# 00:21:09 |
falstaff_ |
joins #crosstool-ng |
# 00:24:09 |
falstaff |
quits : Ping timeout: 264 seconds |
# 05:07:17 |
djerome |
quits : Remote host closed the connection |
# 06:23:09 |
smartin__ |
is now known as: smartin |
# 06:53:54 |
diorcety |
quits : Quit: Leaving. |
# 07:10:49 |
smartin |
quits : Ping timeout: 246 seconds |
# 07:13:51 |
aalv |
joins #crosstool-ng |
# 07:15:09 |
smartin |
joins #crosstool-ng |
# 07:53:27 |
diorcety |
joins #crosstool-ng |
# 09:07:49 |
sh4rm4 |
quits : Ping timeout: 240 seconds |
# 09:09:57 |
diorcety |
quits : Remote host closed the connection |
# 09:15:48 |
diorcety |
joins #crosstool-ng |
# 09:20:30 |
sh4rm4 |
joins #crosstool-ng |
# 09:51:06 |
aalv |
quits : Remote host closed the connection |
# 10:55:08 |
aalv |
joins #crosstool-ng |
# 11:30:28 |
aalv |
quits : Remote host closed the connection |
# 11:51:54 |
aalv |
joins #crosstool-ng |
# 13:21:58 |
Net147 |
joins #crosstool-ng |
# 13:41:01 |
alan_o |
joins #crosstool-ng |
# 13:48:19 |
Net147 |
quits : Quit: HydraIRC -> http://www.hydrairc.com <- Now with extra fish! |
# 14:23:21 |
doc2 |
joins #crosstool-ng |
# 15:35:09 |
aalv |
parts #crosstool-ng |
# 15:53:58 |
diorcety |
quits : Quit: Leaving. |
# 16:39:48 |
plfiorini |
joins #crosstool-ng |
# 16:51:41 |
diorcety |
joins #crosstool-ng |
# 17:02:24 |
y_morin |
joins #crosstool-ng |
# 17:36:49 |
cipherboy |
joins #crosstool-ng |
# 17:39:45 |
blueness |
quits : Ping timeout: 246 seconds |
# 17:43:22 |
cipherboy |
Question, can anyone point me to a tutorial for building working on crosstool-ng? |
# 17:44:12 |
cipherboy |
Also, is the arm-iphone-linux-gnueabi package for iOS iphone-linux? |
# 17:45:40 |
cipherboy |
Your OS flag points to the linux kernel, but iphone-linux project has more or less disappeared in the past few years. |
# 17:47:18 |
cipherboy |
is now known as: cipherboy-afk |
# 17:48:23 |
plfiorini |
quits : Ping timeout: 256 seconds |
# 17:56:03 |
geckos |
joins #crosstool-ng |
# 17:56:22 |
geckos |
why crosscompile is so tricky.. ? :( |
# 17:56:47 |
geckos |
I'm getting this "configure.in:294: error: automatic de-ANSI-fication support has been removed" |
# 17:57:01 |
dalias |
... |
# 17:57:28 |
plfiorini |
joins #crosstool-ng |
# 18:01:27 |
y_morin |
geckos: 'de-ANSI-fication' and ansi2knr have been irrelevant for more than a decade now. Packages that did not upgrade are really lagging behind, and should be fixed. |
# 18:01:33 |
y_morin |
geckos: what package is that? |
# 18:02:06 |
y_morin |
cipherboy-afk: Yes, we should remove the iphone sample from the release, it is bit-rotting... :-/ |
# 18:03:32 |
geckos |
y_morin, here is the output: http://pastebin.com/cK6s55QK |
# 18:03:46 |
geckos |
I think was trying to build glibc |
# 18:03:50 |
cipherboy-afk |
is now known as: cipherboy |
# 18:04:04 |
cipherboy |
y_morin, k, thanks. |
# 18:05:45 |
y_morin |
geckos: strange, glibc-2.17 is recent, so it is strange it exhibits this issue. Can you post your complete build.log file, please? |
# 18:06:07 |
geckos |
y_morin, I found this: http://sourceware.org/ml/crossgcc/2012-09/msg00138.html so I'm trying with different mpfr version |
# 18:07:13 |
y_morin |
geckos: You can try workaround #2 first. |
# 18:07:20 |
y_morin |
s/can/should/ |
# 18:08:43 |
geckos |
workaround #1 already running, I just go to menuconfig and change the MPFR version seems to pass the error |
# 18:08:59 |
y_morin |
cipherboy: In fact, the iphone sample has been removed a while ago, now. What version of ct-ng are you using? Current is 1.18.0 (as stated in the topic of the chan, and on the website). |
# 18:09:05 |
y_morin |
geckos: OK, good! :-) |
