# 00:03:30 |
saedelaere |
quits : Remote host closed the connection |
# 01:52:15 |
diorcety1 |
joins #crosstool-ng |
# 01:55:29 |
diorcety |
quits : Ping timeout: 265 seconds |
# 04:47:54 |
alan_o |
quits : Quit: Leaving |
# 05:48:15 |
smartin |
quits : Ping timeout: 252 seconds |
# 06:40:05 |
b_rare |
joins #crosstool-ng |
# 07:25:53 |
b_rare |
quits : Quit: leaving |
# 07:26:01 |
diorcety1 |
quits : Quit: Leaving. |
# 07:27:54 |
b_rare |
joins #crosstool-ng |
# 08:34:40 |
saedelaere |
joins #crosstool-ng |
# 09:32:43 |
smartin |
joins #crosstool-ng |
# 15:40:33 |
dyblast |
joins #crosstool-ng |
# 15:43:22 |
mingwandroid |
joins #crosstool-ng |
# 15:44:50 |
dyblast |
hi |
# 15:46:47 |
Martell |
joins #crosstool-ng |
# 15:47:04 |
mingwandroid |
Martell: hey. I got a meeting in 15 mins :-( |
# 15:47:07 |
mingwandroid |
dyblast: hi |
# 15:47:19 |
Martell |
No worries |
# 15:47:27 |
Martell |
I'll email you that build log |
# 15:47:59 |
mingwandroid |
ok, I think that MSYS2 is being a bit weird with not overwriting files and reporting an error, and I think that is all that broke my build. |
# 15:48:22 |
mingwandroid |
Martell: dyblast is Yann, Yann, meet Martell. |
# 15:49:18 |
mingwandroid |
Martell: Did your build then progress to Firefox and do anything there? (Yann, Martell has been trying out the MSYS2 / ctng / OS X stuff for now ..) |
# 15:49:45 |
Martell |
Hey :) |
# 15:49:51 |
Martell |
It got as far as here |
# 15:50:36 |
Martell |
~/ctng-firefox-builds/mozilla-esr24.patched-HEAD-i386-dbg-0616c4e4 ~/ctng-firefox-builds |
# 15:50:36 |
Martell |
~/ctng-firefox-builds |
# 15:50:36 |
Martell |
Configuring, to see log, tail -F /home/Martell/ctng-firefox-builds/mozilla-esr24.patched-HEAD-i386-dbg-0616c4e4/configure.log from another terminal |
# 15:50:36 |
Martell |
real 0m58.737s |
# 15:50:37 |
Martell |
user 0m4.251s |
# 15:50:39 |
Martell |
sys 0m5.389s |
# 15:50:41 |
Martell |
configure failed, see /home/Martell/ctng-firefox-builds/mozilla-esr24.patched-HEAD-i386-dbg-0616c4e4/configure.log |
# 15:53:29 |
mingwandroid |
ok, send me that file too if possible. |
# 15:53:36 |
mingwandroid |
I will kick off a fresh new build now. |
# 15:55:13 |
dyblast |
Martell: hello |
# 15:55:30 |
Martell |
Hi dyblast :) |
# 15:55:41 |
mingwandroid |
bbl |
# 15:55:59 |
Martell |
bbl? |
# 15:58:04 |
dyblast |
Martell: he has a meeting |
# 15:58:30 |
Martell |
*be back later? |
# 15:59:45 |
Martell |
I didn't know what the Abbreviation meant |
# 16:00:04 |
dyblast |
Martell: yes one hour at least |
# 16:09:10 |
falstaff_ |
joins #crosstool-ng |
# 16:11:47 |
falstaff |
quits : Ping timeout: 272 seconds |
# 16:42:09 |
Martell_ |
joins #crosstool-ng |
# 16:44:44 |
Martell_ |
quits : Client Quit |
# 16:45:00 |
Martell |
quits : Ping timeout: 245 seconds |
# 16:45:06 |
Martell_ |
joins #crosstool-ng |
# 17:14:18 |
Martell_ |
quits : Ping timeout: 245 seconds |
# 17:16:52 |
Martell |
joins #crosstool-ng |
# 17:16:55 |
mingwandroid |
hey. |
# 17:17:27 |
mingwandroid |
anyway, that merge request seems fine, I built Firefox again on Arch Linux using the changes. |
# 17:18:54 |
mingwandroid |
Martell: "Building openssl ... cryptlib.c:1:0: error: CPU you selected does not support x86-64 instruction set" <- is that the thing you said you got a fix for? |
# 17:19:25 |
Martell |
no I switched over to msys32 |
# 17:19:28 |
Martell |
and it passed |
# 17:19:57 |
mingwandroid |
ah ok. my msys32 is now building clang so fingers crossed. |
# 17:20:15 |
Martell |
Mine did so hopefully |
# 17:20:17 |
Martell |
:) |
# 17:20:23 |
mingwandroid |
Martell: dyblast: what do you think about commit access directly to our fork? |
