| 00:21.43 | sjhill | ibot: .sk |
| 00:21.45 | | parse error: dunno what the heck you're talking about, sjhill |
| 00:21.54 | sjhill | ibot: prefix sk |
| 00:21.56 | | sjhill: are you smoking crack? |
| 00:22.02 | sjhill | *sigh* |
| 00:22.44 | ds-work | slovakia |
| 00:32.37 | *** join/#uclibc kergoth (~kergoth@dsl081-228-056.chi1.dsl.speakeasy.net) |
| 00:33.32 | sjhill | thx |
| 01:28.50 | *** part/#uclibc sjhill (~NOYB@207-191-210-241.cpe.ats.mcleodusa.net) |
| 03:36.20 | *** join/#uclibc BZFlag (timr@rikers.org) |
| 05:05.24 | mjn3 | andersee: wcscoll and wcsxfrm now seem to be matching glibc pretty closely... the differences i'm seeing for various locales all seem to be related to one of two glibc bugs involving either weighting of UNDEFINED wchars or not ensuring that position counts are different from the end-of-pass flag |
| 05:06.50 | mjn3 | andersee: so, all that's left for me to do is use a little preprocessor magic to create strcoll and strxfrm. i'm heading for bed now, so i'll do that tomorrow and then get everything checked in |
| 05:16.32 | *** join/#uclibc sandman_ (~sandman@pD9E8AD0C.dip.t-dialin.net) |
| 05:39.12 | *** join/#uclibc bug1 (~bug1@c17156.sunsh3.vic.optusnet.com.au) |
| 08:05.21 | *** join/#uclibc ChanServ (ChanServ@services.) |
| 08:05.21 | *** join/#uclibc bug1 (~bug1@c17156.sunsh3.vic.optusnet.com.au) [NETSPLIT VICTIM] |
| 08:05.22 | *** join/#uclibc jenga (~jim@jenga.stardot-tech.com) [NETSPLIT VICTIM] |
| 08:05.22 | *** join/#uclibc TheGoop (dragon@evrtwa1-ar1-4-47-239-173.evrtwa1.dsl-verizon.net) [NETSPLIT VICTIM] |
| 08:05.22 | *** join/#uclibc TheMasterMind1 (foobar@h-69-3-153-228.MCLNVA23.covad.net) |
| 08:05.22 | *** join/#uclibc parsley_ ([dUzxMdoUN@galadriel.alfred.edu) [NETSPLIT VICTIM] |
| 08:05.22 | *** join/#uclibc andersee (~andersen@codepoet.org) |
| 08:05.22 | *** join/#uclibc kergoth (~kergoth@dsl081-228-056.chi1.dsl.speakeasy.net) |
| 08:05.22 | *** mode/#uclibc [+o ChanServ] by vinge.freenode.net |
| 08:06.19 | *** join/#uclibc kergoth (~kergoth@dsl081-228-056.chi1.dsl.speakeasy.net) |
| 08:08.37 | *** join/#uclibc ChanServ (ChanServ@services.) |
| 08:08.37 | *** join/#uclibc TheGoop (dragon@evrtwa1-ar1-4-47-239-173.evrtwa1.dsl-verizon.net) [NETSPLIT VICTIM] |
| 08:08.37 | *** mode/#uclibc [+o ChanServ] by vinge.freenode.net |
| 08:09.25 | *** join/#uclibc andersee (~andersen@codepoet.org) |
| 08:09.25 | *** join/#uclibc parsley_ ([dUzxMdoUN@galadriel.alfred.edu) [NETSPLIT VICTIM] |
| 08:09.25 | *** join/#uclibc TheMasterMind1 (foobar@h-69-3-153-228.MCLNVA23.covad.net) [NETSPLIT VICTIM] |
| 08:09.25 | *** join/#uclibc jenga (~jim@jenga.stardot-tech.com) [NETSPLIT VICTIM] |
| 08:23.46 | TheGoop | anyone up? |
| 09:03.59 | *** join/#uclibc bug1 (~bug1@c17156.sunsh3.vic.optusnet.com.au) |
| 09:04.27 | bug1 | make dep isnt working, i think i broke it |
| 09:19.54 | bug1 | after make distclean, make dep reports *** No rule to make target `dep' |
| 09:22.56 | andersee | bug1: this is what you changed |
| 09:22.58 | andersee | -busybox: .depend $(libraries-y) |
| 09:22.58 | andersee | +busybox: depend $(libraries-y) |
| 09:23.19 | andersee | But the dep target is: |
| 09:23.22 | andersee | depend dep: include/config.h .depend |
| 09:23.59 | andersee | So you eliminated the dependancy on include/config.h |
| 09:24.48 | andersee | Try changing the .depend target to |
| 09:24.53 | andersee | depend dep .depend: include/config.h scripts/mkdep |
| 09:25.23 | andersee | And remove the old 'depend dep: include/config.h .depend' target |
