ibotlog2html for #crosstool-ng

<< Previous 2014-07-26 Next >>

# 00:31:25 Guma quits : Quit: Leaving...
# 00:52:48 mingwandroid quits : Remote host closed the connection
# 01:05:56 djerome joins #crosstool-ng
# 03:19:57 Guma joins #crosstool-ng
# 05:33:14 Guma quits : Quit: Leaving...
# 06:25:39 tlwoerner quits : Ping timeout: 255 seconds
# 06:50:16 diorcety joins #crosstool-ng
# 07:16:16 diorcety quits : Ping timeout: 260 seconds
# 08:05:58 RushPL quits : Ping timeout: 260 seconds
# 08:06:23 RushPL joins #crosstool-ng
# 08:36:16 tlwoerner joins #crosstool-ng
# 09:25:23 y_morin joins #crosstool-ng
# 10:06:02 mingwandroid joins #crosstool-ng
# 10:56:04 maxime__ joins #crosstool-ng
# 10:57:33 maxime quits : Ping timeout: 240 seconds
# 11:04:32 maxime__ is now known as: maxime
# 11:04:34 maxime quits : Changing host
# 11:04:34 maxime joins #crosstool-ng
# 13:22:14 maxime__ joins #crosstool-ng
# 13:25:19 maxime quits : Ping timeout: 272 seconds
# 14:29:04 RushPL quits : Ping timeout: 255 seconds
# 14:29:35 RushPL joins #crosstool-ng
# 15:05:59 maxime__ is now known as: maxime
# 15:06:04 maxime quits : Changing host
# 15:06:04 maxime joins #crosstool-ng
# 15:10:04 mingwandroid quits : Remote host closed the connection
# 15:10:26 mingwandroid joins #crosstool-ng
# 17:06:46 mingwandroid quits : Remote host closed the connection
# 19:38:24 y_morin quits : Ping timeout: 250 seconds
# 19:39:36 y_morin joins #crosstool-ng
# 19:41:57 y_morin_ joins #crosstool-ng
# 19:43:22 y_morin_ parts #crosstool-ng
# 20:21:18 feep quits : Ping timeout: 250 seconds
# 21:56:28 bhundven joins #crosstool-ng
# 22:14:33 y_morin bhundven: OK, to summarise our little private conversation:
# 22:14:42 bhundven :)
# 22:14:48 y_morin 1- add musl as a new C library, without the biggish gcc patch
# 22:15:10 y_morin 2- add the biggish gcc patch to a special location, so it does not get applied by default
# 22:15:25 y_morin 3- add a README along with the patch stating its status
# 22:15:43 y_morin Did I miss anything?
# 22:17:37 bhundven 4- add an experimental Kconfig to enable experimental patches, and a Kconfig entry for each experimental patch. I like the idea of a Kconfig for each entry so that the user explicitly selects the patch, and possible for other experimental patches to depend on others.
# 22:18:08 bhundven but in agreement with 1-3
# 22:18:29 bhundven this also helps with getting llvm stuff into the tree as experimental
# 22:18:30 y_morin Yep, but I was merely thinking about a single config knob to enable experimental patches: it's all or nothing.
# 22:18:59 bhundven musl patches won't interfere with llvm patches, in #4
# 22:19:21 y_morin I was just commenting on your #4, not the llvm stuff.
# 22:20:06 bhundven s/this also/#4 also/
# 22:20:23 bhundven ctngbot: fail
# 22:20:35 y_morin :-)
# 22:20:55 y_morin Yes! We can all blame ctngbot for all our little miseries... ;-]
# 22:21:04 bhundven lol
# 22:21:30 y_morin OK, I'm having a little break for a few minutes. Back shortly...
# 22:25:32 y_morin Back
# 22:29:35 bhundven being able to select the individual patch might also help with the buildroot case ( https://bugs.busybox.net/show_bug.cgi?id=7208 ) where the patch conflicts with glibc: Kconfig: select BR2_PACKAGE_GCC_MUSL_PATCH if ! (BR2_PACKAGE_GLIBC || BR2_PACKAGE_EGLIBC)
# 22:30:20 bhundven s/glibc/(e)glibc/
# 22:30:27 bhundven :)
# 22:31:22 y_morin Well, for Buildroot, we have a completely different way of handling patches, so it does not realy apply.
# 22:32:07 bhundven let me write up my patch, and you (you, irc, and ml) can tell me what *you* think.
# 22:33:03 y_morin bhundven: Well, you basically already have most of #1 in your existing patch, no?
# 22:33:24 bhundven yes, but I have to break it out
# 22:33:33 y_morin Yep.
# 22:33:36 y_morin Then, it's about adding the config knob(s) for #4.
# 22:33:50 bhundven well it should be:
# 22:34:05 bhundven 1) Add EXPERIMENTAL_PATCHES (depends on EXPERIMENTAL)
# 22:34:12 y_morin But the config knob should be generated.
# 22:34:22 y_morin 1: yep
# 22:34:23 bhundven 2) add musl
# 22:34:32 bhundven 3) add #4 knobs
# 22:34:52 bhundven 4) add musl patches
# 22:35:04 bhundven 5) add musl patches readme's
# 22:35:16 bhundven I guess squash 5 and 4 into 4
# 22:35:24 y_morin bhundven: Yep, sounds good. 4 and 5 should go together, though, no?
# 22:35:27 y_morin He!
# 22:35:43 bhundven :)
# 22:35:54 bhundven so, I'm on #1
# 22:36:16 y_morin bhundven: I guess you meant the new #1? Hehe! :-)
# 22:36:29 bhundven y_morin: yes!