# 17:20:53 |
Martell |
I've done a pull request already |
# 17:21:01 |
Martell |
Ahh that would be great |
# 17:21:12 |
Martell |
As long as I don't break it :D |
# 17:21:14 |
y_morin |
joins #crosstool-ng |
# 17:21:53 |
mingwandroid |
Martell: yeah, I saw it. but .. if Yann's (D) ok with giving you direct access then I'm ok with it too. You know llvm better than I anyway. |
# 17:22:37 |
mingwandroid |
Martell: Yann's got buildbot on the go now, so breakages can be found easily and the responsible party punished (usually me) |
# 17:23:24 |
Martell |
I have to run for a meeting |
# 17:23:29 |
Martell |
be back in 20mins |
# 17:23:40 |
Martell |
Otherwise that sounds great |
# 17:37:26 |
dyblast |
quits : Quit: Leaving. |
# 17:48:21 |
mingwandroid |
Martell: "ld: in E:/usr/home/Martell/MacOSX10.6.sdk/usr/lib/libSystem.dylib, file too small for architecture i386" .. indicates a problem with your SDK extraction. |
# 18:00:21 |
mingwandroid |
y_morin: hi. |
# 18:03:18 |
y_morin |
mingwandroid: Hello! |
# 18:08:39 |
Martell |
I will fix that then |
# 18:08:42 |
Martell |
:) |
# 18:08:45 |
Martell |
any try again |
# 18:08:48 |
mingwandroid |
y_morin: hi man, long time since we spoke! |
# 18:09:00 |
Martell |
*and |
# 18:09:33 |
Martell |
quits : Read error: Connection reset by peer |
# 18:09:47 |
y_morin |
mingwandroid: Indeed! Any progress in preparing your changes for upstreaming? |
# 18:10:35 |
mingwandroid |
y_morin: well, this is the final goal, made progress on most other things I think .. |
# 18:11:09 |
mingwandroid |
y_morin: we are keeping track of mercurial reasonably well, Martell has just joined us and his first change was to re-sync .. |
# 18:11:29 |
y_morin |
mingwandroid: Good! :-) |
# 18:12:01 |
mingwandroid |
y_morin: I got asked by the Tor Project to help them out with Tor Browser Bundle cross compilation and that's distracted me these last 2 months a little .. but all for the greater good. I can build Firefox 24ESR and also the latest trunk for OS X on Linux. |
# 18:12:30 |
y_morin |
mingwandroid: Woohoo! :-) |
# 18:12:45 |
y_morin |
knows how cross-building firefox can be a pain... |
# 18:13:03 |
mingwandroid |
that's with Clang, older Firefox can be built with GCC 4.2.1. |
# 18:13:58 |
mingwandroid |
y_morin: Also, I've been doing some MSYS2 stuff .. we've adopted Arch Linux's pacman as a package manager for that, so I can also build Clang Darwin compilers on MSYS2 then commence a Firefox build with them. |
# 18:15:07 |
diorcety |
joins #crosstool-ng |
# 18:15:36 |
mingwandroid |
y_morin: I want to be cleaner than clean when we submit the patches to you, technically as well as legally .. so I am planning to work on a Clang plugin ASAP to do a static coverage thing at compile time to work out exactly what parts of what headers are needed .. |
# 18:15:57 |
y_morin_ |
joins #crosstool-ng |
# 18:16:10 |
y_morin |
quits : Remote host closed the connection |
# 18:16:25 |
y_morin_ |
mingwandroid: Sorry, I may have missed your last comments... :-/ |
# 18:17:46 |
mingwandroid |
(6:13:03 PM) mingwandroid: that's with Clang, older Firefox can be built with GCC 4.2.1. |
# 18:17:51 |
mingwandroid |
(6:13:58 PM) mingwandroid: y_morin: Also, I've been doing some MSYS2 stuff .. we've adopted Arch Linux's pacman as a package manager for that, so I can also build Clang Darwin compilers on MSYS2 then commence a Firefox build with them. |
# 18:17:57 |
mingwandroid |
and re: merging: |
# 18:18:02 |
mingwandroid |
(6:15:36 PM) mingwandroid: y_morin: I want to be cleaner than clean when we submit the patches to you, technically as well as legally .. so I am planning to work on a Clang plugin ASAP to do a static coverage thing at compile time to work out exactly what parts of what headers are needed .. |
