00:13.47 | *** join/#uclibc Newsome (n=sorenson@d53-64-13-210.nap.wideopenwest.com) |
00:32.38 | SpanKY | heh |
00:32.40 | SpanKY | nite ! |
01:18.26 | CIA-10 | 03vapier * r14252 10busybox/docs/style-guide.txt: add exception for uncuddling brackets |
01:21.39 | *** join/#uclibc landley (n=landley@pool-70-17-177-38.pitt.east.verizon.net) |
01:21.42 | landley | Stroppy! |
01:24.43 | SpanKY | wanker |
01:27.02 | landley | You try rearchitecting sed to be happy with embedded nul bytes. |
01:27.34 | landley | Dismantling the special casing for optional \n characters along the way and replacing it by propogating the read length to every usage point of the string. |
01:31.53 | landley | (Stroppy!) |
01:46.00 | CIA-10 | 03landley * r14253 10busybox/testsuite/seq.tests: Add a test suite for seq. |
01:50.32 | *** join/#uclibc basit (n=abasit@linuxpakistan/basit) |
01:50.41 | basit | is auto_ptr supported by ucLibc? |
02:00.57 | SpanKY | wtf is auto_ptr |
02:14.43 | *** join/#uclibc thraxisp (n=thraxisp@ottawa-hs-64-26-170-144.d-ip.magma.ca) |
02:17.23 | solar | c++ thing |
02:23.13 | solar | http://www.gotw.ca/publications/using_auto_ptr_effectively.htm |
02:30.42 | CIA-10 | 03landley * r14254 10busybox/ (coreutils/sort.c editors/sed.c libbb/get_line_from_file.c): |
02:30.42 | CIA-10 | A few changes falling out from the effort to make sed handle embedded NUL bytes. |
02:30.42 | CIA-10 | Checking in to reduce the diff between my tree and svn... |
02:35.04 | dalias | i don't see how the _c_ library would support auto_ptr |
02:35.06 | dalias | it's a c++ feature |
02:45.36 | *** join/#uclibc landley (n=landley@pool-70-17-174-215.pitt.east.verizon.net) |
02:45.57 | landley | And now irc is being stroppy. |
02:46.00 | landley | It's that kind of week. |
02:48.51 | CIA-10 | 03landley * r14255 10busybox/docs/busybox.net/subversion.html: Mention patches directory as a read-only svn alternative. |
02:55.26 | CIA-10 | 03landley * r14256 10busybox/docs/busybox.net/ (docs.html header.html): Move FAQ under docs, add programming.html to docs list. |
03:12.05 | CIA-10 | 03landley * r14257 10busybox/include/libbb.h: Missed one in the get_line_from_file() update... |
03:13.48 | *** join/#uclibc landley1 (n=landley1@pool-70-20-97-48.pitt.east.verizon.net) |
03:15.02 | *** join/#uclibc landley (n=landley@pool-70-20-97-48.pitt.east.verizon.net) |
03:15.05 | landley | That's better... |
03:15.35 | CIA-10 | 03landley * r14258 10busybox/ (e2fsprogs/ext2fs/getsize.c shell/hush.c): |
03:15.36 | CIA-10 | I was playing around with a dependency generator a couple weeks ago and I found |
03:15.36 | CIA-10 | some #include files that aren't in the tree. No biggie... |
03:16.05 | landley | Ha, and finally I've got the superflous crap in my tree either reverted or checked in. |
03:16.32 | landley | Now to get back to banging my head against sed.c... |
03:18.56 | ashes | beryllium nickel alloy has a tensile strength of 300,000 psi. its 375% stronger than the best high tensile steel |
03:19.07 | landley | Cool. |
03:19.21 | landley | I tend to go for silicon nitride and carbon nitride myself... |
03:19.30 | landley | Where does one get berylium nickel alloy? |
03:19.40 | landley | (Vapor deposition of diamond's pretty cool too.) |
03:19.52 | landley | I want that for my glasses, but they won't do it in the US. |
03:20.12 | ashes | i found berylium for sale at $20 per gram |
03:20.20 | ashes | 99.9% pure |
03:20.33 | ashes | then just get someone to mix it with nickle |
03:21.34 | ashes | beryllium has a 1287 ╟C melting point |
03:21.46 | ashes | and its lighter than aluminum |
03:22.13 | ashes | its toxic though |