# 22:36:50 bhundven is bootstrapping thought, cross-compiling
# 22:37:27 y_morin bhundven: So, you're OK with a single experimental-patches knob, for all experimental patches or none?
# 22:38:07 y_morin bhundven: While you're hacking #1, I'll reply to your original patch on the list.
# 22:38:14 bhundven I'd like to show you my idea [rfc] patch for more finer tuning
# 22:38:21 y_morin Yep, sure.
# 22:38:43 bhundven hard to explain, easy to show code change
# 22:40:12 y_morin Yep, code is sometimes more explicit than speach.
# 22:44:17 y_morin bhundven: BTW, you did not say: are you OK with a single experimental-patches knob, for all experimental patches or none? Or do you stand for a config knob per patch?
# 22:44:55 bhundven I'm going to make the last change enable a knob for each patch, to show the usefulness. But up to that patch will be all or nothing.
# 22:45:19 bhundven if everyone agrees on the last patch, then it can be used too.
# 22:45:40 y_morin bhundven: OK, fine. But then, the knobs should be generated, so we do not have to maintain coherencey between the patches and the config knobs.
# 22:46:09 bhundven another thought, along the same lines...
# 22:46:25 bhundven was to generate the config help from the patch readme.
# 22:46:34 bhundven but maybe the readme contains the knob
# 22:46:41 y_morin Yep, I was also thinking of that! ;-)
# 22:46:47 y_morin (the former)
# 22:47:05 bhundven so the readme is the knob and config help
# 22:47:18 bhundven so you only maintain the patch and readme
# 22:47:58 y_morin bhundven: Better yet:
# 22:48:11 y_morin bhundven: No README, but a description in the patch itself.
# 22:48:57 bhundven I'll look into it
# 22:48:59 y_morin bhundven: The patch gets named like: NNN-musl.patch so we can generate the knob CT_PATCH_EXPERIMENTAL_MUSL
# 22:49:47 y_morin bhundven: And of course, we do not ttrak dpendencies between patches. It's up to the user to handle that, right?
# 22:49:53 y_morin *track
# 22:50:50 bhundven let me step back to the original 1-4, and we'll see how to solve the fine tuned config knobs, when/if we get there.
# 22:54:22 y_morin quits : Ping timeout: 250 seconds
# 22:56:30 y_morin_ joins #crosstool-ng
# 22:57:02 y_morin_ quits : Client Quit
# 22:57:19 y_morin joins #crosstool-ng
# 23:04:58 memfrob y_morin, heyyyyyy!
# 23:05:51 y_morin memfrob: ?
# 23:06:15 memfrob im the only person so far who used or at least documented crosstool-ng on cubieboard / cubietruck
# 23:06:26 memfrob i fucking love crosstool-ng
# 23:06:32 y_morin Lol! :-)
# 23:06:42 y_morin Thanks for telling! :-)
# 23:07:07 memfrob i have no idea why the hell people use embedian or codesourcery though, i honestly dont see the attraction.
# 23:07:22 memfrob crosstool-ng is basically gentoo for a toolchain.
# 23:07:47 memfrob once you switch to git (if you're still working on that) i'd really like to get more involved in the development.
# 23:07:55 y_morin memfrob: Yep, a standard distro is OK for a PC. But for an embedded board? Not so much. Better use something like Buildroot (or even Yocto, biut that's a bit bloated)
# 23:08:08 y_morin memfrob: WE *are* using git now! :-)
# 23:08:13 memfrob WHAT!???????????
# 23:08:25 y_morin memfrob: http://crosstool-ng.org/git
# 23:08:40 memfrob FUCKING NICE!
# 23:09:06 y_morin memfrob: That's a relatively recent switch, however. Something like, a month ago...
# 23:09:30 memfrob ok
# 23:09:47 memfrob plenty of time to stabilize the servers
# 23:09:58 memfrob rock solid now
# 23:09:58 memfrob ?
# 23:10:15 y_morin memfrob: You mean, is the git tree solid?
# 23:10:22 memfrob correct.
# 23:10:36 y_morin memfrob: Yep, it is stable.
# 23:10:46 memfrob keyrings, connection, etc.
# 23:10:53 y_morin memfrob: Well, that's git. What did you expect? ;-)
# 23:11:07 y_morin memfrob: Well, only I have commit access to the tree.
# 23:11:26 memfrob whats the new method for sending in patches?
# 23:11:29 y_morin memfrob: Otherwise cloning is working. You can send patches to the list! ;-)
# 23:11:37 memfrob ah.. email?
# 23:11:48 y_morin memfrob: Yes: git send-email.
# 23:12:10 memfrob ok
# 23:12:32 memfrob :)
# 23:12:39 memfrob thanks y_morin for hoping over!
# 23:12:57 y_morin memfrob: ;-)
# 23:13:13 memfrob sorry if im a bit jumpy..
# 23:13:58 y_morin No problem.
# 23:26:48 y_morin quits : Quit: Nighty Night!

Generated by ibotlog2html by Yann E. MORIN