| # 00:31:00 | codyps | quits : Ping timeout: 256 seconds |
| # 00:34:43 | djerome | joins #crosstool-ng |
| # 00:45:47 | codyps | joins #crosstool-ng |
| # 01:33:35 | alan_o | quits : Read error: Connection reset by peer |
| # 01:36:42 | smartin_ | quits : Ping timeout: 264 seconds |
| # 01:48:21 | codyps | quits : Ping timeout: 248 seconds |
| # 02:24:30 | derRichard | quits : Read error: Operation timed out |
| # 02:24:37 | derRichard | joins #crosstool-ng |
| # 06:02:16 | ccole_ | joins #crosstool-ng |
| # 06:04:31 | ccole | quits : Remote host closed the connection |
| # 06:37:53 | djerome | quits : Remote host closed the connection |
| # 06:54:14 | diorcety | quits : Quit: Leaving. |
| # 07:10:37 | aalv | joins #crosstool-ng |
| # 07:24:49 | diorcety | joins #crosstool-ng |
| # 07:38:45 | smartin | joins #crosstool-ng |
| # 07:57:15 | Jacky79 | joins #crosstool-ng |
| # 07:57:38 | Jacky79 | quits : Client Quit |
| # 08:29:02 | ccole | joins #crosstool-ng |
| # 08:29:54 | aalv | quits : Ping timeout: 240 seconds |
| # 08:29:58 | ccole_ | quits : Ping timeout: 240 seconds |
| # 08:30:00 | aalv | joins #crosstool-ng |
| # 08:49:55 | diorcety | quits : Read error: No route to host |
| # 08:51:38 | diorcety | joins #crosstool-ng |
| # 08:52:56 | falstaff_ | joins #crosstool-ng |
| # 08:55:18 | falstaff | quits : Ping timeout: 264 seconds |
| # 11:31:26 | Net147 | joins #crosstool-ng |
| # 13:55:47 | Net147 | quits : Quit: HydraIRC -> http://www.hydrairc.com <- The professional IRC Client :D |
| # 14:19:18 | blueness | quits : Ping timeout: 264 seconds |
| # 14:21:15 | blueness | joins #crosstool-ng |
| # 14:38:48 | alan_o | joins #crosstool-ng |
| # 15:42:07 | doc2 | joins #crosstool-ng |
| # 15:46:56 | diorcety | quits : Quit: Leaving. |
| # 16:17:00 | diorcety | joins #crosstool-ng |
| # 16:31:25 | aalv | parts #crosstool-ng |
| # 17:03:44 | y_morin | joins #crosstool-ng |
| # 17:35:34 | codyps | joins #crosstool-ng |
| # 17:48:55 | xenoxaos | joins #crosstool-ng |
| # 18:02:32 | smartin_ | joins #crosstool-ng |
| # 18:57:58 | codyps | quits : Ping timeout: 268 seconds |
| # 19:12:33 | codyps | joins #crosstool-ng |
| # 19:56:27 | alan_o | quits : Quit: Leaving |
| # 19:56:52 | alan_o | joins #crosstool-ng |
| # 20:13:28 | jevin | quits : Quit: Textual IRC Client: www.textualapp.com |
| # 20:27:14 | smartin_ | quits : Quit: leaving |
| # 21:35:34 | agylytyx | joins #crosstool-ng |
| # 21:36:29 | agylytyx | cannot create directory `/arm-al-linux-gnueabihf' from CT_DoExecLog[scripts/functions@257] |
| # 21:37:14 | agylytyx | what variable do i need to change to set a path prefix for this directory? |
| # 21:38:17 | sh4rm4 | y_morin, ^ |
| # 21:45:14 | y_morin | agylytyx: Hello! This is CT_PREFIX_DIR in sub-menu: Paths and misc options ---> Prefix directory |
| # 21:45:43 | y_morin | agylytyx: It defaults to ${HOME}/x-tools/${CT_TARGET} so it being empty means you modified it. It can't be empty. |
| # 21:45:55 | y_morin | (but that may warrant a test, indeed) |
| # 21:46:26 | y_morin | Hmm... Strange, we *do* have a test for that... |
| # 21:46:32 | agylytyx | I confess I altered it, per someones blog steps |
| # 21:46:41 | y_morin | Doh... Bad blog, bad... |
| # 21:46:52 | y_morin | really does not like blog posts... |
| # 21:47:14 | y_morin | Most of those blog posts I've seen are simply wrong. And some in very dangerous ways... |
| # 21:47:29 | agylytyx | I dont want it in HOME but a mounted path |