03:23.03 | ashes | beryllium dust inhaled causes a disease. a scratch from a peice of beryllium causes a rash. it also amplifies alpha radiation |
03:23.31 | landley | Darn it, it's snowing. |
03:23.37 | landley | That's pittsburgh for you... |
03:23.58 | ashes | 245,000 psi yeild capacity |
03:25.01 | ashes | ming |
03:25.28 | ashes | beryllium nickel alloy is only 2% beryllium, 97% nickel |
03:25.56 | ashes | pure beryllium would have several million psi load capacity |
03:30.43 | landley | I'm not sure it scales quite that way. |
03:30.50 | landley | Take irridium: strong but really brittle. |
03:35.15 | ashes | pure beryllium tensile strength is only 58,000 psi |
03:35.25 | ashes | 98.5% pure |
03:36.05 | landley | There's a reason people alloy things. :) |
03:36.48 | landley | The beryllium in nickel sounds like the carbon in steel. |
03:36.51 | landley | A little goes a long way. |
03:38.19 | ashes | carbon has the highest tensile strength of any material we know of. beryllium is much lighter though |
03:38.29 | landley | Carbon can. |
03:38.31 | landley | Soot doesn't. :) |
03:38.46 | landley | And the tensile strength of carbon dioxide is left as an exercise to the reader. |
03:38.58 | landley | (Well, you can freeze it...) |
03:39.25 | landley | And in _theory_, alloying diamond could help it just as much as alloying steel does. |
03:39.45 | landley | Stick stuff in there to break up the crystaline fracture planes diamond cutters take advantage of. |
03:39.59 | landley | (Presumably, more fun with vapor deposition?) |
03:40.23 | landley | A honeycomb is generally stronger than the underlying solid material... |
03:41.04 | ashes | beryllium has a hexegon structure |
03:41.22 | landley | In 2d or 3d? |
03:41.33 | ashes | im not sure |
03:42.03 | landley | Graphite has a marvelous 2d structure. Unfortunately in 3d it fractures darn easily into little plate things... |
03:42.22 | ashes | a hex is the second strongest shape, i think, under the sphere/circle |
03:42.32 | landley | Diamond has a repeatable 3d structure with the tetrahedral bonds and all. |
03:42.50 | landley | In theory silicon can do the same thing. Possibly that's what granite is, I've never looked it up. |
03:43.03 | *** part/#uclibc Newsome (n=sorenson@d53-64-13-210.nap.wideopenwest.com) |
03:43.06 | landley | (And germanium, but I doubt anybody would really care.) |
03:45.15 | landley | I spent a summer researching metals, back in high school. |
03:45.49 | landley | Van Nostrand's scientific encyclopedia can be remarkably entertaining under the right circumstances... |
03:50.37 | ashes | :) |
03:53.53 | landley | (Writing sed test corner cases.) |
03:53.57 | landley | (sed -i with no filename.) |
03:54.21 | landley | Multiple files and the $ last-line indicator. |
03:54.25 | landley | With and without -i... |
04:32.07 | *** join/#uclibc SpanKY (n=UserBah@pool-141-154-55-220.bos.east.verizon.net) |
05:30.11 | *** join/#uclibc Laggy_McGee (n=jchadwic@pool-71-245-114-103.cmdnnj.fios.verizon.net) |
05:30.39 | Laggy_McGee | So, I created the filesystem, and copied to my CF disk... Now what? |
05:30.59 | Laggy_McGee | What kernel do I use, where do I put it...? Where is the installation/booting documentation? |
05:35.54 | SpanKY | uhh there is none |
05:36.03 | SpanKY | that's *highly* target dependent |
05:36.06 | SpanKY | and that's your job |
05:36.26 | Laggy_McGee | hmmm |
05:37.02 | Laggy_McGee | there's nothing about using GRUB on an IDE HDD on an x86?? |
05:37.38 | SpanKY | go read grub documentation |
05:58.48 | dalias | yay, i figured out how to make environ semantics fully correct without weak symbols |
06:11.25 | *** join/#uclibc tty56 (n=johannes@p54A7A224.dip0.t-ipconnect.de) |
