01:04.23 | *** join/#uclibc FireEgl (i=Atlantic@Atlantica.US) |
02:22.29 | CIA-19 | 03landley * r15029 10busybox/util-linux/Config.in: More busting on legacy mtab support in the help text. |
02:53.25 | CIA-19 | 03landley * r15030 10busybox/util-linux/mdev.c: (log message trimmed) |
02:53.25 | CIA-19 | Shrink the code about 50 bytes, allocate less run-time memory, and add a |
02:53.25 | CIA-19 | comment that null terminating the string we sscanf() shouldn't be required |
03:20.53 | CIA-19 | 03vapier * r15031 10busybox/ (include/applets.h networking/udhcp/): add back in udhcp support |
03:24.43 | CIA-19 | 03vapier * r15032 10busybox/ (Makefile networking/Config.in): re-enable udhcp in a few more places |
03:35.32 | *** join/#uclibc Newsome (n=sorenson@adsl-75-0-136-70.dsl.chcgil.sbcglobal.net) |
05:11.57 | *** join/#uclibc landley (n=landley@c-24-3-194-136.hsd1.pa.comcast.net) |
05:12.14 | landley | I just wanted to apologize to Mike Frysinger. |
05:12.32 | landley | (I did so in email, I just feel really bad about overreacting.) |
05:12.50 | landley | When I've been at my laptop for more than about 12 hours, I start to develop a short temper. |
05:13.47 | SpanKY | s/When.*,/ |
05:13.50 | SpanKY | :p |
05:14.22 | landley | I didn't say this was an uncommon occurence. :) |
05:15.48 | landley | The main downside of my otherwise marvelous day job is I'm squeezing busybox development into cracks when I'm otherwise tired. |
05:16.05 | landley | And haven't got the luxury to put it off until I feel like doing it. |
05:16.26 | landley | (I haven't even made it _to_ a coffee shop in... two weeks? Three?) |
05:17.27 | SpanKY | i'm Mike Frysinger btw ;P |
05:18.05 | landley | I suspected that. |
05:18.08 | landley | (So who's vapier?) |
05:18.15 | landley | I know blindvt is bernhard... |
05:18.41 | landley | Ok, /who vapier says SpanKY. |
05:18.45 | landley | That is _so_ not fair. |
05:19.20 | SpanKY | pwnt |
05:21.03 | landley | And now that I've stopped programming, I realize I'm having trouble staying upright. |
05:21.09 | landley | And I have work in the morning. |
05:21.38 | SpanKY | i'm outs |
05:21.49 | landley | ? |
05:23.41 | landley | Rapidly losing hit points. |
05:23.43 | landley | Time for bed. |
05:23.47 | landley | G'night... |
06:29.23 | *** join/#uclibc ashes (n=ashes@modemcable085.56-130-66.mc.videotron.ca) |
06:30.11 | *** join/#uclibc ashes (n=ashes@modemcable085.56-130-66.mc.videotron.ca) |
06:31.01 | *** join/#uclibc ashes (n=ashes@modemcable085.56-130-66.mc.videotron.ca) |
06:31.46 | *** join/#uclibc ashes (n=ashes@modemcable085.56-130-66.mc.videotron.ca) |
06:32.23 | *** join/#uclibc ashes (n=ashes@modemcable085.56-130-66.mc.videotron.ca) |
06:33.10 | *** join/#uclibc ashes (n=ashes@modemcable085.56-130-66.mc.videotron.ca) |
06:33.46 | *** join/#uclibc ashes (n=ashes@modemcable085.56-130-66.mc.videotron.ca) |
06:34.18 | *** join/#uclibc blindvt_ (n=bf@M895P025.adsl.highway.telekom.at) |
06:34.34 | *** join/#uclibc ashes (n=ashes@modemcable085.56-130-66.mc.videotron.ca) |
06:35.23 | *** join/#uclibc ashes (n=ashes@modemcable085.56-130-66.mc.videotron.ca) |
06:36.10 | *** join/#uclibc ashes (n=ashes@modemcable085.56-130-66.mc.videotron.ca) |
06:36.47 | *** join/#uclibc ashes (n=ashes@modemcable085.56-130-66.mc.videotron.ca) |
06:40.59 | *** join/#uclibc ashes (n=ashes@modemcable085.56-130-66.mc.videotron.ca) |
06:46.59 | *** join/#uclibc ashes (n=ashes@modemcable085.56-130-66.mc.videotron.ca) |
09:02.12 | *** join/#uclibc ashes (n=ashes@modemcable085.56-130-66.mc.videotron.ca) |
09:23.29 | *** join/#uclibc psaksa (n=pate@ip212-226-134-137.adsl.kpnqwest.fi) |
11:45.26 | *** join/#uclibc ankur (n=winner@59.95.248.142) |
11:47.17 | *** join/#uclibc Sgt-Donan (n=Loutre@feu30-1-82-242-59-35.fbx.proxad.net) |