# 18:10:34 |
cipherboy |
Interesting, I thought I pulled the latest, but turns out it is version 1.9.3 |
# 18:11:18 |
geckos |
[CFG ] checking whether autoconf works... no... trying workaround #2 |
# 18:11:36 |
geckos |
lol |
# 18:15:43 |
y_morin |
cipherboy: By default the download dir is sorted alphabetically, not chronologically. |
# 18:15:55 |
y_morin |
geckos: What's your host machine? |
# 18:16:26 |
geckos |
x86_64 |
# 18:17:17 |
cipherboy |
y_morin, my bad, thought I checked for that. Do you have ideas then for compiling for iOS, or a general tutorial you could set me up with to let me play around with and try to get something working? |
# 18:17:19 |
y_morin |
geckos: and host distribution? |
# 18:17:38 |
y_morin |
cipherboy: ping diorcety for iOS stuff. |
# 18:17:52 |
geckos |
archlinux |
# 18:18:25 |
y_morin |
diorcety: Which reminds me this WE was rather busy (being very sunny, I was out almost all day long), and I did not have time to look at your patches. Scheduled for tonight. Sorry... :-( |
# 18:18:53 |
y_morin |
geckos: I don't know of specific issues with arch, AFAIR... |
# 18:19:40 |
diorcety |
y_morin: no problem |
# 18:20:48 |
y_morin |
diorcety: But it looks like those two patch series were mostly good! :-) |
# 18:20:58 |
y_morin |
I don't expect much trouble with them. :-) |
# 18:22:15 |
y_morin |
And it's time for dinner! Yum-yum! |
# 18:27:09 |
doc2_ |
joins #crosstool-ng |
# 18:29:31 |
geckos |
I don't know what to do, so I'm trying system update, I hope my machine boots after this, since I'm on little bad luck today |
# 18:31:07 |
geckos |
What is autoconf version of ppl out there? |
# 18:32:01 |
sh4rm4 |
quits : Remote host closed the connection |
# 18:32:45 |
sh4rm4 |
joins #crosstool-ng |
# 18:42:56 |
Thomas1 |
joins #crosstool-ng |
# 18:46:10 |
geckos |
/usr/bin/bash: line 4: 15216 Segmentation fault (core dumped) ${dir}$tst... OMG is getting worse |
# 19:07:00 |
geckos |
quits : Read error: Connection reset by peer |
# 19:36:59 |
smartin_ |
joins #crosstool-ng |
# 20:01:10 |
cipherboy |
Hey, so once I get a crosstool toolchain built ([65:48]...), could I copy it to another computer, or is it preferable to build it again on each host? Also, thoughts on distccing builds for the target arch? |
# 20:03:25 |
y_morin |
cipherboy: Moving a toolchain across machines is not well supoprted. If you *really* want to do that, you must build a staticaly-linked toolchain: Toolchain options ---> [ ] Build Static Toolchain |
# 20:04:10 |
y_morin |
cipherboy: As for ditccing, I never attempted it. I guess it will be a bit involved, but not undoable... |
# 20:04:17 |
y_morin |
got to go, back in a few minutes... |
# 20:04:55 |
cipherboy |
Thanks, I will just build it again then. |
# 20:09:56 |
smartin_ |
y_morin: about statically-linked toolchain? |
# 20:10:08 |
y_morin |
cipherboy: It depends on how many machines you want to deploy it. A few, you better build it on each machine; more and you may want to build a statically-linked toolchain and copy across machines. |
# 20:10:13 |
y_morin |
smartin_: Yes? |
# 20:10:21 |
smartin_ |
y_morin: hi |
# 20:10:28 |
y_morin |
cipherboy: And alternative is to build it once, and NFS-mount it from all hosts. |
# 20:10:33 |
y_morin |
smartin_: Hey! :-) |
# 20:11:30 |
cipherboy |
y_morin, Just two, both dated by a bit, but I can spend the time to build it on the other once I get this disk copy done. |
# 20:11:33 |
smartin_ |
y_morin, cipherboy: at work i only built statically-linked toolchain, and now ~120 people use them without complaining about anything |
# 20:11:51 |
y_morin |
smartin_: Woohoo! :-) |
# 20:12:00 |
smartin_ |
