--- Log opened Sun Feb 27 00:00:02 2005 --- Day changed Sun Feb 27 2005 00:00 < rxr> then fixfilediff will create the patch ;-) 00:00 < lars__> fixfile? 00:00 < rxr> a comforable function 00:00 < mnemoc_> fixfile? 00:00 < lars__> how to use it? 00:00 < rxr> otherwise you need to take care of .orig fils yourself ... 00:00 < rxr> if you already modded the files it is too late 00:00 < rxr> untar a 2nd copy then ... 00:01 < rxr> othersiwe for the next time you can use fixfile some-file the next time 00:01 < rxr> it will create a .vanilla file suitable for fixfilediff if it does not yet exist ... 00:02 < lars__> ok, I start a new build of xpdf. 00:02 < lars__> I found the problem with the xpdf build. It is somewhat subtil. 00:02 < rxr> just untar a 2nd copy would be enough 00:03 < lars__> too late, the build is running and will fail soon... 00:03 < lars__> Maybe there are other packages suffering from the same problem with tetex 3 00:03 < lars__> in tetex 3 the default engine for latex changed from etex to pdfetex 00:04 < lars__> this causes constructs like \ifx\pdfoutput\undefined not to behave as before 00:05 < lars__> It is used when building a file which is delivered as pdf as well as as ps - which often is the case for documentation of programs 00:07 < lars__> I'll commit the fix soon. If someone encounters a similiar problem, he can refer to the xpdf fix 00:07 < rxr> ok 00:08 < lars__> (my diploma thesis for example suffered from the same thing...) 00:08 < mnemoc_> *g* 00:09 < mnemoc_> when do you have to deliver your diploma thesis? 00:10 < rxr> he already did 00:11 < lars__> I delivered it 5 month ago, but I used it when I was looking for the problem with the new tetex. 00:12 < lars__> Fortunately I had the same tex-construction in it. Otherwise I would not have had an idea what could have been the problem with xpdf 00:15 < rxr> did actually anything work with this uclibc beast ... 00:15 * rxr getting frustated ... 00:16 < CIA-9> rene * r6830 /trunk/package/base/uclibc/uclibc.conf: 00:16 < CIA-9> * fixed the cross prefix to end with - since the programs are named 00:16 < CIA-9> e.g. arm-t2-linux-gnu-strip and not arm-t2-linux-gnustrip ... 00:16 < rxr> look at this: 00:16 < rxr> ^- this even ... 00:16 < rxr> CIA was to fast for the comment *g* 00:17 < rxr> ok - uclibc basically building now ... 00:17 < CIA-9> amery * r6831 /trunk/package/network/nss_ldap/ (. nss_ldap.conf nss_ldap.desc): * added nss_ldap (234) 00:17 < rxr> just that the .config generation needs to be fixed 00:18 < rxr> lars__: I found quite a cool package today 00:19 < rxr> dropbear - a tiny ssh implementation 00:19 < rxr> already packaged 00:20 < lars__> what is the advantage compared to openssh? 00:21 < rxr> way tinier ... 00:21 < rxr> and fits on my PDA targets ... 00:21 < rxr> ssh is 2x bigger and than even depends on openssl which has a binary size of > 1MB ! 00:21 < rxr> that makes that package consume less than 10% of the space of openssh ... 00:22 < lars__> ok, when a reactivate my old laptop some day again... 00:22 < mnemoc_> rxr: uclibc has never ever been tested cross compiling 00:22 < rxr> and openssh is not cross compilable without heavy patching 00:22 < rxr> because those BSD people wrote a configure that is way more harmful than usual ... 00:22 < rxr> mnemoc_: the conditionals have been there ... 00:22 < rxr> Finally finished compiling... 00:22 < rxr> ^- good that uClibc itselfs classified this as "Finally" .... 00:22 < mnemoc_> rxr: the conditionals were including just using imagination :p 00:23 < rxr> 309344 Feb 27 00:22 libuClibc-0.9.27.so 00:23 < rxr> libuClibc-0.9.27.so: ELF 32-bit LSB shared object, ARM, version 1 (ARM), stripped 00:23 < mnemoc_> :D 00:23 < rxr> ok - /me looking into this .config code and then I'm in bed I think ... 00:23 < mnemoc_> go go go 00:24 < rxr> and I thik i have to decustmain uclibc tomorrow ... 00:24 < mnemoc_> what was the i386 thing? 00:24 < rxr> and xpdf, too ... 00:24 < mnemoc_> i have seen no patches to fix pkgsel-like processing 00:25 < rxr> I guess we can declare t2 as final when we touched each rock inherited line once *g* 00:25 < rxr> mnemoc_: yes - that i386 stays in the config even whem arm should be sleected ... 