11:52.24 | bk_ankur | Hello friends i am having problem in running zebra on uClinux. any one has tested it. |
11:55.29 | *** join/#uclibc tux1800 (n=eric@c66.110.147-33.clta.globetrotter.net) |
12:29.41 | *** join/#uclibc ulf_k (n=ulf_kypk@p54BD16C0.dip0.t-ipconnect.de) |
12:41.02 | *** join/#uclibc prpplague (n=billybob@216.110.117.16) |
12:53.48 | *** join/#uclibc sjhill (n=sjhill@eth13.com-link.com) |
14:38.17 | *** join/#uclibc blindvt_ (n=bf@M978P016.adsl.highway.telekom.at) |
15:08.27 | CIA-19 | 03landley * r15033 10busybox/libbb/Makefile.in: Remove a link that leaked in from the pending llist_t changes. |
15:30.55 | *** join/#uclibc andersee (n=andersee@codepoet.org) |
15:35.49 | CIA-19 | 03landley * r15034 10busybox/miscutils/hdparm.c: |
15:35.49 | CIA-19 | The if() is cheaper than the extra function parameter. According to make |
15:35.49 | CIA-19 | bloatcheck, this saves 112 bytes on x86... |
15:39.36 | *** join/#uclibc ashes (n=ashes@modemcable085.56-130-66.mc.videotron.ca) |
15:55.02 | *** join/#uclibc samrobb-work (n=samrobb_@65-117-135-105.dia.static.qwest.net) |
17:17.16 | *** join/#uclibc Sgt-Donan (n=Loutre@mto30-1-82-242-53-8.fbx.proxad.net) |
18:18.05 | CIA-19 | 03landley * r15035 10busybox/Rules.mak: (log message trimmed) |
18:18.05 | CIA-19 | The changes svn 14960 made to Rules.mak broke building out of tree |
18:18.05 | CIA-19 | (with O=directory), and reverting them fixes building out of tree. I'd be |
18:18.30 | *** join/#uclibc landley (n=landley@65-117-135-105.dia.static.qwest.net) |
18:18.46 | landley | I just reverted the Rules.mak part of 14960, to fix building out of tree. |
18:19.04 | landley | I gave up on trying to understand what it was doing. SpanKY: could you fix it up and re-submit? |
18:20.17 | ashes | i made ssp use arc4random, but it looks like ssp is in ld.so and arc4random is not |
18:20.27 | ashes | so arc4random is undefined |
18:24.52 | solar | yeah. |
18:25.01 | solar | don't use arc4 anymore. |
18:25.21 | solar | pull entropy via the auxv vector if you want to use less system entropy |
18:27.30 | ashes | how about putting an arc4random dummy in ld.so |
18:28.25 | landley | solar: I've got the salt patch queued up for 1.1.3, by the way. |
18:30.36 | CIA-19 | 03landley * r15036 10busybox/scripts/bloat-o-meter: |
18:30.36 | CIA-19 | Tweak suggested by Matt Mackall when I submitted the bloat-o-meter changes |
18:30.36 | CIA-19 | back to linux-kernel... |
18:34.39 | solar | ashes: no. putting ssp in ld.so is touchy enough. |
18:34.46 | solar | just be normal :p |
18:35.06 | solar | landley: your updated one? cuz I already pushed the one from taviso |
18:38.52 | landley | solar: the one from taviso. |
18:39.00 | landley | The updated one isn't -stable material. |
18:39.36 | landley | I should have put taviso's in earlier rather than trying to push the new version faster. |
18:48.38 | solar | landley: then it's already in. |
18:49.06 | landley | In -devel, yes. I added what's in now to the 1.1.3 queue. |
18:49.24 | solar | oh I see. I did not know you were branching all over the place |
18:49.30 | landley | And I'll probably hold off on the passwd rewrite until after 1.2.0 at this point. Needs a longer test period for something that major... |
18:49.38 | landley | The 1.1.x are stable fixes-only releases. |
18:49.47 | landley | 1.2.0 will be the next development stabilization. |
18:49.58 | solar | oh. btw.. getting reports of newest busybox hating ppc |
18:50.02 | landley | I try to comeout with a new stabilization of the development branch every 3 months. |
18:50.04 | landley | solar: darn it. |
18:50.07 | solar | let me see if I can find a report |
18:50.33 | landley | I got ubuntu's knoppix-like cd to boot under qemu PPC, but not the install one... |
18:50.43 | landley | Fighting with the simulated firmware. :( |