07:12.22 | *** join/#uclibc blindvt__ (n=bf@M961P021.adsl.highway.telekom.at) |
07:14.51 | *** join/#uclibc blindvt (n=bf@M988P010.adsl.highway.telekom.at) |
07:24.20 | *** join/#uclibc woody___ (n=woody@bi01p1.nc.us.ibm.com) |
07:38.30 | *** join/#uclibc blindvt_ (n=bf@M986P001.adsl.highway.telekom.at) |
08:02.50 | *** join/#uclibc vrm (n=vrm@151.55.101-84.rev.gaoland.net) |
08:27.24 | *** join/#uclibc mjn3 (n=mjn3@c-71-196-255-107.hsd1.co.comcast.net) |
08:44.22 | *** join/#uclibc psaksa (n=pate@ip212-226-134-137.adsl.kpnqwest.fi) |
08:53.44 | *** join/#uclibc vodz (i=1000@dzo.simtreas.ru) |
08:54.07 | vodz | http://drugoi.livejournal.com/1699622.html |
08:54.13 | vodz | Moscow 23 Feb 2006 |
08:56.19 | vodz | In this photos have not russians, caucasian or near east only |
08:56.38 | blindvt` | vodz, i can trigger a cornercase where strlen goes into an infinite loop. Can't we just use __builtin_strlen in bb_strlen() ? For the non-IMA case, we can -fno-builtin-strlen, but i'm not sure if you intended to have strlen builtin in bb_strlen or if you want to use the strlen from libc there. |
08:58.17 | vodz | blindvt`: again infinite loop? Hmm. May be for CONFIG_BUILD_AT_ONCE only |
08:58.59 | blindvt` | vodz, yes, it's IMA only. |
08:59.23 | vodz | blindvt`: Ohh, i be afraid of this :( |
08:59.39 | vodz | blindvt`: -fno-builtin-strlen Ah, good idea! |
09:03.26 | vodz | blindvt`: hmm. But why infinite loop? I can`t understand withg quick analize. :( Before xfunc used strlen->bb_strlen, after xfunc use builtin strlen... |
09:06.58 | vodz | haw translate (form bb mailist): nano forever? nano is 10e-9? |
09:07.02 | vodz | how |
09:09.20 | blindvt` | vodz, nano is an editor. (pico, nano, joe, etc) |
09:09.38 | vodz | aha |
09:10.03 | vodz | About, give then in to cultivate a complex of fault before east people. The Europe has well succeeded in it. And still it is possible to arrange a genocide to itself, that to them here would be is better, aha. People certainly it is a pity. But I do not see sense in поднимании similar politcorrect that. |
09:10.37 | vodz | поднимании = lionize |
09:11.00 | vodz | from comment to URL drogoi... |
09:12.03 | vodz | traslate bu mashine |
09:14.26 | blindvt` | vodz, we do want to use the strlen() from libc, yes? If so, we can generally try if CC accepts -fno-builtin-strlen and all should be well, ok? |
09:16.10 | blindvt` | vodz, well all would be well then even for IMA then if we'd move all strlen() in the applets to use bb_strlen() that is. |
09:16.48 | vodz | blindvt`: for me, bb_strlen must die. before my patch in IMA mode my gcc 3.3.3 generate warning: strlen undeclared, may be #define strlen(s) bb_strlen(s) hide strlen->__.... a declaration |
09:17.31 | blindvt` | vodz, ah! The better if we just remove it. Will you do this, please? |
09:19.09 | *** join/#uclibc quitte__ (n=quitte@stgt-d9be1310.pool.mediaWays.net) |
09:20.41 | vodz | blindvt`: The current maintainer chase us a urinate cloths. ;-) |
09:21.49 | vodz | blindvt`: can you rename "blindvt_" to other? b<tab> is not work, must typed full :( |
09:22.06 | blindvt` | vodz, is there any fundamental reason why we have bb_printf, by the way? IIRC it increased the size of an applet compared to using plain printf |
09:23.24 | blindvt` | vodz, i can't rename him currently, sorry. Just talk to "him" instead of me, epic doesn't highlight anything anyway :) |
09:25.07 | blindvt` | vodz, i don't quite understand what you mean with the current maintainer and our clothes.. You mean he is intimidating? |
09:27.07 | vodz | blindvt`: bb_printf by mj3, he is more made. |
09:27.56 | blindvt` | mjn3, ping |
09:28.16 | dalias | bb_printf is evil :/ |