00:25 < rxr> selected even ... 00:25 < rxr> strange somehow I make more typing mistakes on this Apple keyboard ... 00:26 < lars__> rxr: xpdf builds here. But I can not find an xpdf binary. It installs lots of other programs (link pdftops) but no xpdf. 00:26 < rxr> or maybe it is just the delay of the ssh to my bome box ... 00:26 < mnemoc_> hehe 00:26 < rxr> lars__: I remeber it only builds the gui when some old motif or around 00:26 < rxr> if nto only the processing and rendering libs are installed 00:28 < lars__> rxr: I need! some working pdf viewer! Kpdf does not work on many pdfs, acroread does not exist for ppc... how can I get that gui? 00:28 < rxr> lars__: ./configure --help ? 00:29 < rxr> or whatever configure complains what old gui lib is not around = 00:29 < rxr> I do not remeber off hand ... 00:30 < lars__> ok, the new kpdf in kde3.4 is better --- it should use xpdf as well, but in KDE3.3 there are still some fonts problems. 00:30 < rxr> yes 3.4 one is very nice ... 00:30 < lars__> is it really better ? 00:30 < rxr> yep - way 00:30 < jsaw> re 00:31 < rxr> lars__: took a look right now - no idea what xpdf likes 00:31 < rxr> maybe pass --with-x and try what it does ... 00:31 < rxr> hi jsaw ! 00:31 < lars__> rxr: perhaps I should simple emerge KDE3.4 stuff 00:31 < rxr> but 3.4 takes lot of CPU cycles to build ... 00:32 < lars__> How long on an ibook? 00:32 < rxr> openmotif 00:32 < rxr> the .cache lists ... 00:32 < mnemoc_> uhm, i just noticed stuff like $pkg_openldap_prefix which is injected into parse-config on build time, so they don't exist on stage 9 00:33 < jsaw> hi rxr 00:33 < jsaw> hi lars__ , mnemoc_ 00:34 < mnemoc_> hi jsaw 00:34 < mnemoc_> how are your girls? 00:34 < jsaw> all sleeping :) 00:34 < mnemoc_> great :D 00:34 < jsaw> ...currently at least :p... 00:34 < jsaw> :D 00:34 < mnemoc_> you have 15 minutes.... code code code 00:34 < jsaw> I do,... 00:35 < jsaw> I'm testing hotplugging ideas... 00:35 < jsaw> hotplug-ng is nice but... it only works with 2.6, and actually defers the matching to modprobe... with probably use fnmatch... 00:35 < mnemoc_> btw, later i can give you access to a pentium were hotplug frozes 00:35 < mnemoc_> where* 00:36 < jsaw> ouhm, rockplug you mean? 00:36 < mnemoc_> err... rockplug i mean 00:36 < mnemoc_> yes 00:36 < jsaw> :) 00:38 < mnemoc_> rxr,jsaw: what do you think about recording '$prefix' on var/adm/package/$pkg ? 00:39 < mnemoc_> or do you imagine an easy way to 'guess' prefixes of another already installed packages? 00:39 < rxr> not package ... 00:39 < jsaw> + a function like "pkg_prefix " 00:39 < rxr> ack 00:39 < mnemoc_> jsaw: doing what? 00:39 < rxr> mnemoc_: obtaining it rom some "well known location" 00:39 < jsaw> mnemoc_: instead of using pkg_openldap_prefix use "pkg_prefix openldap" 00:39 < rxr> i#m just not that sure that adm/package/ is the right place 00:40 < mnemoc_> prefixes chan change on target on parse-config on {pre,post}.conf etc... 00:40 < mnemoc_> jsaw: i wonder how the guess the value no how to ask for it :) 00:40 < rxr> mnemoc_: that function would just return it from the well known location we need to define 00:40 < mnemoc_> there are no well known location 00:40 < rxr> where it is stored on build time 00:41 < jsaw> so, set_confopt records the prefix.. or? 00:41 < rxr> mnemoc_: I know 00:41 < rxr> we need to define it ... 00:41 < rxr> jsaw: no 00:41 < rxr> the generic cleanup _after_ the build 00:41 < rxr> where /var/adm is generated 00:41 < CIA-9> lars * r6832 /trunk/package/graphic/xpdf/tetex3.patch.t1lib: 00:41 < CIA-9> * Added a patch for xpdf package that makes t1lib build its 00:41 < CIA-9> documentation with tetex-3.0. 00:41 < jsaw> rxr: ok 00:42 < rxr> mnemoc_: ok - the packages looks best .. 00:42 < rxr> mnemoc_: go for it ;-) 00:42 < mnemoc_> no no 00:42 < mnemoc_> i think we can 'extract' it from flist 00:42 < rxr> no! 00:42 < mnemoc_> $prefix/share 00:42 < rxr> do not try those wild guesses ... 00:43 < jsaw> I'd also vote for packages/ 00:43 < rxr> mnemoc_: those guesses often produce random results ... 00:43 < lars__> good night, bye. 00:43 < mnemoc_> prefix or all *dir ? 00:43 < rxr> n8 lars__ ! 00:43 < jsaw> gn8 lars__ 00:43 -!