03:53:50 | bug1 | im not sure about the new udhcp patch |
03:54:22 | bug1 | on one hand it easier to maintain because its insync with upstream code, and it provided udhcpd as well |
03:55:02 | bug1 | on the other hand ther may be optimisations that are justifiable for busybox but not for upstream |
05:45:23 | bug1 | i just fixed (but didnt test) uncompress |
05:45:57 | bug1 | re-integrated it i hsould have said |
12:50:35 | andersee | morning Manuel |
12:50:53 | andersee | Did you see davidms commits? |
12:51:00 | mjn3 | morning |
12:52:10 | mjn3 | i saw he made a number of commits. was that to the main repository? i thought not... |
12:53:35 | andersee | I hope not! |
12:53:38 | andersee | Looking now |
12:53:43 | andersee | He deleted 1/2 the lib! |
12:55:52 | andersee | Hmm |
12:56:04 | andersee | Seems to indeed be some other repository. |
12:56:07 | andersee | Phew |
12:58:36 | mjn3 | yes. to the uClinux-dist repository. was just looking at it http://cvs.uclinux.org/cgi-bin/cvsweb/uClinux-dist/uClibc |
13:21:05 | mjn3 | andersee: over the shock yet? |
13:36:30 | andersee | yup |
13:36:54 | andersee | I was a bit worried there for a sec. It was an old script on uclinux.org... Fixed now. |
14:29:21 | prpplague | howdy all, just going to hang out and listen to the convesation ( won't ask any stupid questions) |
15:09:08 | andersee | prpplague: ask away. Its pretty quiet today. |
15:09:42 | prpplague | andersee: nothing specific today, just figured i needed to start listening in |
15:09:56 | prpplague | andersee: i've got start using uclibc more as well as uclinux |
15:11:46 | andersee | k |
15:22:55 | andersee | is trying to track down the lsh and fressh problems reported on the mailing list |
15:29:28 | andersee | mjn3: thoughts on adding isastream() |
15:29:34 | andersee | mjn3: http://www.opengroup.org/onlinepubs/007904975/functions/isastream.html |
15:29:54 | andersee | mjn3: On the one hand, STREAMS are an abomination in the sight of god |
15:30:09 | andersee | mjn3: On the other hand, it is in SuSv3 |
15:34:57 | mjn3 | andersee: looking... |
15:38:59 | mjn3 | andersee: well, it would have to either return 0 or -1 as i can't see adding STREAMS support. i remember something coming up in the mailing list about some app wanting it. the only complication i see is that configure scripts might see it and assume STREAMS support |
15:39:10 | andersee | mjn3: A trivial implementation would be { return 0; } |
15:41:13 | andersee | mjn3: I suppose we could steal the stubs from glibc: sysdeps/generic/isastream.c, sysdeps/generic/getpmsg.c, etc. |
15:42:25 | mjn3 | andersee: agreed. the question becomes, should we return EBADF if the arg isn't a valid file descriptor. i guess it doesn't really matter much though. |
15:42:57 | mjn3 | andersee: (i was agreeing about the trivial implmentation. haven't looked at glibc stubs) |
15:45:53 | andersee | This is the slightly more elaborate glibc stub: |
15:45:56 | andersee | int isastream(int fildes) |
15:45:56 | andersee | { |
15:45:56 | andersee | /* In general we do not have a STREAMS implementation and therefore |
15:45:56 | andersee | return 0. But for invalid file descriptors we have to return an |
15:45:56 | andersee | error. */ |
15:45:56 | andersee | if (__fcntl (fildes, F_GETFD) < 0) |
15:45:58 | andersee | return -1; |
15:46:00 | andersee | /* No STREAM. */ |
15:46:02 | andersee | return 0; |
15:46:04 | andersee | } |
15:46:47 | mjn3 | andersee: works for me |
23:37:35 | | andersee was last seen on #uclibc 7 hours, 51 minutes and 31 seconds ago, saying: } [Fri Sep 27 16:46:04 2002] |
23:37:35 | jacques | ibot: seen andersee |
23:37:45 | jacques | BZFlag: you awake? |
23:37:55 | jacques | and do you have the ability to modify buildroot cvs? |
23:39:28 | mjn3 | jacques: what do you need done? |
23:40:10 | jacques | I'm in final verification, but I tink I got tinyx building again |
23:40:25 | jacques | I would need the tinyx patch modified |
23:41:19 | mjn3 | how big of a change is it? |
23:41:36 | jacques | very small, a sub-patch needs to be removed |
23:42:11 | jacques | essentially the hack for wide char support is no longer needed and was preventing the build |
23:43:03 | mjn3 | ah, wait... which buildroot are we talking about? |
23:43:09 | jacques | the tuxscreen one |
23:43:28 | mjn3 | oh.. sorry. can't help you there |
23:43:32 | jacques | is there another one? :-) |
23:44:12 | mjn3 | erik keeps on to build a uClibc-based user mode linux demo |
23:44:28 | mjn3 | s/ on / one / |
23:44:42 | jacques | oh cool |
23:44:48 | jacques | I didnt know that existed |
23:46:33 | mjn3 | it is in the uclibc.org cvs |
23:47:03 | jacques | cool |
23:48:26 | mjn3 | btw, i'm working towards more complete wide char and locale support by mid Nov |
23:53:39 | jacques | mjn3: excellent |
23:53:59 | jacques | I'm happy that tinyx compiled without complaint about the side char |
23:54:01 | jacques | wide |
23:54:11 | jacques | only thing is a warning at link step of X server |
23:54:24 | jacques | something about some fnuction being hardwired to posix locale |
23:54:30 | jacques | localeconv() |
23:55:17 | mjn3 | the mb<->wc stuff should work ok. currently the stdio stuff isn't multibyte and wide char aware for the most part though, and really only the time functions are locale aware |
23:55:45 | mjn3 | the warning is there to cut down on the "locale doesn't work" posts to the list. ;-) |
23:56:01 | jacques | well I'm sure there are lots of ppl clamoring for the more complete support :-) |
23:57:16 | mjn3 | surprisingly not... only a few have tried, and in its current state it is really only a "proof of concept" to see how small i could get the data |
23:58:05 | jacques | hmm |
23:58:47 | jacques | we;; the data can definitely use size reduction - I think it's like 16MB for a full locale install in glibc |
23:59:04 | jacques | that's just the locale files |
23:59:11 | mjn3 | i'll be cleaning that stuff up some more and implementing collation stuff. |