09:28.43 | blindvt` | vodz, ok. Will you take care of removing the bb_strlen wrapper throughout the tree soon? |
09:29.42 | vodz | blindvt`: I think now: may be set #define strlen(s) bb_strlen(s) after bb_strlen(..) { } body remove bug in IMA mode |
09:30.41 | vodz | but can`t test it. I can`t show problem always |
09:31.53 | blindvt` | vodz, i saw it with mount -t nfs. |
09:32.36 | *** join/#uclibc woody__ (n=woody@129.33.1.37) |
09:33.47 | vodz | blindvt`: hmm. Also. For remove problem with require -o nolock for mount -t nfs - require portmap. But in mailist have wrong opinion: "portmap require for NFS servers only" |
09:34.59 | blindvt` | vodz, i'll try that. If it doesn't work then i suggest we just ex - $(find ./ -name "*.c") g/bb_strlen/s//strlen/g |
09:36.12 | vodz | blindvt`: you say: find . -name "*.[hc]" | xargs sed g/bb_strlen/s//strlen/g ? |
09:40.21 | dalias | is there any chance of getting a patch accepted so that busybox doesn't have to munge with the internals of FILE* for bb_*printf? |
09:42.33 | vodz | dalias: may be wait before mj3n awake |
09:44.34 | dalias | wow excellent |
09:44.35 | dalias | 11720 dalias 16 0 64 60 44 S 0.0 0.0 0:00 busybox |
09:48.19 | blindvt` | vodz, i admit i normally say for i in $(find ./ -name "*.c");do ex - $i <<done;done |
09:48.21 | blindvt` | g/bb_strlen/s//strlen/g |
09:48.22 | blindvt` | x |
09:48.26 | blindvt` | done |
09:49.03 | blindvt` | vodz, but find ./ -name "*.c" | xargs sed -i -e "s/bb_strlen/strlen/g;x" should do the same. The former is interactive, which is quite conventient at times. |
09:50.31 | dalias | gcc is so stupid for inlining strlen |
09:50.41 | dalias | the inline version is certainly slower than an optimized version |
09:52.29 | blindvt` | dalias, -fno-builtin[-funcname] and use __builtin_funcname for the gcc-variant of funcname |
09:55.04 | dalias | yeah i know |
09:55.14 | dalias | i'm just saying it's a stupid anti-optimization |
09:56.48 | blindvt` | dalias, yes. Perhaps it should default to off for -Os *shrug* |
09:57.39 | dalias | for everything except -fexpensive-optimizations it should be off |
09:58.01 | blindvt` | reminds me of tcc getting strlen wrong. I ment to take a short glance at that.. hm |
09:58.08 | dalias | heh |
09:59.22 | blindvt` | ah no, it was sizeof. whatever |
10:01.30 | dalias | :) |
10:01.51 | dalias | hmm wtf why does gcc insist on putting .align before my string literals?? |
10:05.24 | vodz | why not? |
10:07.15 | dalias | because it's putting .align 5 (32 byte alignment!) and wasting tons of space |
10:07.37 | dalias | and there's no point in _any_ alignment for string data |
10:09.09 | vodz | write(1, "str", strlen(str)) if "str" aligned - very good idea |
10:10.43 | vodz | mem*, str* inline function very fast if begin str is aligned |
10:11.50 | vodz | else - spaghetti if+goto :( |
10:13.25 | dalias | for write it makes no difference |
10:13.41 | dalias | if it's not aligned the first few bytes will just be read byte-by-byte instead |
10:14.28 | dalias | it's almost impossible for str* functions to take advantage of alignment |
10:14.42 | dalias | they can, but they need to use very fancy bit-twiddling to catch the null termination |
10:14.53 | dalias | certainly not feasible inline |
10:14.56 | dalias | anyway |
10:15.06 | dalias | i'm just trying to figure out how to disable this.. |
10:19.30 | vodz | from glibc: strcpy_small (...) switch ((unsigned int) __srclen) { case 1: __u->__uc = '\0'; case 3: _u = __extension__ ((void *) __u + 2); case 5: __u = __extension__ ((void *) __u + 4); .... |
10:19.58 | vodz | and etc etc a tricks |
10:20.56 | vodz | and its c-like dumb demo, not asm gcc inlines |