- lars__ [~lars@port-212-202-42-223.dynamic.qsc.de] has quit ["leaving"] 00:43 < mnemoc_> gn8 lars 00:44 < jsaw> I'd say we only record non-standard locations (meaning prefix != /usr etc..) 00:45 < jsaw> or if libdir != ${prefix}/lib 00:45 < jsaw> or so 00:45 < rxr> jsaw: a field in packags is cheap 00:45 < rxr> we could always store it ... 00:45 < rxr> we also encode the standard license GPL ;-) 00:46 < jsaw> or we create a pkg-config for each package not doing it... no... 00:46 < rxr> .oO 00:46 < mnemoc_> panic @ jsaw 00:46 < jsaw> only a thought from long time ago 00:46 < rxr> mnemoc_: you already implement store on packages and the pkg_prefix function ? 00:46 < mnemoc_> nope 00:46 < rxr> you have my vote on that l... 00:47 < mnemoc_> prefix or everything from set_confopt ? 00:47 < rxr> only prefix! 00:47 < mnemoc_> libdir? 00:47 < rxr> no 00:47 < mnemoc_> :) 00:47 < rxr> Place it before the Status: 00:47 < rxr> Build [5] at 02/09/2005 from 05:02:56 to 05:07:29 CET 00:48 < rxr> Prefix: /usr 00:48 < rxr> Status: Stable, License: GPL 00:48 < rxr> ^- artificial /var/adm/packages/reiserfsprogs 00:48 < mnemoc_> ok 00:49 < jsaw> But I'd still record stuff like libdir,sysconfdir only if it is non-standard... readability is my main concern here 00:49 < rxr> I would not encode them at all for now 00:50 < mnemoc_> the main problem today are packages already built without it 00:51 < mnemoc_> i think i agree of only adding non default 00:51 < rxr> make it fail with "abort 'Can not determine prefix from package $pkg" 00:51 < rxr> I would add all 00:51 < mnemoc_> ok 00:51 < rxr> otherwise you can not say if it is an old packag eor just a std. prefix one 00:51 < mnemoc_> adding all is cheap, detecting non-standard-ness is expensive 00:52 < mnemoc_> true 00:52 < rxr> add them to all 00:52 < mnemoc_> ok 00:52 < rxr> and error when it is not there so people rebuild the package to update it ... ;-) 00:52 < rxr> if you are at it - you would be a hero to remove all the current hacks ... 00:52 < rxr> in that area ... 00:53 < rxr> and I fix your config parser to really honor this rule: 00:53 < rxr> apply_conffiles: rule unset TARGET_*. 00:53 < rxr> since it leaves: 00:53 < rxr> TARGET_i386=y 00:53 < mnemoc_> oh 00:53 < mnemoc_> .oO( i want that for kernel.... removing all SCSI for example )o 00:54 < mnemoc_> i'll do prefix, *dir thing today 00:55 * mnemoc_ moving home, back in 30 minutes 00:55 < jsaw> cu 00:56 < rxr> cu mnemoc_ 00:56 < rxr> jsaw: and - do y 00:56 < rxr> ou already know how often you will be here in the next weeks ? *g* 00:57 < jsaw> I'll be here very often, only that I won't commit sth. right now 00:57 < rxr> oh nice ,-) 00:58 < jsaw> I'll have to stay up a while in the night and wait if Charlotte wakes up... then I'll either continue writing thesis or inject my daily dose of hacking ;) 00:58 < rxr> ,-) 01:04 < CIA-9> rene * r6833 /trunk/package/base/uclibc/recent-tar.patch: * fixed uclibc for recent tar versions 01:05 < CIA-9> rene * r6834 /trunk/package/base/uclibc/clean_arch.config: 01:05 < CIA-9> * quick fixed the uclibc/clean_arch.config to use real sed regex to 01:05 < CIA-9> remove config lines - if the auto_config is intended to take shell 01:05 < CIA-9> like globbing it must be fixed to expand * to .* instead ... 01:13 < rxr> mnemoc_: I think I remove the optimization optiosn from the uclibc package 01:13 < rxr> they only pass gcc optoins 01:13 < rxr> which our wrapper will substitute anyways ... 01:16 < rxr> the last changes on that lines was: 01:16 < rxr> r2783 | clifford | 2004-04-23 01:01:48 +0200 (Fri, 23 Apr 2004) | 10 lines 01:16 < rxr> Alejandro Mery: 01:16 < rxr> uClibc 0.9.26: no custmain, gcc33 only (for the moment) 01:16 < rxr> autofix config.sub, locale based on DISABLE_NLS 01:16 < rxr> stages 0-1 complete; root usable; doesn't enter stage 2 :( 01:16 < rxr> parse-config-fixarch to $toolsdir 01:18 < CIA-9> rene * r6835 /trunk/package/base/uclibc/uclibc.conf: 01:18 < CIA-9> * removed fine grained optimization settings from uclibc which where 01:18 < CIA-9> only implemented for i386. Initiall I started to add the arm ones 01:18 < CIA-9> however I greppped the uClibc source and they only set gcc options 01:18 < CIA-9> which our gcc wrapper will substitude anyways - so all the code 01:18 < CIA-9> throwing the detailed options into uclibc's configure is totally 01:18 < CIA-9> useless. 