ibotlog2html for #crosstool-ng

<< Previous 2012-04-03 Next >>

# 00:32:37 mnt_real quits : Quit: Ex-Chat
# 06:55:31 smartin joins #crosstool-ng
# 09:08:25 ben1066_ joins #crosstool-ng
# 09:10:05 ben1066 quits : Ping timeout: 246 seconds
# 09:13:36 ben1066 joins #crosstool-ng
# 09:15:20 ben1066_ quits : Ping timeout: 246 seconds
# 09:35:59 ben1066 quits : Ping timeout: 246 seconds
# 09:37:17 ben1066 joins #crosstool-ng
# 11:50:58 TheHound dsmith-work: I decided to just ignore it and it seems it did find the svn binary while building the toolchain, just not in the menuconfig
# 11:59:30 ben1066_ joins #crosstool-ng
# 12:00:11 ben1066 quits : Ping timeout: 246 seconds
# 12:00:41 ben1066_ quits : Client Quit
# 12:01:12 ben1066 joins #crosstool-ng
# 12:02:00 ben1066 quits : Changing host
# 12:02:01 ben1066 joins #crosstool-ng
# 13:03:48 ben1066 quits : Read error: Connection reset by peer
# 13:03:55 ben1066 joins #crosstool-ng
# 13:05:00 ben1066 quits : Changing host
# 13:05:01 ben1066 joins #crosstool-ng
# 13:35:59 dsmith-work So I asked this the other day, but there was no one around I guess.
# 13:36:30 dsmith-work I want to exeute ct-ng from a makefile. It doesn't like that. Thinks it's being called recursively.
# 13:37:14 TheHound because it is? I'm new to this but I think I read that ct-ng is a front for make
# 13:37:39 dsmith-work I got it working by running it with env with an option to remove MAKELEVEL from the environment.
# 13:37:51 dsmith-work IS there a better way?
# 13:38:25 TheHound what else do you do inside the makefile that is calling ct-ng?
# 13:39:03 dsmith-work Compliing other stuff, like bootloader, kernel, userland.
# 13:39:49 TheHound hm. those things don't have their own make-file do they? I am thinking of a simple bash script to run make for everything you have to compile
# 13:39:58 dsmith-work Yeah, ct-ng is just a makefile with a shebang
# 13:40:47 dsmith-work TheHound: Well, make runs the whole build. No need to rebuild the toolchain every time if it already has been built.
# 13:40:58 dsmith-work Think BSD packages, or buildroot.
# 13:41:11 dsmith-work s/packages/ports/
# 13:41:22 dsmith-work Yow
# 13:41:54 dsmith-work ctngbot: botsnack
# 13:41:54 ctngbot dsmith-work: thanks
# 13:42:35 TheHound eh true enough
# 13:42:47 TheHound well. In that case I'm afraid I don't know
# 14:56:13 alan_o joins #crosstool-ng
# 15:19:50 ben1066_ joins #crosstool-ng
# 15:22:10 ben1066 quits : Ping timeout: 276 seconds
# 15:24:56 ben1066_ quits : Ping timeout: 246 seconds
# 15:42:03 ben1066 joins #crosstool-ng
# 15:46:38 ben1066 quits : Ping timeout: 246 seconds
# 15:53:46 ben1066 joins #crosstool-ng
# 16:47:00 y_morin joins #crosstool-ng
# 16:54:30 y_morin dsmith-work: hey! Yes, there is a "protection" in the ct-ng frontend against recursion
# 16:56:33 y_morin dsmith-work: I'm lookig at the code right now to refresh my memory...
# 16:59:38 y_morin dsmith-work: correction: the protection is not in the ct-ng frontend, it's in the Makefile to build the crosstool-NG package.
# 17:01:03 y_morin dsmith-work: this is to work around a bug in make: http://savannah.gnu.org/bugs/?20501
# 17:02:36 y_morin dsmith-work: in your case, I woudl suggest that your upperl-layer Makefile calls make with a "sanitised" environment, as you apparently do (eg. env -u MAKEFLEVEL make).
# 17:02:55 dsmith-work Yeah, it's in the Makefile, around line 113
# 17:03:30 y_morin dsmith-work: I hope to be able to fully transition to autotools-based infrastructure in the future, and then it *might* be possible to avoid this...
# 17:03:38 y_morin dsmith-work: yes, that's it.
# 17:03:55 y_morin dsmith-work: and t has nothing to do with the fact that ct-ng is a make-script
# 17:05:02 y_morin dsmith-work: an alternative solution wold be to manually build and install crosstool-NG once and for all on your system, so you would not be bothered by this problem. Then, calling the 'ct-ng' would not be impacted by this limitation.