10:21.11 | dalias | i just rtfs'd gcc |
10:21.40 | dalias | it's fucking optimization for idiots who write code like strcpy(a, "fucking long string here"); |
10:21.49 | dalias | where the optimizer knows the source is a constant |
10:22.05 | dalias | huge waste of space |
10:22.15 | vodz | aha |
10:22.18 | dalias | fortunately i just found the tine of i386.h to remove it |
10:22.26 | dalias | line |
10:24.23 | vodz | __builtin_constant_p used before call strcpy_small(...) ;-) And call memcpy with know size |
10:25.52 | dalias | yes |
10:26.15 | dalias | evil gnu extension crap for 1% performance gain at 15% size increase |
10:27.02 | dalias | and in all but trivial benchmarking software it's actually a performance _loss_ |
10:27.06 | dalias | because size == speed |
10:27.23 | dalias | basically all realworld code is memory-io/cache-bound |
10:27.33 | dalias | so bigger code == slower code |
10:30.09 | vodz | ok, must buy computer with memory cache eq busybox size ;-))) |
10:31.20 | dalias | :) |
10:35.14 | vodz | from long time dispute ru.os.cmp: must buy computer with memory cache eq JRE size ;-) |
10:36.58 | dalias | haha |
10:37.17 | dalias | damn java kids.. |
10:40.34 | vodz | http://dirty.ru/news/62550 |
10:40.41 | vodz | muahahah |
10:41.12 | vodz | excuse me, its russian-context :( |
10:45.26 | vodz | http://leprosorium.ru/dirty/pics/1/2006-1-14/3229.210103.1137261220.jpg |
10:45.28 | vodz | Rabbit breeding. Advice to the farmer. Cultivation and leaving. |
10:45.36 | vodz | example translated |
10:48.26 | vodz | Inscription: Daily feeling of the superiority: http://leprosorium.ru/dirty/pics/1/2006-1-17/9356.152057.1137499208.jpg |
11:05.51 | vodz | <PROTECTED> |
11:05.56 | vodz | My current location |
11:06.05 | vodz | http://zabyg17.livejournal.com/71674.html |
11:06.36 | *** join/#uclibc psm (n=mps@host-6.mikroweb.hu) |
11:13.17 | vodz | ah, http://www.spelio.net.ru/vs/eng/ have more |
11:16.23 | blindvt` | vodz, for your amusement. last week i was here: http://highgate-cemetery.org/images/fullSize/B&WOrigAvenue760x420.jpg I didn't knew that before, but Marx is buried there. See http://images.google.at/images?&num=50&hl=en&btnG=Google+Search&lr=&as_ft=i&as_qdr=all&as_dt=i&as_rights=&safe=images&sa=N&tab=wi&q=highgate%20cemetary%20marx |
11:16.54 | vodz | http://www.spelio.net.ru/vs/eng/i3.html - My current location of building |
11:17.47 | vodz | left second the building, but www.simtreas.ru have this building of top |
11:20.51 | vodz | blindvt`: wow, beautifully |
11:21.25 | vodz | http://www.ltcconline.net/lukas/winter/marx.jpg |
11:22.16 | vodz | I did not think, that with Marx so respect. |
11:22.44 | vodz | But US fight communists |
11:23.15 | blindvt` | vodz, it's an impressive cemetary. It looks more like a forest, only the tops of the gravestones are looking through ivy |
11:26.51 | vodz | http://journal.vz.ru/rest/2006/2/17/498.html |
11:26.56 | vodz | A russians bath |
11:28.44 | *** join/#uclibc Kwon (n=Kwon@CBL217-132-72-96.bb.netvision.net.il) |
11:28.46 | Kwon | hi |
11:48.01 | Kwon | anyone is using or used Boot Everywhere linux distro? |
11:51.35 | vodz | http://www.liftes.ru/lifts/superlux2.htm Lift of the President of Russian Federation. Ohh |
11:55.42 | vodz | http://www.torino2006.org/ENG/IDF/MDL/MDL_Big.html Hmm Austria is like winners sport? Interesting |
11:58.37 | vodz | Englishmen have recognized Livanov (the Russian actor) as the best Holmes |
11:59.01 | vodz | Wow! We know always ! |
12:00.37 | blindvt` | vodz, just in wintergames and due to skiing counting as N different sports, i guess |
12:01.03 | blindvt` | Kwon, no, i haven't |