| # 21:47:55 | y_morin | agylytyx: Then just enter the path you want to install your toolchain |
| # 21:48:21 | agylytyx | it doesnt edit very easily IIRC, no back arrow support |
| # 21:48:24 | y_morin | agylytyx: the default is in ${HOME} because it is the only place we are *sure* is writeable by the current user. |
| # 21:48:37 | y_morin | agylytyx: Try: ct-ng nconfig |
| # 21:48:49 | agylytyx | kk |
| # 21:48:49 | y_morin | nconfig is an alternate frontend |
| # 21:49:13 | y_morin | agylytyx: There is *one* rule you should abide to: never, *ever* build as root. |
| # 21:49:34 | agylytyx | using a wheel user |
| # 21:49:45 | agylytyx | but not under sudo |
| # 21:49:54 | y_morin | agylytyx: I suggest you do it as a normal user first. |
| # 21:52:13 | agylytyx | its feelingmuch better thx |
| # 21:52:29 | fester | y_morin: did you see my suggestion above? |
| # 21:53:10 | agylytyx | y user is a regular user just in the wheel & adm groups |
| # 21:53:23 | agylytyx | nconfig is nice |
| # 21:54:16 | y_morin | agylytyx: Beware that wheel/adm groups do not have exessive rights (eg. to remove a whole lot of files) in case you foobar your settings. |
| # 21:54:49 | agylytyx | I also putzed with the .config to use 4.8.1 compile built on U12.04 |
| # 21:55:22 | y_morin | fester: the subversion issue? |
| # 21:55:36 | agylytyx | understood |
| # 21:57:01 | y_morin | agylytyx: As with all software, bugs are lurking around for the unsuspecting user to trip in their trap, and bite him hard! ;-) |
| # 21:58:02 | fester | y_morin: yes, please |
| # 21:58:23 | y_morin | agylytyx: Also, editing the .config by hand is *wrong*. Either an option was selected in the menuconfig when it should not, or there is some fixes to do somewhere... |
| # 21:58:46 | y_morin | fester: I was considering making subversion a mandatory dependency, rather than an optional one. |
| # 21:59:22 | fester | y_morin: that's fine too, but it's not obvious the current way |
| # 21:59:46 | y_morin | fester: actually, if you look at the result of ./configure, you'll see subversion was missed: checking for svn... no |
| # 21:59:49 | agylytyx | @y_morin, re svn, please no, git if a vcs must be used |
| # 22:00:13 | y_morin | agylytyx: no, svn it is, as eglibc is using svn. |
| # 22:00:36 | y_morin | agylytyx: and the only official way to get eglibc is with a svn checkout |
| # 22:00:37 | fester | y_morin: then it should probably fail configure |
| # 22:01:12 | y_morin | fester: The idea was to not fail, just hide eglibc in the menu. In retrospect, it seems it fails when using a sample. |
| # 22:01:33 | y_morin | fester: Making it a mandatory dependency means it will then fail. |
| # 22:01:44 | agylytyx | speaking of which i beleive I just hit that issue, sigh "Could not retrieve 'eglibc-2_17'" |
| # 22:01:53 | y_morin | agylytyx: Hehe! |
| # 22:01:58 | fester | install svn! |
| # 22:02:17 | fester | if nothing else just add it to the known issues |
| # 22:02:30 | y_morin | agylytyx: do you think svn should be a hard-dependency, ad fester suggests? |
| # 22:02:41 | y_morin | *as |
| # 22:03:33 | agylytyx | @y_morin ply, it would give a more explicit/relevent err message |
| # 22:04:09 | agylytyx | hmm apt-get install fails |
| # 22:04:55 | fester | apt-get install subversion |
| # 22:06:34 | agylytyx | y just pulled that froma websearch, no svn alias in apt |
| # 22:09:40 | agylytyx | heh, watching ncis on dvd "vicios pixie" used to describe lina hunts char, sounds like a cover band for th gogo's |