| 09:26.45 | bug1 | andersee: yea, that fixes it :) |
| 09:30.10 | bug1 | andersee: hmm, i dont think i should need scripts/mkdep at the end as it is run by the .depend target above |
| 09:30.52 | bug1 | actually i misunderstood you i think |
| 09:31.25 | bug1 | -depend dep: include/config.h .depend |
| 09:31.33 | bug1 | +depend dep .depend: include/config.h .depend |
| 09:31.42 | bug1 | that seems to work |
| 09:41.15 | andersee | yup |
| 09:41.32 | andersee | Only woke up since the baby was crying |
| 09:41.38 | andersee | She's asleep now |
| 09:41.46 | andersee | night |
| 10:57.25 | bug1 | bah! Makefile is stufed, dpkg-buildpackage is broken, i cant fix it |
| 13:57.29 | *** join/#uclibc mjn3 (~mjn3@cdm-208-59-243-laft.cox-internet.com) |
| 16:26.46 | *** join/#uclibc sjhill (~NOYB@207-191-210-241.cpe.ats.mcleodusa.net) |
| 16:34.24 | mjn3 | sjhill: morning |
| 16:44.34 | sjhill | hi mjn3 |
| 17:25.10 | *** join/#uclibc ds-work (ds@condor.ucsf.edu) |
| 18:28.55 | andersee_asleep | mjn3: morning |
| 18:29.33 | parsley_ | andersee: sleep in this morning? ;-) |
| 18:31.30 | mjn3 | andersee: hey. strcoll and strxfrm are working now too. i'm fixing the stub locale version of nl_langinfo at the moment. i also need to look into a wprintf buf |
| 18:31.59 | andersee | parsley_: was up quite late with my wife doing christmas stuff |
| 18:32.21 | andersee | mjn3: wohoo! |
| 18:32.55 | andersee | mjn3: Guess I'd better fix handling of crt1 for powerpc and the others |
| 18:40.43 | mjn3 | andersee: the files with the collation rule descriptions are about 2Mb. but you need them to generate the collation data. do you want them in cvs? |
| 18:58.51 | andersee | mjn3: We either put them in CVS, or create a seperate uClibc-locales package (and put that in CVS too) |
| 18:59.02 | andersee | Lets just put it in CVS |
| 19:00.57 | mjn3 | andersee: i think that the pregenerated locale data should be a seperate download though. unlike the collation source tables, i expect the pregenerated tables to change a good bit and the .c source is > 1Mb |
| 19:02.38 | andersee | mjn3: Ok. |
| 19:03.11 | andersee | mjn3: I suppose, since it is generated, we can add an option to either generate it, or download it buildroot style from codepoet.org |
| 19:03.17 | andersee | err. |
| 19:03.19 | andersee | uclibc.org |
| 19:05.54 | mjn3 | andersee: it is generated, but in order to generate it you have to build > 300 glibc locales consuming about 40Mb of space. and i had to tweak some of the locale source to build a few of them |
| 19:06.48 | mjn3 | andersee: if the user wants to do that, then all the have to do (assuming they have the locales) is make headers ; make -C extra/locale ; make |
| 19:16.51 | andersee | k |
| 19:17.15 | andersee | Does downloading buildroot-style sound reasonable? |
| 19:18.40 | mjn3 | i suppose. doesn't matter to me one way or the other. i'll give you a tar of the data and you can put it wherever for download |
| 19:41.08 | mjn3 | andersee: i just put a file pregen-locale-021220.tgz in my home dir on winder. that should be all the data needed, but the Makefile in extra/locale insists on building everything and i don't have time to look at it at the moment. we need to run some errands. back later |
| 20:23.23 | *** join/#uclibc BZFlag (timr@rikers.org) |
| 21:28.29 | *** join/#uclibc bug1 (~bug1@c17156.sunsh3.vic.optusnet.com.au) |
| 21:33.08 | mjn3 | andersee: i'm back |
| 21:40.00 | *** join/#uclibc BZFlag (timr@rikers.org) |