01:22 * rxr off soon 01:30 < rxr> when I scroll over submaster I have to say "we are already there" ... 01:31 < jsaw> yep 01:32 < rxr> the digikam stuff got even added to t2 in last december ... *g* 01:32 < jsaw> rev count... 01:33 < rxr> squid SABLE8 is in trunk since 2005-02-12 01:33 < rxr> the submaster ID is 5 days older ... 01:34 < rxr> + the apply delay *g 01:34 < rxr> * 01:35 < rxr> the initrd is the same 01:35 < rxr> I fixed our for 2.6 in _november_ last year ... 01:35 < rxr> and make my mind up with a real -ng rewrite 01:35 < rxr> and submaster junk can be called "from regen in die traufe" ... 01:36 < jsaw> :) 01:36 < jsaw> -ng rewrite? 01:37 < rxr> aka next generation 01:38 < jsaw> that I know, but what do you ng-rewrite? 01:38 < jsaw> oh, you mean the scripts 01:38 < jsaw> is rock discussing this too? 01:42 < rxr> no - I was only talking about the initrd at that moment 01:42 < rxr> damn is the uclibs building long ... 01:42 < jsaw> ah 01:42 < rxr> are you sure it is uc ? 01:43 < jsaw> yes 01:43 < rxr> oh - the auto_config is running endlessly .. 01:43 < _Ragnar_> re 01:44 < jsaw> ? 01:44 < _Ragnar_> which package? 01:44 < rxr> uc libc 01:45 < jsaw> guess I'm gone... Charlotte sends sigstop sigcome_to_me 01:45 < rxr> hehe - cu 01:47 < _Ragnar_> doesn't look endlessly for me 01:48 < rxr> I changed it quite a bit this night to: 01:48 < rxr> a) bild on non x86 01:48 < rxr> and 01:48 < rxr> b) cross build ... 01:50 < _Ragnar_> ;p 01:51 < _Ragnar_> did you change parse-config-9? 01:51 < rxr> don't think so ... 01:51 < _Ragnar_> okay 01:55 < _Ragnar_> have you commited your changes? 01:56 < rxr> sure 01:58 < _Ragnar_> :) 02:02 < mnemoc_> rxr: don't remove them 02:03 < rxr> the optmization ? 02:03 < mnemoc_> rxr: that's how uclubc detects it 02:03 < rxr> already done ... 02:03 < mnemoc_> :'( 02:03 < rxr> detect what ? 02:03 < mnemoc_> the arch 02:04 < mnemoc_> uhm 02:04 * mnemoc_ read commit message 02:04 < rxr> they are set byt he wrapper anyway ... 02:04 < rxr> there is not point of this error prone (since never updated code there ... 02:05 < _Ragnar_> doesn't it use the defines in some header files, too? 02:05 < rxr> I grepped for the config optoins 02:05 < rxr> they only define -mcpu and code 02:05 < _Ragnar_> ah okay 02:05 < rxr> there are some occurances qwh 02:06 < rxr> they then use the gcc internal defines for #ifdefs (shiffer) - but our gcc wrapper already injects the wanted optimization so, ... 02:06 < _Ragnar_> okay 02:08 < mnemoc_> i'm more concerned about 6834, but i can't see the diff yet 02:08 < mnemoc_> i'll revert that 02:08 < rxr> why ? 02:08 < mnemoc_> because that's not the purpose 02:08 < rxr> what is it then ? 02:09 < mnemoc_> * should be properly translated 02:09 < rxr> that the parser expans the * ? 02:09 < mnemoc_> like pkgsel 02:09 < rxr> ok - but then not reveert it butimplement it properly in the same changeset 02:09 < mnemoc_> yes 02:09 < mnemoc_> i had a different TODO for tonight but i'll have to fix this first 02:10 < rxr> should be trivial 02:10 < mnemoc_> i have to discover why [^ ]* doesn't work 02:10 < rxr> that does work 02:10 < rxr> IIRC 02:11 < rxr> but you dump the config tag directy into the sed expression 02:11 < mnemoc_> i'll have to check that 02:11 < rxr> so the sed exprssion is e.g. TARGET_* and not TARGET_.* 02:11 < mnemoc_> but the delay is high at this hour 02:12 < rxr> mnemoc_: do your other stuff 02:12 < mnemoc_> my neighbor is leeching from emule :\ 02:12 < rxr> it works right now and you can fix it when you have free time 02:12 < rxr> or I do it tomorrow ... 02:12 < mnemoc_> TARGET_* ---> TARGET_[^ ]* 02:12 < mnemoc_> not .* 02:12 < mnemoc_> i have to fix that 02:12 < rxr> I had TARGET_* in my file 02:12 < mnemoc_> but i have to wait until the leecher go to sleep 02:13 < mnemoc_> the file have to have that 02:13 < rxr> traffic shape hinm *g* 02:13 < mnemoc_> it should be translated on the fly 02:13 < rxr> yes yes 02:13 < rxr> I meant the translated file ... 