| # 22:13:08 | agylytyx | @y_morin back to my puting with the config gcc 4.8.1 is not an available option, the only reasom im building a tc is to use the latest compiler |
| # 22:14:24 | sh4rm4 | y_morin, mandatory svn seems like a bad idea |
| # 22:14:33 | y_morin | agylytyx: then you may want to use the tip of the repository: there is gcc 4.8.1 in there. ct-ng from the repository is pretty stable right now. |
| # 22:14:43 | sh4rm4 | why not just detect if its installed and deny eglibc build without it ? |
| # 22:15:23 | y_morin | sh4rm4: Yes, I'm a bit reluctant at making it mandatory, but eglibc is a major component these days, and we can't really consider it being optional. |
| # 22:15:26 | y_morin | sh4rm4: That's what is done now. |
| # 22:15:29 | sh4rm4 | it's optional for uclibc |
| # 22:15:44 | sh4rm4 | y_morin, but with a horribly misleading error message |
| # 22:15:53 | sh4rm4 | it's not obvious that its due to svn |
| # 22:16:08 | agylytyx | @y_morin kk, if my changes hose me over i'll give that a try, config change seemed pretty straight forward |
| # 22:16:39 | y_morin | sh4rm4: It's not obvious when using a sample, where the check is by-passed. |
| # 22:16:53 | y_morin | sh4rm4: But the check is properly done if you go in the menuconfig. |
| # 22:17:14 | y_morin | sh4rm4: But there's no easy way to do the test when using the sample. |
| # 22:17:20 | sh4rm4 | y_morin, i see |
| # 22:17:21 | agylytyx | sh4rm4 what are the merits of eglibc vs ulibc re rpi? |
| # 22:17:35 | sh4rm4 | eglibc is horribly bloated |
| # 22:17:42 | sh4rm4 | uclibc is much better |
| # 22:17:51 | sh4rm4 | but by far worse than musl libc |
| # 22:18:40 | y_morin | sh4rm4: Fact is, uClibc is losing traction now that "embedded" is getting gebi-bytes of storage and RAM. |
| # 22:18:50 | y_morin | sh4rm4: and musl is not really well known so far |
| # 22:19:06 | sh4rm4 | y_morin, that's changing rapidly atm |
| # 22:19:13 | sh4rm4 | a lot of distros are switching to musl |
| # 22:19:21 | y_morin | sh4rm4: eglibc might be "bloated", but it is "feature-complete" |
| # 22:19:22 | sh4rm4 | probably even CLFS |
| # 22:19:30 | agylytyx | sh4rm4 interesting looks like i should reconfigure/build |
| # 22:19:34 | sh4rm4 | musl is nearly feature-complete |
| # 22:19:38 | sh4rm4 | agylytyx, https://bitbucket.org/GregorR/musl-cross |
| # 22:19:39 | y_morin | sh4rm4: Oh, Debian, Fedora, SuSE, and Co are switching to musl? ;-) |
| # 22:19:48 | sh4rm4 | you can get pre-built cross compilers there |
| # 22:19:56 | sh4rm4 | y_morin, ha ha. |
| # 22:20:23 | y_morin | sh4rm4: just kidding! :-p |
| # 22:20:42 | sh4rm4 | no, alpine, aboriginal, dragora and some others are switching to musl |
| # 22:20:48 | sh4rm4 | probably even gentoo |
| # 22:20:49 | y_morin | sh4rm4: still, musl does not (yet!) have the track-record of glibc/eglibc, and even uClibc. |
| # 22:21:13 | sh4rm4 | a track record of bad maintenance ? |
| # 22:21:15 | sh4rm4 | indeed |
| # 22:21:34 | y_morin | sh4rm4: Ah, if Gentoo is indeed switching (and not only providing it as a non-default alternative), then *that* would be a good starting point. :-) |
| # 22:21:53 | sh4rm4 | problem is, the gentoo guys are kinda slow |
| # 22:22:11 | sh4rm4 | oops, i see some of them are here... |
| # 22:22:13 | y_morin | sh4rm4: the day of bad maintenance in glibc/eglibc world is over now, and things are getting better. |
| # 22:22:32 | sh4rm4 | yeah glibc is *slowly* getting better |
| # 22:22:45 | sh4rm4 | however uclibc's maintainership seems very questionable |