| 21:49.58 | bug1 | waldi: i cant build busybox-cvs from source |
| 21:50.35 | bug1 | 0.60.99.cvs20021209-2 |
| 21:50.56 | bug1 | make[1]: *** No rule to make target `dep'. Stop. |
| 21:51.18 | andersee | mjn3: k |
| 21:51.38 | bug1 | Ive treid all the old Makefile from CVS i get the same thing |
| 21:51.49 | bug1 | well, not all, hehe, the last 4 |
| 21:52.23 | bug1 | im sure they were working... i dunno whats going on |
| 21:55.05 | bug1 | and when i try compiling busybox against uclibc i get |
| 21:55.07 | bug1 | /usr/i386-linux-uclibc/lib/crt0.o(.text+0x13): In function `_start': |
| 21:55.07 | bug1 | : undefined reference to `_fini' |
| 21:55.32 | bug1 | uclibc as provided by the unstable deb |
| 21:56.11 | bug1 | undefined reference to _init as well |
| 22:00.21 | bug1 | i havent tried compiling against uclibc from cvs, so i dont know if it has anything to do with the deb |
| 22:00.53 | mjn3 | andersee: i'm going to try building on nail... |
| 22:17.09 | andersee | mjn3: I need to fixup the powerpc makefile to copy crt0 as crt1 |
| 22:17.28 | andersee | mjn3: Or even better, fix it to use the new ctor setup |
| 22:21.01 | mjn3 | andersee: well, it's building now... |
| 22:21.59 | andersee | mjn3: I hear it |
| 22:22.01 | andersee | :) |
| 22:29.02 | mjn3 | andersee: well, looks like locales build on powerpc. collation is working there too, as well as nl_langinfo |
| 22:33.47 | andersee | mjn3: excellent news! |
| 22:34.17 | andersee | mjn3: Guess its time for me to finish up the last things I've been procrastinating, and work on pushing out a release |
| 22:36.49 | andersee | mjn3: thanks! |
| 22:37.02 | mjn3 | andersee: i noted a couple of bugs i need to fix when i updated my todo list. but we'll be on the road tomorrow |
| 22:37.43 | mjn3 | i wouldn't expect to get to them until after christmas |
| 22:38.03 | mjn3 | andersee: a release with profiling support? |
| 22:40.27 | andersee | mjn3: yeah, I can finish that off |
| 22:40.38 | andersee | mjn3: I've got it mostly working already |
| 22:41.29 | mjn3 | andersee: cool |
| 22:43.11 | mjn3 | andersee: i just built a static hello world. size is about 280k, 250k of which is the locale data |
| 22:49.33 | andersee | mjn3: printf is dragging in the whole pile |
| 22:51.34 | mjn3 | andersee: if you want to get an idea of the locale data supported, there's a little app extra/locale/tst_nl_langinfo.c which will dump out all the supported langinfo entries for a given locale. pretty neat with the UTF-8 locales in an xterm in utf-8 mode with the appropriate fonts |
| 22:52.53 | mjn3 | andersee: yes, but anything using the ctype functions would do it too since all the locale data is in one big structure in a single object |
| 22:54.45 | mjn3 | andersee: currently, if you want to build static apps and you don't want locale support, then you should build them against a stub-locale-configured uClibc. that will change eventually, but i wouldn't look for that until early feb at least. |
| 22:55.51 | andersee | mjn3: We can live with that for now |
| 22:56.24 | mjn3 | andersee: to do it, i need to rewrite the locale structure to put some of the hard-coded values into the structure header for runtime configurability, as well as rework the {w}ctype functions to not use wired-in constants |
| 22:56.53 | mjn3 | andersee: that's stuff on my list, but finishing *wprintf and doing *wscanf come first |
| 23:04.10 | andersee | sure |
| 23:04.16 | andersee | As does making a release |
| 23:04.19 | andersee | :) |