02:13 < mnemoc_> he provides me the inet connection 02:13 < mnemoc_> he has the router, and it is a dlink/arm thing 02:14 < mnemoc_> it doesn't even does dhcp properly 02:14 < mnemoc_> crappy thing 02:14 < rxr> and he is willingly doing ir - or did you hijack him *g* 02:14 < mnemoc_> :p 02:14 < mnemoc_> he thinks he has a 'firewall' 02:15 < mnemoc_> actually it does packet filtering but also gives me access :p 02:16 < mnemoc_> damn... i need to sleep 02:18 * rxr too 02:18 < rxr> but this uclibc keeps me awake ... 02:18 < mnemoc_> 23° at 22:18 is too much 02:18 < rxr> I know why I prefer dietlibc .. 02:18 < _Ragnar_> *lol* 02:18 < _Ragnar_> what are you building? 02:18 < rxr> cross build arm 02:19 < _Ragnar_> oh okay 02:19 < rxr> had to notice noone ever tried uclibc on non x86 nor cross build it on ROCK /t2 ... 02:19 < _Ragnar_> dietlibc doesn't do arm? 02:19 < rxr> sure 02:19 * mnemoc_ doing a fresh checkout to his lapto 02:19 < rxr> I have already running and relased it ... 02:19 < rxr> but I want to test uclibc 02:19 < mnemoc_> working remote is imposible 02:19 < _Ragnar_> ah okay 02:19 < rxr> and the dynamic stuff on dietlibc is not really advanced ... 02:19 < rxr> if ti would I would not need anything else ... 02:20 -!- valentin [~valentin@port-212-202-42-223.dynamic.qsc.de] has quit [Read error: 60 (Operation timed out)] 02:20 < rxr> .oO 02:21 < mnemoc_> svn: PROPFIND of '/t2/trunk': Could not read status line: Connection reset by peer (http://svn.exactcode.de) 02:21 < mnemoc_> .oO( i hate emule )o 02:21 < rxr> mnemoc_: see valentins disconnect above ... 02:21 < mnemoc_> oh 02:21 < mnemoc_> :) 02:21 < rxr> hm - online again ? 02:21 < rxr> I have access ... 02:21 < rxr> try again ... 02:21 < mnemoc_> yes... started 02:21 < mnemoc_> :)} 02:21 -!- valentin [~valentin@port-212-202-46-153.dynamic.qsc.de] has joined #t2 02:24 < CIA-9> rene * r6836 /trunk/package/base/uclibc/uclibc.conf: * of couse $bindir has also nothing to do on the CROSS option of uclibc 02:25 < _Ragnar_> lag;p 02:25 < mnemoc_> rxr: uclibc/functions.in:11 02:26 < rxr> maybe overquoted for bash3 ? 02:26 < CIA-9> rene * r6837 /trunk/package/base/uclibc/functions.in: 02:26 < CIA-9> * do not run a final apply_conffiles to prevent incomplete config files 02:26 < CIA-9> * do run config with yes to make sure it does not stall (compare 02:26 < CIA-9> lx_config.sh) 02:26 < rxr> btw please never use a b c and so on ... 02:27 < rxr> use useful names so one knows what tag of the file it is ;-) 02:28 < rxr> so I think uclibc will finally pass a build for cross-arm ... 02:28 -!- _Ragnar__ [loki@66-146-166-62.skyriver.net] has joined #t2 02:29 -!- _Ragnar_ [loki@66-146-166-62.skyriver.net] has quit [Remote closed the connection] 02:30 * mnemoc_ rewriting apply_conffiles 02:34 < rxr> == 02:22:57 =[1]=> Building base/uclibc [0.9.27 2.1.0-beta4]. 02:34 < rxr> == 02/27/05 02:28:02 =[1]=> Finished building package uclibc. 02:34 < rxr> == 02:28:04 =[1]=> Building archiver/gzip [1.3.5 2.1.0-beta4]. 02:34 < rxr> == 02/27/05 02:28:19 =[1]=> Finished building package gzip. 02:34 < rxr> == 02:31:41 =[1]=> Building base/ncurses [5.4 2.1.0-beta4]. 02:34 < rxr> ok /me in bed ... 02:35 < rxr> looking at the result in the morning ... ;-) 02:36 < mnemoc_> :) 02:36 < mnemoc_> gn8 rene 02:45 < _Ragnar__> nini rxr 02:52 -!- valentin_ [~valentin@port-212-202-41-157.dynamic.qsc.de] has joined #t2 03:02 -!- valentin [~valentin@port-212-202-46-153.dynamic.qsc.de] has quit [Read error: 110 (Connection timed out)] 03:13 < CIA-9> amery * r6838 /trunk/package/base/uclibc/functions.in: * changed uclibc/functions.in's apply_conffile to use variables with meaning 03:15 < mnemoc_> rxr: stupid error.... i used \\ to escape * :( 03:27 -!- valentin_ is now known as valentin 03:27 < valentin> re 03:27 < mnemoc_> hi valentin 03:29 < mnemoc_> valentin: what's the t2-logo-rc1? 03:31 < valentin> t2-logo-rc1 ? 03:32 < mnemoc_> the logo you where writing^Wdrawing 03:32 < mnemoc_> were* 03:33 < valentin> the last version i did was http://exactcode.de/valentin/t2logoValentin.png some hours ago 03:34 < valentin> rxr has a different version, i do not know if he worked on that 03:39 < CIA-9> amery * r6839 /trunk/package/base/uclibc/clean_arch.config: * reverted r6834 - this is an exploit not a fix 03:41 < CIA-9> amery * r6840 /trunk/package/base/uclibc/functions.in: 03:41 < CIA-9> * fixed uclibc's apply_conffiles to do globbing properly. 