18:51.11 | landley | I'm trying to add testing on ppc and arm to my standard checks... |
18:51.32 | landley | We had a couple of endianness issues, but the ones I know about were fixed in 1.1.2, I think... |
18:51.49 | solar | actually. I was wrong. Not ppc but ppc64 (which is not exactly supper supported) |
18:51.59 | landley | Ah. That I have no test environment for. |
18:52.06 | solar | http://bugs.gentoo.org/show_bug.cgi?id=130943 |
18:52.07 | landley | Does it work on x86-64? (That I can sort of test...) |
18:53.33 | solar | ahh n/m looks like a toolchain problem |
18:54.20 | landley | sibling call optimization to `__vmx__sigsetjmp_ent' does not allow automatic |
18:54.20 | landley | multiple TOCs; recompile with -mminimal-toc or -fno-optimize-sibling-calls, or |
18:54.20 | landley | make `__vmx__sigsetjmp_ent' extern |
18:54.20 | landley | I don't understand that at _all_... |
18:54.20 | solar | which is not shocking. ppc64/ppc do a pretty crappy job at maintaining proper toolchains |
18:54.27 | solar | toolchains + kernels |
18:54.48 | landley | I'm just admiring the error message. |
18:54.58 | landley | I've called Garrett (the uclibc++ guy) over to my desk to look at it. |
18:55.03 | landley | I think I may want a copy of that framed. |
18:55.18 | solar | oh I did not know you guys worked together. |
18:55.20 | landley | What the heck is sibling call optimization? |
18:55.26 | landley | He's at the next desk. |
18:55.37 | landley | We knew each other through email for a year or so before winding up at the same company. |
18:55.47 | dalias | landley, looks like broken c++ abi crap... |
18:56.18 | landley | Garrett says it's not c++, and he would know. |
18:56.22 | landley | Looks like setjmp to me. |
18:56.31 | dalias | looks like c++ abi applied to c |
18:56.36 | blindvt` | why on earth is it displaying to build busybox_unstripped *twice*? |
18:56.38 | dalias | gcc developers' favorite passtime |
18:56.48 | landley | The __vmx__ bit? |
18:57.06 | dalias | sigsetjmp should not need any symbols aside from sigsetjmp itself |
18:57.10 | dalias | for a proper implementation |
18:57.22 | landley | Yes, but alas glibc stopped having a proper implementation recently. |
18:57.42 | dalias | yes i imagine they merged it with the c++/exception handling crap... |
18:57.48 | landley | I remember Ulrich Dr. Pepper flaming the User Mode Linux guys for wanting to grab the registers out of the jmp_buf. |
18:57.55 | landley | He broke their ability to do that, and told them to lump it. |
18:58.10 | dalias | well grabbing regs out of the jmp_buf is broken |
18:58.11 | dalias | however |
18:58.25 | dalias | using bloated complex structures for something as simple as setjmp is idiotic |
18:58.29 | dalias | and equally broken |
18:58.33 | landley | *shrug* |
18:59.41 | dalias | glibc setjmp probably depends on malloc now... |
19:00.10 | landley | I'm just not going there. |
19:00.36 | landley | I can't stop glibc from being broken. I can just stick to the standards and make sure it works on something sane (like uClibc) too. |
19:00.58 | dalias | it == busybox? |
19:01.04 | *** join/#uclibc FireEgl (n=FireEgl@Atlantica.US) |
19:01.10 | landley | whatever I'm working on. |
19:01.16 | landley | Often busybox. :) |
19:03.11 | CIA-19 | 03landley * r15037 10busybox/ (5 files in 4 dirs): (log message trimmed) |
19:03.11 | CIA-19 | Fiddling with llist to make memory management easier. Specifically, the |
19:03.11 | CIA-19 | option to delete the contents of the list when we delete the list is a |
19:03.38 | dalias | well yeah glibc brokenness just never ceases to piss me off beyond my wildest expectations :p |
19:04.22 | landley | dalias: I've never met a tool that wasn't broken. |
19:04.25 | landley | Including ones I wrote. |
19:04.28 | landley | Especially ones I wrote. |