12:06.11 | *** join/#uclibc thraxisp (n=thraxisp@ottawa-hs-64-26-170-144.d-ip.magma.ca) |
12:14.19 | vodz | blindvt`: heh, Well and Russian always does not carry with Bobsleigh, sceleton etc. It is the expensive sports. |
12:21.39 | vodz | http://leprosorium.ru/dirty/pics/1/2006-2-22/1643.233959.1140639933.jpg Livanov - Holmes |
12:40.42 | *** join/#uclibc leoncamel (n=leoncame@219.142.132.189) |
12:49.51 | *** join/#uclibc prpplague (n=billybob@72.22.131.136) |
12:55.14 | Kwon | how i change the same value on the screen? like a percentage number that is changing from %1 to %2 to %3 etc |
12:58.46 | *** join/#uclibc woglinde (n=woglinde@e178093100.adsl.alicedsl.de) |
13:03.02 | blindvt` | Kwon, hm? Please explain. |
13:03.47 | Kwon | i want to code a program in c that shows a number on the screen, and then changing this number |
13:04.20 | blindvt` | Kwon, not sure, perhaps you mean i=1;echo %$i;echo %$(($i+1)) |
13:04.50 | Kwon | echo is a c function? |
13:04.56 | blindvt` | Kwon, where you'd say i=$(/my/program) |
13:06.02 | blindvt` | Kwon, no. You're asking a basic C question. May i suggest to ask in #c or read a C FAQ or comp.lang.c or the like? |
13:06.34 | Kwon | i didnt get an answer from any of these resources |
13:08.49 | vodz | busybox xhvt applet |
13:08.53 | vodz | chvt |
13:08.59 | vodz | very trivial |
13:10.03 | blindvt` | Kwon, I don't believe that. Search for simple C code example increment Sounds like you're asking how to write something like void main(void){int i=0;i+=1;printf("%i\n",i);} |
13:10.20 | blindvt` | Kwon, your '%' should be printable via %% iirc |
13:15.55 | Kwon | blindvt`: but this will print %1, right? |
13:16.16 | Kwon | i want the 1 to be changed to 2, and 2 to be changed to 3, do you understand me? |
13:17.09 | blindvt` | Kwon, that would be a simple while or for loop around the increment and print |
13:17.54 | Kwon | but that wont print a new number to the screen? i want to change the current number that is already on the screen |
13:18.12 | solar | use \r vs \n then |
13:19.05 | Kwon | what \r does? |
13:19.17 | solar | man ascii |
13:19.30 | Kwon | and i must use printf and not puts or another function? |
13:19.42 | Kwon | i know what ascii is |
13:19.51 | solar | then why dod you ask me |
13:20.11 | solar | the answer to your question about \r is in the ascii manpage |
13:20.39 | solar | and a little trial and error never hurt anybody for a hello-world.c |
13:21.35 | prpplague | hehe |
13:22.11 | blindvt` | if you want to delete the already printed stuff then you have to use backspace, obviously. for(i=0;i<9;i++){printf("\b\b\b\b%%%i", myinteger);sleep(1);}putc('\n'); hope that helps |
13:24.12 | blindvt` | getting the number of backspaces right depending on how many digits myinteger has is left to you as an exercise ;) |
13:24.17 | vodz | blindvt`: printf("%%%-3d", i); sleep(1); printf("\b\b\b\b"); |
13:24.54 | vodz | or n = printf("%%%d", i); while(n--) putchar('\b') |
13:25.23 | blindvt` | vodz, it was a silly example. Note that you calling printf twice is wasteful compared to just printing once :P |
13:27.22 | vodz | blindvt`: I like not move cursor to before number. Only before change number :-P |
13:28.41 | vodz | cursor can have full size and shadow first numeral |
13:29.28 | Kwon | can i identify the main hard disk for opening it and writing bytes to it? |
13:30.39 | blindvt` | Kwon, of course. look at what blockdev slash is on and operate on that. |
13:30.59 | Kwon | what is blockdev slash? |
13:34.47 | vodz | Kwon: libbb/find_root_device.c |
13:36.15 | vodz | and main hardisk is strange a term |
13:37.02 | vodz | root may be /dev/hda, but other paths may be /dev/sdf10 |