03:41 < CIA-9> * improved uclibc's apply_conffiles to cache .sed file 03:44 < valentin> this is an exploit not a fix ? 03:45 < mnemoc_> :p 03:45 < mnemoc_> he changed the source to work with a broken processor instead of fixing the processor 03:50 < _Ragnar__> ooh? 03:50 < _Ragnar__> exploit?:) 03:50 < _Ragnar__> *rofl* 03:51 < mnemoc_> :p 03:52 < mnemoc_> he exploited my broken rules processor to get what he wanted 03:56 < mnemoc_> i need a name for a function 03:57 < mnemoc_> pkgprefix [] 03:57 < mnemoc_> pkggetdir [] 03:58 < mnemoc_> i don't like those 03:58 < mnemoc_> type={bindir,libdir,sbindir,includedir,sysconfdir,localstatedir,....} 04:01 < mnemoc_> _Ragnar__: ? 04:01 < _Ragnar__> ? 04:01 < mnemoc_> comments? 04:02 < _Ragnar__> about what? 04:02 < mnemoc_> i need two functions but i'm not sure about the names 04:03 < valentin> i have to sleep now, cu 04:03 < mnemoc_> cu valentin 04:04 < mnemoc_> --with-ldap-libs=`pkgprefix libdir openldap` 04:04 < mnemoc_> --enable-ssl=$root/`pkgprefix openssl` 04:08 < _Ragnar__> sounds good 04:09 < mnemoc_> libdir=$root/`pkggetdir libdir $prefix` 04:09 < mnemoc_> and that? 04:10 < _Ragnar__> uhm I'm not sure what that should do 04:10 < _Ragnar__> the $prefix part 04:10 < mnemoc_> for that prefix, which is the default libdir 04:11 < mnemoc_> [ "`pkggetdir libdir`" != "$libdir" ] && echo "libdir has been redefined" 04:18 < _Ragnar__> hmmmm okay 04:19 < mnemoc_> that's an internal helper 04:19 < mnemoc_> but i don't like the name :\ 04:21 < _Ragnar__> hmm 04:26 < mnemoc_> what this silence means? 05:05 -!- _Ragnar__ is now known as _Ragnar_ 05:05 < _Ragnar_> mnemoc_: distracted 05:05 < _Ragnar_> I can't think of a better name right now 05:05 < mnemoc_> ok :) 05:58 -!- rxr_ [~rene@p213.54.199.35.tisdip.tiscali.de] has joined #t2 05:58 -!- Topic for #t2: T2 | 2.1.0-beta3 RELEASED | The next generation of System Development Enviroments (SDE) | http://www.t2-project.org/ | CIA, don't ever leave us again! 05:58 -!- Topic set by mnemoc [] [Sat Feb 19 18:03:55 2005] 05:58 [Users #t2] 05:58 [ _Ragnar_] [ jsaw ] [ mtr_] [ Postal] [ rxr_ ] [ valentin] 05:58 [ CIA-9 ] [ mnemoc_] [ nzg ] [ rxr ] [ SerWou] 05:58 -!- Irssi: #t2: Total of 11 nicks [0 ops, 0 halfops, 0 voices, 11 normal] 05:58 -!- Channel #t2 created Sun Aug 8 21:15:33 2004 05:58 -!- [freenode-info] help freenode weed out clonebots, please register your IRC nick and auto-identify: http://freenode.net/faq.shtml#nicksetup 05:58 -!- Irssi: Join to #t2 was synced in 11 secs 06:14 -!- rxr [~rene@p213.54.193.177.tisdip.tiscali.de] has quit [Read error: 110 (Connection timed out)] 06:15 < CIA-9> amery * r6841 /trunk/scripts/functions: 06:15 < CIA-9> * introduced helper function `pkggetdir ` to generate locations for the different $foodir considering $prefix and $xpkg 06:15 < CIA-9> * cleaned set_confopt using pkggetdir 07:00 < CIA-9> amery * r6842 /trunk/scripts/Build-Pkg: * improved Build-Pkg to store on package/$pkg info about prefix and relevant foodir-s which have been redefined 07:25 < CIA-9> amery * r6843 /trunk/scripts/functions: * introduced pkgprefix() function which returns the prefix or a foodir of an already installed package 07:28 < mnemoc_> rxr_,jsaw: look at 6841, 6842 and 6843 please, and comment :) 07:37 < mnemoc_> i'll start cleaning hacks on that area after your vote to the functions 07:43 < CIA-9> amery * r6844 /trunk/scripts/functions: * fixed Build-Pkg to honor srctar even if desc_SRC is defined (usefull for postlinux.conf) 07:54 < mnemoc_> aaarg... abort doesn't work on pkgprefix becasue it's catched by `foo` as simple output :| 07:55 < mnemoc_> ideas? 07:55 < mnemoc_> i'll go to sleep now 07:55 < mnemoc_> i hope you can found something 08:08 < _Ragnar_> hmmm seems rxr broke uclibc quite well ;p 08:47 -!- SerWou [~SerWou@lafilaire-3-82-224-107-105.fbx.proxad.net] has quit [] 09:26 -!- mipe [~mipe@dsl10040.japo.fi] has joined #t2 09:43 < rxr_> re 09:43 < mipe> hi rxr 09:46 < _Ragnar_> rxr_: what have you done to the uclibc config? 