i'm talking about linux-to-linux toolchain |
# 20:12:18 |
y_morin |
smartin_: What is there out there except Linux? ;-) |
# 20:13:18 |
smartin_ |
for mac-to-linux, i cannot build statically-linked toolchain, so i run a post-build script that render them relocatable |
# 20:13:41 |
y_morin |
smartin_: Hu? The toolchains built by ct-ng *are* relocatable |
# 20:13:48 |
y_morin |
smartin_: If not, it is a bug. |
# 20:14:21 |
y_morin |
diorcety: Hey! Around? |
# 20:14:44 |
smartin_ |
yep, but i don't want to force my mac users to install port and gcc to be able to use my toolchains |
# 20:15:06 |
y_morin |
smartin_: What does that have to do with 'relocatable' toolchains? |
# 20:15:20 |
blueness |
joins #crosstool-ng |
# 20:15:55 |
smartin_ |
y_morin: well, it's more self-contained than relocatable |
# 20:17:01 |
y_morin |
smartin_: OK, what do you mean by 'self-contained'? Do you add stuff (and what stuff) before you deploy them? |
# 20:18:40 |
smartin_ |
y_morin: just libgcc_s and libstdc++ iirc |
# 20:20:08 |
y_morin |
smartin_: Oh, the ones from macports, I guess? How do you manage to get your cross tools (eg. XXX-gcc XXX-ar, XXX-ld et al) to find them at runtime? |
# 20:20:14 |
y_morin |
smartin_: LD_LIBRARY_PATH? |
# 20:21:22 |
smartin_ |
y_morin: i used the same method as qt |
# 20:21:45 |
y_morin |
smartin_: Which is?... |
# 20:22:07 |
smartin_ |
is searching the link... |
# 20:22:11 |
diorcety |
y_morin: yes i am |
# 20:22:38 |
y_morin |
diorcety: I don't understand this CT_CC_FACILITY_LIST stuff? |
# 20:22:41 |
smartin_ |
y_morin: fixing rpath |
# 20:23:34 |
y_morin |
smartin_: OK, so you fidlle with the executables to shoehorn your rpath. Waht tool are you using? |
# 20:24:26 |
diorcety |
y_morin: i use the same code than companion stuff no? :D |
# 20:24:44 |
y_morin |
diorcety: Well, I understand the code, but not what it is for. |
# 20:24:48 |
diorcety |
y_morin: it will be allow to build multiple compiler |
# 20:25:03 |
y_morin |
diorcety: mutliple compilers? At once? |
# 20:25:41 |
diorcety |
y_morin: yes, like this http://pastebin.com/Bmidh7GA |
# 20:26:03 |
diorcety |
y_morin: gcc and llvm-gcc (even it's obsolete) |
# 20:27:09 |
y_morin |
diorcety: Does one requires the other, or are they both stand-alone? |
# 20:27:17 |
y_morin |
*require |
# 20:27:38 |
diorcety |
y_morin: it depends on the compiler for example gcc and llvm are stand alone ... clang depends on some gcc binaries |
# 20:28:47 |
diorcety |
y_morin: at the end gcc is almost mandatory (specially with kernel (for the moment)) |
# 20:28:48 |
y_morin |
does not like that very much... :-( |
# 20:28:55 |
diorcety |
y_morin: i can understand |
# 20:28:55 |
diorcety |
:D |
# 20:29:05 |
y_morin |
diorcety: We're speaking about targetting Darwin, right? |
# 20:29:56 |
diorcety |
y_morin: not specially ... it's generic ... |
# 20:30:52 |
diorcety |
y_morin: what are your doubts? |
# 20:30:59 |
mingwandroid |
joins #crosstool-ng |
# 20:31:43 |
y_morin |
diorcety: I'm not sure we want to introduce sub-components for the compiler. It looks dubious to me. At leat at first sight. |
# 20:32:18 |
diorcety |
sub-components? |
# 20:32:52 |
y_morin |
diorcety: well, you said "multip[le compilers". This means that 'cc' will have sub-components (eg. gcc llvm-gcc...) |
# 20:32:55 |
Thomas1 |
quits : Quit: Thomas1 |
# 20:33:04 |
y_morin |
diorcety: While today 'cc' is a single component. |
# 20:33:22 |
diorcety |
yes |
# 20:33:58 |
y_morin |
really think he should have gone with a dependency-based infrastructure a-la make, rather than a script-based infra... :-( |
# 20:35:25 |
mingwandroid |
could we use comma separated lists or somesuch, e.g. CT_CC=llvmgcc,gcc,clang |
