin quits : *.net *.split
# 19:12:48 y_morin joins #crosstool-ng
# 19:12:48 wbx_ joins #crosstool-ng
# 19:12:48 stefanct joins #crosstool-ng
# 19:12:48 philenotfound joins #crosstool-ng
# 19:12:48 smartin_ joins #crosstool-ng
# 19:12:48 Net147 joins #crosstool-ng
# 19:12:49 enunes_ joins #crosstool-ng
# 19:12:49 wycats joins #crosstool-ng
# 19:12:49 ovf joins #crosstool-ng
# 19:12:49 muep joins #crosstool-ng
# 19:12:49 blueness joins #crosstool-ng
# 19:12:49 xenoxaos joins #crosstool-ng
# 19:12:49 feepbot joins #crosstool-ng
# 19:12:49 devster31 joins #crosstool-ng
# 19:12:49 ChanServ joins #crosstool-ng
# 19:14:43 aneyman_ joins #crosstool-ng
# 19:14:43 imMute^ joins #crosstool-ng
# 19:14:43 roolebo joins #crosstool-ng
# 19:15:07 roolebo quits : Max SendQ exceeded
# 19:27:36 wbx_ is now known as: wbx
# 19:30:26 bhundven_ is now known as: bhundven
# 19:30:35 bhundven quits : Changing host
# 19:30:35 bhundven joins #crosstool-ng
# 19:39:05 enunes_ quits : *.net *.split
# 19:39:11 devster31 quits : *.net *.split
# 20:17:20 ctngbot joins #crosstool-ng
# 20:47:18 wbx_ joins #crosstool-ng
# 20:47:56 y_morin_ joins #crosstool-ng
# 20:57:15 wbx quits : Ping timeout: 246 seconds
# 20:57:16 y_morin quits : Ping timeout: 246 seconds
# 20:58:35 y_morin_ is now known as: y_morinm
# 20:58:38 y_morinm is now known as: y_morin
# 20:59:22 bhundven yea, freenode is having some issues, and people are being pretty dumb in #freenode
# 21:49:31 ragedragon joins #crosstool-ng
# 22:44:40 y_morin quits : Quit: Nighty Night!
# 22:46:26 ragedragon quits : Ping timeout: 240 seconds
# 23:54:06 roolebo quits : *.net *.split
# 23:54:08 aneyman_ quits : *.net *.split
# 23:54:08 imMute^ quits : *.net *.split
# 23:54:10 bhundven quits : *.net *.split
# 23:54:12 smartin quits : *.net *.split
# 23:54:12 kos_tom quits : *.net *.split
# 23:54:13 enunes quits : *.net *.split
# 23:54:13 Net147 quits : *.net *.split
# 23:54:14 smartin_ quits : *.net *.split
# 23:54:14 philenotfound quits : *.net *.split
# 23:54:14 stefanct quits : *.net *.split
# 23:54:16 muep quits : *.net *.split
# 23:54:16 blueness quits : *.net *.split
# 23:54:17 feepbot quits : *.net *.split
# 23:54:18 wbx_ quits : *.net *.split
# 23:54:18 devster31 quits : *.net *.split
# 23:54:18 wycats quits : *.net *.split
# 23:54:19 ovf quits : *.net *.split
# 23:54:20 xenoxaos quits : *.net *.split
# 23:54:21 ChanServ quits : *.net *.split

Generated by ibotlog2html by Yann E. MORIN