09:47 < rxr_> I fixed it ;-) 09:47 < _Ragnar_> .config:2: trying to assign nonexistent symbol Automatically 09:47 < _Ragnar_> .config:24: trying to assign nonexistent symbol Target 09:47 < _Ragnar_> .config:59: trying to assign nonexistent symbol General 09:47 < _Ragnar_> .config:92: trying to assign nonexistent symbol Networking 09:47 < _Ragnar_> .config:98: trying to assign nonexistent symbol String 09:47 < _Ragnar_> .config:112: symbol value 'y' invalid for UCLIBC_PRINTF_SCANF_POSITIONAL_ARGS 09:47 < rxr_> and mnemoc tinkered on it after me ... 09:47 < _Ragnar_> that, afaik, was the same before and after mnemoc_ modified sth 09:48 < rxr_> _Ragnar_: did rsync built for you with uclibc ? 09:48 < _Ragnar_> this is when building uclibc! 09:49 < _Ragnar_> so: no, since I can't build uclibc 09:49 < rxr_> yes - I just asked because my build stopped on rsync ... 09:49 < rxr_> *g* 09:49 < rxr_> and before? 09:49 < rxr_> _Ragnar_: when is this 1-uclibc ? 09:49 < rxr_> for native x86 ? 09:49 < _Ragnar_> 0-uclibc 09:49 < rxr_> 0 ?! 09:49 < _Ragnar_> native x86 yes 09:49 < rxr_> ok - I test here ... 09:49 < _Ragnar_> yes 09:49 < rxr_> did rsync build in the past for you? 09:50 < _Ragnar_> yes 09:50 < _Ragnar_> I think so 09:52 < rxr_> _Ragnar_: I take a look at your uclibc regression after breakfast, ok = 09:52 < rxr_> ? even 09:52 < CIA-9> rene * r6845 /trunk/package/ (2 files in 2 dirs): 09:52 < CIA-9> * moved the dietlibc workaround for rsync to rsync itself, since it 09:52 < CIA-9> also fixes rsync for uclibc (on arm) 09:52 < _Ragnar_> okay 09:53 < rxr_> cu then ,-) 09:53 < _Ragnar_> ciao 10:24 < rxr_> re 10:25 -!- You're now known as rxr 10:27 < rxr> # Note: this leaves around /usr/lib/libgpm.so.1 and /usr/lib/libgpm.so.1.nn 10:27 < rxr> # This is probably the right thing, because programs may be linked to it 10:27 < rxr> uninstall: 10:27 < rxr> ^- what a uninstall target :-) 10:30 < CIA-9> rene * r6846 /trunk/package/base/gpm/gpm.conf: * fix gpm for cross builds 10:33 < _Ragnar_> re 10:33 < _Ragnar_> rxr: anything more with the uclibc bug? 10:33 < rxr> I start on it in some minutes 10:35 < _Ragnar_> from src.uc*/uC*/.config: 10:35 < _Ragnar_> # 10:35 < _Ragnar_> Automatically=y 10:35 < _Ragnar_> # 10:36 < rxr> fefe's embutils are quite dietlibc centric 10:37 < _Ragnar_> *g* 10:40 < _Ragnar_> well I need to get to bed 10:41 < rxr> cu _Ragnar_ 10:43 < rxr> hm - arm 0-uclibc still builds ... 10:45 < rxr> The Global File System 6.1-pre3 10:49 < rxr> hm - a x86 build also passed 0-uclibc ... 11:17 < CIA-9> rene * r6847 /trunk/package/base/busybox/busybox.conf: * fixed busybox for cross stage builds 11:19 -!- mnemoc [~amery@200.75.27.5] has joined #t2 11:20 < rxr> hi mnemoc_ 11:33 -!- mnemoc_ [~amery@200.75.27.59] has quit [Read error: 110 (Connection timed out)] 11:43 -!- mipe_ [~mipe@dsl10040.japo.fi] has joined #t2 11:51 -!- mipe [~mipe@dsl10040.japo.fi] has quit [Read error: 110 (Connection timed out)] 12:12 < rxr> 1585617 Feb 27 12:11 image/initrd.gz 12:12 < rxr> uclibc based PDA target ;-) 12:13 -!- mipe_ is now known as mipe 12:15 -!- mtr__ [~michael@H99b6.h.pppool.de] has joined #t2 12:16 < rxr> moin mipe and mtr__ 12:17 < rxr> dropbear: 12:17 < rxr> Using uClibc - login() and logout() probably don't work 12:17 < rxr> .... ? 12:17 < mipe> hi rxr 12:18 < rxr> hi mipe 12:19 * rxr needs to move the desk - to sunny here ... 12:19 < mipe> hmm,is it fhs compatible if i do /lib32 and /lib64 and /lib is symlink to /lib32, i know that it specifies that /lib has 32bit libraries 12:19 < CIA-9> rene * r6848 /trunk/package/security/dropbear/dropbear.conf: * fixed dropbear.conf LD= typo introduced last night ... 12:19 < rxr> mipe: why do you want to do this ? 12:19 < mipe> ? the dirs? 12:20 < rxr> yes 12:20 < rxr> why lib32 and lib64 when you symlink lib to lib32 anyway ... 12:21 < rxr> # file usr/bin/dbclient 12:21 < rxr> usr/bin/dbclient: ELF 32-bit LSB executable, ARM, version 1 (ARM), for GNU/Linux 2.0.0, dynamically linked (uses shared libs), stripped 12:21 < rxr> 106816 Feb 27 12:19 usr/bin/dbclient 12:21 < rxr> ^- that is a reasonable size for a ssh client ... 12:22 < rxr> not the 280kB of openssh + the 1.xMB openssl ... ;-)! 