| # 22:22:46 | y_morin | sh4rm4: Being slow on such a critical core component is not a bad idea per-se. |
| # 22:22:51 | agylytyx | lookin to build an edgy rpi, tried gentoo 67 dependecies for ntp lol why does ntp need jpeg? |
| # 22:23:10 | agylytyx | dependencies for layman are equally insane |
| # 22:23:55 | sh4rm4 | y_morin, for example posix_spawn was added to uclibc in march 2012 |
| # 22:24:02 | sh4rm4 | it's a pure *addition* |
| # 22:24:09 | sh4rm4 | i.e. nothing can break by adding it |
| # 22:24:19 | sh4rm4 | still, it's not in the latest release from may 2012 |
| # 22:24:25 | agylytyx | working on Archlinux ATM the sdcard DD bz2 i got doent have a compiler with it, so here I am |
| # 22:24:26 | sh4rm4 | that's a big WTF |
| # 22:24:29 | y_morin | sh4rm4: yes, uClibc is somewhat moribund these days. As I said above, it lost traction ~three years ago with the advent of biggish devices with tons of RAM/flash. |
| # 22:25:32 | sh4rm4 | there's still a ton of embedded devices with less than 16 MB RAM |
| # 22:26:08 | sh4rm4 | musl is perfect for them (my musl-based distribution, sabotage linux, uses *2 MB* by default on amd64) |
| # 22:26:59 | y_morin | sh4rm4: Yet, show an actual x86-64 device with so few storage. |
| # 22:27:12 | sh4rm4 | i mean RAM |
| # 22:28:03 | sh4rm4 | i'd rather have 16382 of my 16384 MB ram available for stuff i'm actually running myself, instead of "boilerplate" code |
| # 22:28:30 | y_morin | sh4rm4: 16MiB RAM on an x86_64 device? You must be kidding! |
| # 22:28:37 | sh4rm4 | 16GB |
| # 22:28:38 | y_morin | Oh, 16GiB.. |
| # 22:29:30 | y_morin | Well, even for 16GiB, why bother a few MiB? It's cheaper to add RAM that re-qualify a new software stack. |
| # 22:29:40 | sh4rm4 | few ? |
| # 22:30:39 | sh4rm4 | if i start ubuntu 13.04 on i386, 450 MB are used for a plain desktop |
| # 22:30:49 | sh4rm4 | the difference from 2 -> 450 MB is not what i call little |
| # 22:30:58 | y_morin | Considering the kind of software running on x86-64 (eg. server-class or workstation-class machines), the RAM used is not in the C library, but in the software (eg. desktop, http server...) |
| # 22:31:19 | sh4rm4 | and musl considerably lowers that memory usage |
| # 22:31:34 | y_morin | sh4rm4: Try running Unity+firefox+... What is eating this memory, do you think? The C lbrary, or GTK? |
| # 22:31:36 | sh4rm4 | because it uses nearly no .data |
| # 22:31:56 | sh4rm4 | and does very efficient mapping of shared libraries |
| # 22:32:15 | sh4rm4 | y_morin, the full LXDE destop started on sabotage uses 12 MB |
| # 22:32:25 | sh4rm4 | and that including shared gtk+2 and glib dsos |
| # 22:34:08 | sh4rm4 | the startup overhead when using glibc is bigger for a *single binary* |
| # 22:34:14 | sh4rm4 | it's the epitome of bloat |
| # 22:36:49 | sh4rm4 | admittedly, sabotage uses Xfbdev |
| # 22:37:03 | y_morin | Doh... eglibc is *not* guarded by HAS_svn... Sh!t... :-( |
| # 22:38:05 | sh4rm4 | http://ftp.barfooze.de/pub/sabotage/ in case you're bored, you can fetch sabotage qemu images there and try it yourself |
| # 22:38:18 | y_morin | wonders how we lost the eglibc dependency on svn... :-( |
| # 22:39:22 | agylytyx | @y_morin /aaa/work/tools/arm-al-linux-gnueabihf/arm-al-linux-gnueabihf/lib': File exists looks like it is missing -p |
| # 22:39:51 | agylytyx | on the mkdir *sigh* |
| # 22:48:37 | y_morin | quits : Quit: Nighty Night! (very sleepy...) |
| # 23:01:29 | doc2 | quits : Remote host closed the connection |