19:04.36 | landley | You just have to keep shoveling... |
19:05.10 | dalias | well glibc keeps shoveling in the wrong direction... :) |
19:05.31 | landley | That's why uClibc was invented. |
19:44.19 | landley | I hate makefiles. |
19:44.21 | landley | I really do... |
19:45.28 | landley | My tree is building fine, even after a distclean. |
19:45.46 | landley | But the tree I just anonymously checked out is not. And it's breaking trying to build bb_mkdep. |
19:46.31 | landley | And all the changes in my tree are to C files or the testsuite, so it's got to be some temporary file or permissions or something... |
19:47.25 | prpplague | hehe |
19:52.31 | landley | Well, on the bright side, I think I've figured out what Mike was trying to fix in the change I reverted... |
19:57.08 | blindvt` | by now, this is so fundamentally broken, not funny anymore :( |
19:57.16 | blindvt` | you know, i used to build like this: |
19:57.30 | blindvt` | /tmp/obj$ make -f ../busybox/Makefile top_srcdir=/tmp/src/busybox defconfig ; make |
19:57.59 | blindvt` | i.e. i was in the empty obj-dir and built there. |
19:59.09 | *** join/#uclibc andersee (n=andersee@codepoet.org) |
20:00.43 | blindvt` | erm. top_srcdir=/tmp/busybox does the trick. |
20:00.59 | landley | I had "make defconfig O=otherdir" working for a bit. |
20:00.59 | blindvt` | landley, so how did you "fix" building usage? |
20:01.12 | blindvt` | landley, by reverting vapiers fix, i mean. |
20:01.23 | landley | I reverted the half of 15035 that made it build in my tree. |
20:01.40 | landley | Sorry, 14960. |
20:02.01 | landley | I made sure it built out of tree, and I even did a clean checkout and it built when I tried it. |
20:02.06 | landley | Not a clue what I'm doing differently now. |
20:02.17 | blindvt` | landley, i already said that on the list, but please try this yourself: make an empty objdir, step into it and do: make -f $src/Makefile top_srcdir=$src defconfig |
20:02.39 | landley | Building _in_ the tree stopped working on the clean checkout when I reverted half that fix, though. |
20:02.51 | landley | (Although it still builds in my tree just fine after a distclean, and I have no idea why.) |
20:03.22 | landley | does $src have to be an absolute path here? |
20:03.35 | blindvt` | landley, and see how there is no bb_config.h in the *source-tree* so IT CANNOT BE INCLUDED! no way! Just look at the crap args for building usage |
20:03.44 | blindvt` | landley, yes, it has to be. |
20:04.11 | blindvt` | landley, you reverted vapiers real fix only partial, you still call $(do_link.h) for bb_mkdep. |
20:04.19 | landley | blindvt: I noticed. |
20:04.30 | blindvt` | i'm pretty sick of this by now, sorry |
20:04.31 | landley | I've been poking at putting the half I reverted back, only fixed so it'll build out of tree. |
20:04.42 | landley | blindvt: I try to avoid touching makefiles at all unless forced to. |
20:04.56 | landley | I want this fixed, but this is _not_ my area of expertise. |
20:05.15 | blindvt` | landley, nobody forces you to checkin broken stuff or partially revert real fixes, so that everything is broken. |
20:06.16 | blindvt` | landley, building out of tree is now broken for well over a week, just because you can't be made to turn on threading-mode in you email-client (otherwise you would have noticed that i was talking with vda to trick him into fixing his usage patch) |
20:06.36 | landley | I read each and every email that comes in. |
20:07.28 | blindvt` | landley, maybe, but you rush into checking half-finished stuff in although you would have seen that i was talking care of it so it would work for everyone |
20:08.01 | blindvt` | at least trying to take care. |