# 18:20:49 |
y_morin_ |
long overdue... :-/ |
# 18:20:58 |
y_morin_ |
mingwandroid: OK, sounds good! :-) |
# 18:22:32 |
mingwandroid |
y_morin_: I hope that the review process isn't too bad when we get it done. I know diorcety is keen to get on with that side of things too. |
# 18:24:13 |
y_morin_ |
mingwandroid: I think you could/should post as soon as you have something relativly 'clean', so you do not over-engineer the stuff, or go the wrong direction. |
# 18:26:12 |
mingwandroid |
y_morin_: well, it's kind of tricky, I mean, most of the hard work is to the external projects like clang and LLVM .. and the patches tend to be fairly minor there, somewhat painful to get just right .. |
# 18:27:00 |
mingwandroid |
y_morin_: e.g. I'd not be particularly happy to submit until the compilers can be built on all 3 major build OSes both i686 and x86_64. |
# 18:27:31 |
mingwandroid |
y_morin_: and I *really* don't want to expose any potential legal issues with SDK headers. |
# 18:28:19 |
y_morin_ |
mingwandroid: Would that be legal uncertainty for crosstool-NG, or for the generated toolchains? |
# 18:28:29 |
y_morin_ |
is now known as: y_morin |
# 18:29:10 |
mingwandroid |
y_morin_: well, for exampe one of the cctools patches (Apple's equiv. of binutils) dumps a lot of Apple's SDK headers in there, now to me they'd fall under fair usage for computer interoperability, but I am no lawyer .. |
# 18:30:08 |
mingwandroid |
y_morin: so that's the only real concern, and that's the bit I want to minimise. I want to get that down to the essential parts only, because dumping them all in there goes against the xcode EULA .. |
# 18:30:22 |
y_morin |
mingwandroid: Would it be possible to download the Apple SDK at runtime instead? |
# 18:30:42 |
y_morin |
mingwandroid: and extract whatever we need? |
# 18:31:02 |
mingwandroid |
y_morin: yes, that's the other option, I mean, we already require it for other parts of the build .. but I do some (slight) transformations on the headers too .. |
# 18:31:46 |
y_morin |
mingwandroid: then what about downloading and extracting it, and do your magic at run time? |
# 18:31:47 |
mingwandroid |
but those transformations are done in shell code anyway, so it could be something that I move into ctng. |
# 18:32:23 |
mingwandroid |
we require the user to pass the path to their Darwin SDK, so we can just extract and do slight magic at build time yes. |
# 18:32:39 |
mingwandroid |
by extract I mean cp. |
# 18:32:40 |
y_morin |
mingwandroid: Although I am not a fan of this kind of mangling, it is far better than putting ct-ng in a legal limbo... |
# 18:33:01 |
mingwandroid |
yes, I agree with that. |
# 18:33:02 |
y_morin |
mingwandroid: Oh, the user has to DL the SDK i=himself, we can't downlaod it auto;atically? |
# 18:33:36 |
mingwandroid |
well, it is available on launchpad (search for flosoft), but I don't think even putting a link in there is a great idea. |
# 18:34:00 |
mingwandroid |
I think all legal stuff should be pushed back to the user, even to the extent of having a DL URL. |
# 18:34:24 |
mingwandroid |
Apple don't provide any URL to get this stuff except via their App Store. |
# 18:35:16 |
mingwandroid |
well, I can work on the massage script then if you think that's preferable. The other choice was to minimize the code down then get someone else to clean room re-implement those headers from a english language description. |
# 18:43:51 |
y_morin |