# 20:36:09 |
mingwandroid |
then commas to spaces then for CC in $CT_CC? |
# 20:37:36 |
y_morin |
mingwandroid: Hello! Kconfig does not allow building lists. |
# 20:39:04 |
mingwandroid |
y_morin: hey. shame, but otherwise you don't mind the idea? |
# 20:39:31 |
y_morin |
mingwandroid: So *if* we have to build a list, the way it's done is correct. |
# 20:40:37 |
y_morin |
mingwandroid: No, I'm not opposed to the idea. I just want to get it all well understood, so we don't go at length doing complex changes to the infra for no good reason. |
# 20:41:35 |
smartin_ |
y_morin: i wrote a script that scans the {bin,lib}dirs and fix the executable_path using otool an install_name_tool, similar to http://doc.qt.digia.com/qq/qq09-mac-deployment.html |
# 20:42:32 |
y_morin |
diorcety, mingwandroid: what I find fishy is that your patch builds all 'facilities' in sequence, but pass-by-pass. Ie. gcc-core-1, clang-core-1; gcc-core-2, clang-core-2, gcc-final, clang-final. |
# 20:43:10 |
y_morin |
diorcety, mingwandroid: what says we won't have a case where we want: gcc-core-1, gcc-core-2, gcc-final; clang-core-1, clang-core-2, clang-final ? |
# 20:44:11 |
cipherboy |
is now known as: cipherboy_afk |
# 20:44:50 |
diorcety |
y_morin: each case will work |
# 20:45:15 |
diorcety |
y_morin: clang depends on gcc at RT not compile tiem |
# 20:45:17 |
diorcety |
time |
# 20:45:25 |
mingwandroid |
smartin: I think everyone who's done Mac software has written the same script ;-) ? |
# 20:46:30 |
mingwandroid |
diorcety: in what way does clang depend on gcc? (Yann's the clang master ;-)) |
# 20:46:43 |
smartin_ |
mingwandroid: true! :) |
# 20:48:05 |
diorcety |
mingwandroid: y_morin: if i correcly remember it's for the linking part. llvm/clang haven't his own linker (yet, or maybe in the last version. i'am not up to date) |
# 20:48:32 |
y_morin |
smartin_: mingwandroid No, I am no clang master. I know almost nothing about clang (except what it stands for). |
# 20:48:49 |
mingwandroid |
diorcety: In theory ld64 is the linker, are you saying collect2 is used though? urgh, let's hope it's tidied up by now. |
# 20:48:51 |
y_morin |
smartin_: sorry... |
# 20:49:25 |
mingwandroid |
y_morin: wrong Yann, but I think diorcety is also disclaiming expertise, as am I.. |
# 20:49:34 |
y_morin |
mingwandroid: Hehe! :-) |
# 20:49:58 |
diorcety |
mingwandroid: hum we need to check on this but if i correcly remember clang explicity call gcc for linking ... |
# 20:51:20 |
mingwandroid |
diorcety: y_morin: well, the 3 compilers thing can be reduced to 2 without too much complaint from me since llvmgcc never worked well (just ask CPython or Qt project), but still 2 is > 1. |
# 20:53:15 |
mingwandroid |
.. but I think we should press on and we can re-arrange either by-pass or by-cc-variant as preferred I think? |
# 20:54:21 |
y_morin |
mingwandroid: the by-cc-variant is not really possible with the current infrastructure. |
# 20:56:14 |
mingwandroid |
well, if by-pass can't work for some reason or is really not lliked thenthere's also the option of executing the whole thing twice but being able to share existing build artefacts I suppose? |
# 21:23:44 |
smartin_ |
quits : Quit: leaving |
# 21:29:32 |
cipherboy_afk |
is now known as: cipherboy |
# 21:34:13 |
y_morin |
quits : Quit: Nighty Night! |
# 21:51:29 |
doc2_ |
quits : Remote host closed the connection |
# 22:08:06 |
cipherboy |
is now known as: cipherboy_offlin |
# 22:08:11 |
cipherboy_offlin |
is now known as: cipherboyoffline |
# 23:32:23 |
plfiorini |
quits : Ping timeout: 240 seconds |
# 23:45:28 |
mingwandroid |
quits : Quit: Leaving. |