20:09.14 | blindvt` | the ML archives iirc do suggest that i was talking to vda, pointing out broken stuff and even sending him an incremental fix which he incooperated while i was looking further, but you just applied the first bit he sent, disregarding the rest of the thread. |
20:09.51 | landley | which svn #? |
20:10.28 | blindvt` | anyway. i'll try to calm down and will do my best to find the time to fix this crap up during the next few days, so i can at least build again |
20:10.40 | landley | So I put the Rules.mak hunk back, and it didn't fix the make -f case either. |
20:10.56 | landley | I'm trying to fix it right now. I hadn't even known it was broken until you pointed it out over the weekend. |
20:11.09 | blindvt` | landley, i don't know. It was the patch wich added usage (the the corresponding thread with the "35k") |
20:11.55 | landley | Why usage couldn't use whatever bb_mkdep.c was using is beyond me. |
20:12.07 | blindvt` | landley, just put back in vapier's fix and add -I$(topp_builddir)/include to the definition of do_link.h and/or compile.h |
20:12.11 | landley | At this point, bb_mkdep.c isn't building in the new checkout either, and I'm trying to find out what changed there... |
20:12.30 | blindvt` | landley, see above, i already told you |
20:13.22 | landley | blindvt:Putting vapier's fix back in didn't fix your make -f invocation. |
20:13.44 | landley | It's still looking for a makefile in the top_srcdir directory on the second make invocation. |
20:14.18 | landley | If you'd told me "that patch broke building out of tree" when it first went in, I'd have happily reverted it until a fixed version was available. |
20:14.20 | blindvt` | landley, right, but putting back in fixes bb_mkdep compilation *proper* and adding -I$(topp_builddir)/include to the build of usage even fixes that one |
20:14.22 | landley | But I don't remember anybody telling me that. |
20:14.53 | blindvt` | landley, i asked you to checkin the fixed-up version, which you didn't do |
20:15.29 | blindvt` | i would have taken care of building out of tree then, but i'm not willing to fix up after you at that rate ;) |
20:15.34 | landley | I was waiting for a delta against what was in the tree, and I had no explanation of what specifically was fixed up about it to give it any urgency. |
20:15.50 | landley | "Here's a patch to make things better". |
20:16.42 | landley | Ok, building in the tree is working again... |
20:16.53 | landley | (Here, I'll check it in once I've tested the other two cases...) |
20:17.23 | landley | I'll need to make bloatcheck work out of tree too, I expect... |
20:19.24 | landley | The make -f invocation is still horked. |
20:20.25 | landley | make O=../woot3 defconfig seems to be working, though... |
20:20.37 | CIA-19 | 03aldot * r15038 10busybox/Rules.mak: - fix build |
20:22.33 | blindvt` | landley, should work now for all and everybody.. |
20:22.40 | blindvt` | g'night |
20:22.43 | landley | aldot just stomped my change. Sigh... |
20:23.02 | landley | Did you fix bloat-o-meter? |
20:23.07 | landley | g'night... |
20:48.14 | CIA-19 | 03landley * r15039 10busybox/Makefile: Make bloatcheck work when building out of tree. |
21:00.26 | *** part/#uclibc carlg (i=clrsrv@www.clearcore.com) |
21:37.12 | *** join/#uclibc gcl (n=gcl@S0106001346b91f0f.cg.shawcable.net) |
22:17.26 | CIA-19 | 03landley * r15040 10busybox/editors/sed.c: Move sed over to the generic llist_t for append. Saves about 90 bytes. |
22:22.07 | *** part/#uclibc khem (n=khem@gateway-1237.mvista.com) |
22:32.20 | *** join/#uclibc khem (n=khem@gateway-1237.mvista.com) |
22:35.14 | *** join/#uclibc andersee (n=andersee@codepoet.org) |
22:39.47 | *** join/#uclibc blindvt_ (n=bf@M1011P031.adsl.highway.telekom.at) |
22:59.00 | *** join/#uclibc Fire_Egl (i=Atlantic@Atlantica.US.TO) |
23:14.50 | *** join/#uclibc prpplague (n=dave@69.73.206.157) |
23:37.49 | *** join/#uclibc sjhill (n=sjhill@eth13.com-link.com) |