mingwandroid: OK, so automatic download is not possible. So, yes, just have the user point to the pristine SDK, etract whatever you need and munge it from a script in ct-ng. |
# 18:44:02 |
y_morin |
s/etract/extract/ |
# 19:05:45 |
mingwandroid |
yup. |
# 19:43:04 |
saedelaere |
quits : Remote host closed the connection |
# 19:50:02 |
mingwandroid |
parts #crosstool-ng |
# 20:00:50 |
saedelaere |
joins #crosstool-ng |
# 20:06:05 |
roland |
joins #crosstool-ng |
# 20:06:10 |
roland |
is now known as: ruabmbua |
# 20:06:23 |
ruabmbua |
quits : Client Quit |
# 20:07:22 |
ruabmbua |
joins #crosstool-ng |
# 20:09:46 |
ruabmbua |
Hello, i have a little question. Where should i install additional libraries, if i want to use them in my ct-ng toolchain? Probably in ${PATHTOMYTOOLCHAIN}/${CT_TARGET}/sysroot/usr? |
# 20:18:42 |
ruabmbua |
I mean where in the toolchain directory. |
# 20:37:46 |
y_morin |
ruabmbua: IMHO you should not modify the toolchain with additional libraries. |
# 20:38:12 |
y_morin |
ruabmbua: Doing so means you won't be able to use the toolchain for anything else (eg. if you want to use a different set of libraries in the future). |
# 20:38:29 |
y_morin |
ruabmbua: Usually, in this case, you have two options: |
# 20:38:57 |
y_morin |
ruabmbua: 1) use a complete build system such as #buildroot |
# 20:39:35 |
y_morin |
ruabmbua: 2) if you insist on doing manual builds: copy the toolchain, and use the sysroot ot the toolchain as DESTDIR for your additional libraries. |
# 20:40:52 |
y_morin |
ruabmbua: But I would highly suggest you go for a complete build system, it makes things so much easier |
# 20:50:07 |
ray_ |
joins #crosstool-ng |
# 20:57:51 |
ray_ |
y_morin: I've got a question about CT_GetGit |
# 20:59:22 |
ruabmbua |
Thanks y_morin |
# 20:59:35 |
ray_ |
is now known as: mingwandroid_ |
# 20:59:56 |
ruabmbua |
I will use ct-ng, and create a "staging directory". |
# 21:00:21 |
mingwandroid_ |
y_morin: why does it clone to .build/tarballs instead of $HOME/src ? |
# 21:00:22 |
ruabmbua |
This should be enough to build a little ncurses app for arm |
# 21:01:04 |
y_morin |
mingwandroid_: Oh come on, don't change nick, it's disturbing! :-p |
# 21:01:14 |
y_morin |
mingwandroid_: And don't ask to aks, just ask! :-) |
# 21:01:21 |
y_morin |
Ah, OK |
# 21:01:32 |
ruabmbua |
Thanks :D |
# 21:01:38 |
y_morin |
ruabmbua: Cheers! :-) |
# 21:02:02 |
ruabmbua |
quits : Quit: Konversation terminated! |
# 21:02:05 |
y_morin |
ruabmbua: Of course, if you plan to use the toolchain once and trash it, you can do whatever you want. |
# 21:02:09 |
y_morin |
Doh, too late... |
# 21:02:28 |
y_morin |
mingwandroid_: Hold on... |
# 21:03:09 |
mingwandroid__ |
joins #crosstool-ng |
# 21:03:34 |
y_morin |
mingwandroid_: Hold on... |
# 21:04:19 |
y_morin |
no longer knows who he is talking to: ray_, mingwandroid_, or mingwandroid__ ? ;-) |
# 21:05:28 |
mingwandroid__ |
I'm all one in the same. |
# 21:05:42 |
mingwandroid_ |
quits : Ping timeout: 248 seconds |
# 21:06:26 |
mingwandroid__ |
now my ethernet cable fell out most recently :-( |
# 21:06:39 |
mingwandroid__ |
by midnight I will me mingwandroid__________ |
# 21:06:45 |
mingwandroid__ |
*I will be |
# 21:07:10 |
y_morin |
mingwandroid__: And your keyboard is falling apart, too! :-) |
# 21:07:31 |
y_morin |
knows about keyboard-dyslexia... :-] |
# 21:07:38 |
mingwandroid__ |
that is true actually. It overheated because I leave my laptop compiling compilers all night every night |
# 21:07:56 |
mingwandroid__ |
the buttons stopped working so I got a new one, and now it is failing too :-( |
# 21:09:59 |
y_morin |
mingwandroid__:So, the idea behind CT_GetGit was to create a local clone, generate a tarball (with git archive) from that, and store the resulting tarball in ${CT_LOCAL_TARBALLS_DIR} |