# 17:07:25 y_morin dsmith-work: as for buildroot (as you talk about it earlier), the crosstool-NG integration explicitly patches ct-ng to disable that check.
# 17:08:17 dsmith-work Ahh.
# 17:08:38 y_morin dsmith-work: the check is here because of auto-completion for make, that calls the Makefile in debug mode to get all the rules, but because the Makefile calls itself back, we have to find a cut at some point, hence the MAKELEVEL limitation.
# 17:08:50 dsmith-work Actually, I am using buildroot, but I'm building ct-ng outside of it.
# 17:09:15 y_morin dsmith-work: for buildroot, ct-ng would never, ever be compiled manually, so we simply disabled the MAKELEVEL check (I know, I did the integration1 ;-) )
# 17:09:29 y_morin dsmith-work: so you are using the toolchain as external toolchains ?
# 17:09:34 dsmith-work Yes
# 17:09:50 y_morin GTB, BBL...
# 17:10:32 dsmith-work Using env is fine. I was just curious if there was something I was missing.
# 17:17:39 y_morin dsmith-work: nope, that is expected behavior.
# 17:17:54 smartin quits
# 17:17:59 y_morin Not that I am very fond of it, but it is a work around to that make bug... :-(
# 17:21:18 y_morin dsmith-work: just for some more info: http://crosstool-ng.org/hg/crosstool-ng/rev/f87f02b00cc7
# 17:27:59 dsmith-work y_morin: And now for something comepletely different.
# 17:28:09 y_morin dsmith-work: shoot! ;-)
# 17:29:02 dsmith-work Using eglibc, I'm not getting the TZ files. I haven't looked to deep yet, but I didn't see an option to crosstoools
# 17:29:30 dsmith-work I'm using ct-ng "in place"
# 17:29:39 y_morin dsmith-work: in my understanding, the TZ files comes from the tzdata package.
# 17:29:59 y_morin *tz-database
# 17:31:01 dsmith-work Oh, and one reason I'm building the complier as part of the build is so a certain toolchain is poart of a release.
# 17:32:16 y_morin dsmith-work: --enable-local is not relevant to the sistuation (at least, it shouldn't! ;-) )
# 17:33:12 dsmith-work True.
# 17:33:51 dsmith-work wanders off in search of food...
# 17:33:57 y_morin dsmith-work: source for the tz database: https://www.iana.org/time-zones/repository/tz-link.html
# 17:34:07 y_morin dsmith-work: enjoy your lunch! :-)
# 17:35:53 y_morin goes bookmarking the TZ database homepage...
# 18:38:05 dsmith-work is now known as: foo`
# 18:38:26 foo` is now known as: dsmith-work
# 18:59:17 dsmith-work Looks like ftp://munnari.oz.au/pub/tzdata2012c.tar.gz (and ftp://munnari.oz.au/pub/tzcode2012b.tar.gz ?)
# 19:09:40 sh4rm4 quits : Ping timeout: 276 seconds
# 19:21:30 sh4rm4 joins #crosstool-ng
# 19:26:03 y_morin dsmith-work: these are the legacy URLs. The current ones are at: https://www.iana.org/time-zones
# 19:26:16 y_morin dsmith-work: sorry for the confusion.
# 19:42:06 dsmith-work y_morin: Thanks!
# 19:42:53 y_morin Cheers.
# 20:52:01 alberthrocks joins #crosstool-ng
# 20:52:43 alberthrocks hey! I'm trying to cross compile the SH3(eb) Newlib, and I'm running into issues with OBJEXT (in configure and Makefile) being undefined.
# 20:52:58 alberthrocks This is, btw, not using your tool - I'm trying to do it myself :)
# 21:29:41 y_morin alberthrocks: ask the newlib guys, maybe. They have a mailing list: http://sourceware.org/newlib/mailing.html
# 21:32:13 alberthrocks oops, missed that, thanks! :)
# 21:32:29 alberthrocks (I tried building sh3eb with your tool, it failed :P)
# 21:47:54 y_morin alberthrocks: AFAICR, crosstool-NG only builds sh4...
# 21:55:56 alberthrocks It's an experimental option, IIRC
# 21:56:11 alberthrocks it tried (and failed) to build sh3eb-unknown-elf for me
# 21:59:42 mnt_real joins #crosstool-ng
# 22:04:35 y_morin alberthrocks: yep experimental means what it means! ;-)
# 22:08:05 alberthrocks heh :)
# 22:09:21 alberthrocks I'm too lazy to contact them right now, so I'm going to sed 'OBJEXT = ' and replace it with 'OBJEXT = o'
# 22:21:19 alberthrocks quits : Read error: Connection reset by peer
# 22:21:44 alberthrocks joins #crosstool-ng
# 23:34:44 y_morin quits : Quit: Night!

Generated by ibotlog2html by Yann E. MORIN