13:37.10 | vodz | what main disk? |
13:39.56 | Kwon | ok, all hard disks |
13:42.24 | *** join/#uclibc carlg (i=clrsrv@www.clearcore.com) |
13:42.24 | *** join/#uclibc ccjoe (i=clrsrv@www.clearcore.com) |
13:46.43 | Kwon | what is root device> |
13:47.25 | Kwon | ? |
13:47.40 | psm | Kwon: some un*x basic knowledge would be good ... |
13:47.53 | Kwon | the boot device? |
13:48.02 | psm | '/' |
13:48.02 | Kwon | what if the boot device is floppy? |
13:49.20 | blindvt` | Kwon, i encourage you to really read up on unix semantics. Never looked at it but perhaps it's nice: http://www.faqs.org/faqs/unix-faq/ |
13:53.18 | vodz | In my live I used a boot devices: punch tape, punch card, magnetic tape, 1 ton disks etc etc etc |
13:56.01 | vodz | In my table have lie VHS magnetic tape - its russian arch device: spec ISA card with video system ;-) |
13:56.50 | blindvt` | vodz, eew punch tape :) don't forget netbooting via PXE |
13:57.14 | vodz | hmm, why lie :-O placed may be more right? |
14:00.33 | vodz | blindvt`: PXE have little sense: M$ do not like memory main disk (Kwon term) ;-)) |
14:04.17 | vodz | http://gorchev.lib.ru/images/pugalo.jpg |
14:06.49 | vodz | by popular russian fiction writer |
14:06.51 | blindvt` | vodz, my first really comfortable PC was a dual P2@400MHz with 2GB ramdisk for the system and scratch. Took a while to boot but was a pleasure to work on |
14:09.08 | vodz | blindvt`: 400MHz with 2GB ram? Hmm. I do not believe this computer configuration |
14:09.30 | vodz | else if not Sun hardware |
14:10.27 | blindvt` | or was it 1.5GB? most likely. It was alot, anyway |
14:11.20 | vodz | my first Linux have 386SX25 with 2Mb memory |
14:12.19 | vodz | I do removed and removed code from kernel, libc and utilites for can comfortable with it |
14:13.15 | CIA-10 | 03pkj * r14259 10uClibc/libc/sysdeps/linux/bfin/bits/syscalls.h: Fix the _syscall6() macro (I hope). |
14:13.21 | vodz | And FreeBSD 2.0 in this PC is not started anyway |
14:14.19 | vodz | after 1.2.13 I can relief |
14:14.53 | vodz | hmmmm, or 1.1.13? Ohh. get old |
14:16.13 | psm | 1.2.13 sounds ok ... |
14:19.20 | vodz | may be, but i wokr with linux pre 1.0 kernels |
14:34.00 | psm | I meant 1.2.13 was a good, usable kernel |
14:35.35 | vodz | O yes, 2 series is a spaghetti. I have ceased there something to understand. |
14:36.21 | vodz | smp, threads ... bloat and slow |
14:38.35 | vodz | after added threads to libc, size 0.4 up 1.2 Mb |
14:39.10 | vodz | after added wcwdth - 2.3 Mb |
14:39.27 | vodz | its progress? :( |
14:55.00 | mjn3 | someone was looking for me? |
14:55.16 | psm | mjn3: long time ago ;-( |
14:55.23 | vodz | mjn3: blindvt` ask about bb_printf |
14:57.06 | mjn3 | what did he want to know? it was a hack to set the stream error indicator for non-stream related errors. that way, an applet could check ferror() for both types of errors rather than having to check the return value of every printf call |
14:57.53 | blindvt` | mjn3, i read it and i think i don't have a question anymore. Thanks anyway |
14:58.01 | mjn3 | ok |
14:59.54 | psm | mjn3: there seems to be some conflict w/ INT_MAX < LONG_MAX/UINT_MAX == ULONG_MAX in ifdef L_abs and L_labs (libc/stdlib/stdlib.c) |
15:04.49 | mjn3 | conflict? |
15:05.25 | psm | lets assume I compile stdlib.c w/ -DL_abs -DL_labs |
15:05.48 | psm | the conflict will show up in strong_alias(labs,abs) |
15:06.13 | psm | because both the earlier mentioned conditions will met |
15:07.21 | psm | the same happens w/ atoi/atol |
15:09.05 | psm | what is the correct approach, dont build atoi and abs if UINT_MAX==ULONG_MAX? |
15:09.54 | psm | or dont strong_alias if INT_MAX<LONG_MAX |