# 21:10:59 |
y_morin |
mingwandroid__: Since the local clone would not be used afterwards, and only the tarball we care about, we don't do the clone in CT_LOCAL_TARBALLS_DIR. |
# 21:11:21 |
y_morin |
mingwandroid__: And we should just 'rm -rf' that clone after we've done the tarball. |
# 21:11:50 |
mingwandroid__ |
y_morin: well, for me, I regularly build from scratch, and having tarballs in $HOME/src saves time, but I added git cloning for LLVM and Clang and they take *ages* to clone .. |
# 21:12:05 |
y_morin |
mingwandroid__: Which ultimately means the clone should be done in a temporary place, not in .build/tarballs. |
# 21:12:42 |
y_morin |
mingwandroid__: I understand. But once you have made the tarball once, it sjhould by re-used without re-cloning. |
# 21:13:40 |
mingwandroid__ |
y_morin: ok, but if I am tracking head I generally would like it to pull --rebase each time, but maybe this should be optional? |
# 21:13:49 |
y_morin |
admits the whole download+extract stuff in ct-ng is way too complex, and seemingly borked... |
# 21:14:39 |
y_morin |
mingwandroid__: If you want to actively track HEAD, then you should manage the sources outside of ct-ng, and use CT_CUSTOM_LOCATION_ROOT_DIR |
# 21:15:01 |
mingwandroid__ |
y_morin: how about cloning Git to $HOME/src, then making tarball in .build/tarballs that includes the sha1sum in the tarball filename .. then an option to pull --rebase the git repos and clever detection to avoid remaking tarballs that have same sha1sum as HEAD? |
# 21:16:31 |
y_morin |
mingwandroid__: no, that's not going to work easily. ct-ng is meant to do reproducible toolchains. Once should not use HEAD (or thrunk, or...). But if you want to track such a volatile HEAD, you should point ct-ng to your local copy of the sources that you manage yourself. |
# 21:16:44 |
y_morin |
s/Once/One/ |
# 21:17:21 |
mingwandroid__ |
y_morin: ok fair enough, I will investigate CT_CUSTOM_LOCATION_ROOT_DIR then. |
# 21:18:46 |
mingwandroid__ |
FWIW, my suggestion wouldn't preclude reproducability as CT_GIT_PULL would default to n .. so the sha1 would be same as the tarball, so the tarball would be reused (or if someone did a pull themselves then a new tarball would get made and the sha1sum appended to its name) |
# 21:19:13 |
mingwandroid__ |
but it is fine, I'm happy to go with CT_CUSTOM_LOCATION_ROOT_DIR |
# 21:28:09 |
mingwandroid_ |
joins #crosstool-ng |
# 21:28:46 |
mingwandroid_ |
urgh, how can I work with this laptop :-( |
# 21:29:12 |
y_morin |
mingwandroid_: WiFi ? |
# 21:29:56 |
mingwandroid_ |
yes, I should disconnect the wire that keeps falling out, WiFi would work better, but remember what I was saying about the git cloning of clang and llvm .. that's why I plugged in ;-) |
# 21:31:05 |
mingwandroid__ |
quits : Ping timeout: 272 seconds |
# 21:40:35 |
mingwandroid__ |
joins #crosstool-ng |
# 21:40:37 |
mingwandroid_ |
quits : Ping timeout: 240 seconds |
# 22:11:56 |
saedelaere |
quits : Remote host closed the connection |
# 22:15:17 |
smartin_ |
joins #crosstool-ng |
# 23:00:41 |
smartin_ |
quits : Quit: leaving |
# 23:01:58 |
mingwandroid__ |
quits : Read error: Connection reset by peer |
# 23:03:11 |
mingwandroid_ |
joins #crosstool-ng |
# 23:10:27 |
y_morin |
quits : Quit: /me calls it a day... ;-) |
# 23:13:12 |
rZr |
is now known as: RzR |
# 23:46:59 |
mingwandroid__ |
joins #crosstool-ng |
# 23:48:20 |
mingwandroid_ |
quits : Ping timeout: 245 seconds |
# 23:53:12 |
mingwandroid__ |
quits : Read error: Connection reset by peer |
# 23:58:26 |
mingwandroid_ |
joins #crosstool-ng |