15:16.43 | mjn3 | so you are saying that you have a platform on which both conditions are met? what are the values of the macros? |
15:21.18 | psm | I see INT_MAX and LONG_MAX being defined by gcc (difference in value being only appended L) |
15:22.02 | psm | limits.h has same values for UINT_MAX(appended U)/ULONG_MAX(appended UL) |
15:23.24 | psm | unsigned long = unsigned ? |
15:25.08 | mjn3 | those are just tests to compare the size of long vs the size of int. since sizeof() is a C operator, it can not be used in preprocessor conditionals. i'm curious about how both conditions could be met |
15:30.03 | psm | x86: sizeof int 4/long 4 |
15:32.47 | mjn3 | yes. and on x86 i get only one of the tests evaluating to true |
15:34.02 | mjn3 | need to head to work. i'll be there in about 15 min |
15:38.27 | *** join/#uclibc blindvt__ (n=bf@M2753P002.adsl.highway.telekom.at) |
15:46.05 | blindvt` | Kwon, i have to run now. bbl |
15:52.06 | *** join/#uclibc tty56_ (n=johannes@p54A783D1.dip0.t-ipconnect.de) |
16:03.30 | *** join/#uclibc thraxisp (n=thraxisp@ottgate.precidia.com) |
16:15.42 | blindvt` | with an old make when building out-of tree, i furthermore need to conditionally include the toplevel stuff to avoid endless loops when including files in make. Fix is: |
16:15.44 | blindvt` | find */ -name Makefile | xargs sed -i -e " |
16:16.09 | blindvt` | /top_srcdir=../i ifndef top_srcdir |
16:16.12 | blindvt` | /top_srcdir=../a endif |
16:16.18 | blindvt` | /top_builddir=../i ifndef top_builddir |
16:16.51 | blindvt` | /top_builddir=../a endif" |
16:17.09 | blindvt` | i'll check that in later, fwiw. |
16:19.17 | psm | mjn3: I double checked, the problem exhibits due to different prototypes of abs/labs and atoi/atol probably, you are right about that check, it does not happen at the same time |
16:29.03 | CIA-10 | 03landley * r14260 10busybox/util-linux/fdisk.c: Cleanup from Denis Vlasenko. |
17:18.29 | *** join/#uclibc psm (n=mps@host-6.mikroweb.hu) |
17:30.01 | *** join/#uclibc prpplague (n=billybob@72.22.131.136) |
17:42.00 | vodz | Conditions in the list becomes intolerable. I already cry. After reception of a maintainer Rob became completely deranged. |
17:42.18 | vodz | list = bb maillist |
17:48.51 | dalias | mjn3, i posted to the bugtracker with several major reasons why the printf error hack is a bad idea (and dangerous) |
17:50.20 | dalias | and two easy alternatives that are portable and don't rely on hacking with the internals of an opaque structure |
18:16.22 | CIA-10 | 03psm * r14261 10uClibc/libc/stdlib/stdlib.c: guard __ignore_x correctly, move undefs to their proper place, add prototypes. Should I do all the other *max aliases as well |
18:28.13 | psm | mjn3: the warning about DECIMAL_DIG, wouldn't it be enough to use gcc's __DECIMAL_DIG__ ? or at least use it if DECIMAL_DIG is undefined? |
18:31.40 | CIA-10 | 03psm * r14262 10uClibc/libc/sysdeps/linux/common/bits/uClibc_fpmax.h: Correct comments |
18:59.42 | *** join/#uclibc prpplague (n=billybob@72.22.131.136) |
19:01.25 | *** join/#uclibc zerdith (n=zerdith@c-24-56-229-145.chrlmi.cablespeed.com) |
20:43.39 | SpanKY | ah, vodz left ... i was gonna send him this http://despair.com/dysfunction.html |
20:51.36 | mjn3-work | anyone with an account on mace using lots of hard drive space? there's none currently free |
20:52.18 | mjn3-work | and erik is still out of town |
20:53.30 | SpanKY | i only use morris |
21:01.14 | CIA-10 | 03psm * r14263 10uClibc/libc/stdlib/stdlib.c: Do not use __XL_NPP macro |
22:06.47 | *** join/#uclibc sphaerynx (n=sphaeryn@24.115.86.20.res-cmts.eph.ptd.net) |
22:10.33 | *** part/#uclibc sphaerynx (n=sphaeryn@24.115.86.20.res-cmts.eph.ptd.net) |