12:22 < rxr> _and_ it cross compiles - not like the stinking openssh configure .... 12:22 < mipe> i dont know :), for future? fhs says:Using lib for 32-bit binaries allows existing binaries from the 32-bit systems to work without any changes: such binaries are expected to be numerous. 12:22 < mipe> i dont them to be numerous,only boot stuff atm as far as i can see 12:23 < rxr> but why symlink when just lib and lib64 works as well? 12:23 < rxr> what does lib -> lib32 give you ? 12:23 < rxr> 1806329 Feb 27 12:22 initrd.gz 12:23 < mipe> atm nothing. 12:23 < rxr> ^- including a tiny ssh ,-) 12:23 < mipe> so gonna change that... 12:24 * mipe is at lunch 12:25 < rxr> cu 12:29 -!- mtr_ [~michael@H9b07.h.pppool.de] has quit [Read error: 110 (Connection timed out)] 12:33 < CIA-9> rene * r6849 /trunk/package/multimedia/libmpeg/libmpeg.desc: * marked libmpeg NOPARALLEL 12:35 < CIA-9> rene * r6850 /trunk/package/develop/avrdude/avrdude.conf: * fixed avrdude to not build some documentatoin files we do not need 12:35 < CIA-9> rene * r6851 /trunk/package/textproc/libxml1/libxml1.desc: * marked libxml1 as NOPARALLEL 12:36 < CIA-9> rene * r6852 /trunk/package/x11/xorg/postlinux.conf: 12:36 < CIA-9> * fixed typo in xorg/postlinux.conf to use postmake instead of premake 12:36 < CIA-9> for the install 12:37 < CIA-9> rene * r6853 /trunk/package/graphic/magicpoint/magicpoint.desc: * marked magicpoint as NOPARALLEL 12:37 < CIA-9> rene * r6854 /trunk/package/base/fgetty/x-de-diet-fication.patch: * fixed fgetty for non dietlibc builds 12:38 < CIA-9> rene * r6855 /trunk/package/base/linux24/linux24.conf: * fixed linux24 for rebuilds by forcing a symlink made 12:53 < CIA-9> rene * r6856 /trunk/target/psion-pda/ (build_image.sh config.in fstab group issue passwd profile): 12:53 < CIA-9> * added soe hardcoded files and migrated the psion-pda target to uclibc 12:53 < CIA-9> I'll later implement an option to toggle between uclibc and dietlibc 12:53 < CIA-9> and the target will not yet build fully since it relies on statically 12:53 < CIA-9> minit files I do not ceck in - however I'll add real minit support 12:53 < CIA-9> over the next days ... 14:25 -!- nosfod [~nosfod@p213.54.223.190.tisdip.tiscali.de] has joined #t2 14:25 < nosfod> hello 14:26 < nosfod> whats the status of t2 for x86-64? 14:27 < rxr> compiles normally .. 14:27 < rxr> only the lib32 does some trouble from time to time ... ;-) 14:27 < nosfod> i got an error in stage 1 14:27 < rxr> mipe is the specialist for it ;-) 14:27 < nosfod> gcc does not find the libs in */lib64 14:28 < rxr> mipe: ? ;-) 14:28 < rxr> nosfod: best stay around abit and wait for mipe 14:28 < nosfod> rxr: ok 14:28 < rxr> I have no x86-64 hardware around ... 14:33 * mipe is back 14:33 < mipe> nosfod:whats the error? 14:34 < nosfod> ld: crti.o: No such file: No such file or directory 14:35 < nosfod> mipe: the libs are in /usr/lib64 but 14:37 < mipe> nosfod:http://koti.japo.fi/~pmika/diff.patch 14:37 < mipe> nosfod:apply that patch to your t2 trunk 14:37 < nosfod> mipe: thanks :) 14:38 < mipe> nosfod:and one more patch coming up,mom 14:41 < mipe> nosfod:http://koti.japo.fi/~pmika/amd64.patch 14:41 < mipe> nosfod:copy that to your package/base/gcc folder 14:42 < nosfod> mipe: ok 14:43 < mipe> nosfod:those are work in progress patches, i did get my t2 amd64 build that way but it still needs some cleaning up before they can go to svn 14:48 < mipe> nosfod:tell me if that works,you may need to rebuild your glibc,glibc32 and gcc, /scripts/Create-Errlist -cfg -remove 14:48 < mipe> well gcc is for sure but may need some glibc and glibc32 rebuilding 14:49 < nosfod> mipe: i will start from 0 15:08 < mipe> kdelibs fail with shared files with other packages error :( 15:11 < mipe> rxr: did you build new kde to your ibook? did you get "shared files" error then? i get kdebase and kdelibs conflict 15:16 < rxr> I got one error - I do not remeber for which package ... 16:21 < rxr> ouhm - building arm uclibc on my 800Mhz iBook takes 13 minutes ... 16:23 < mnemoc> _Ragnar_: hi, can you show me your .config ? 16:23 < mnemoc> moin 16:24 < mnemoc> rxr: did you saw var/adm/package/$pkg thing ? 16:25 < rxr> oh no - not yet 16:31 < mnemoc> i had a problem with the abort 16:31 < mnemoc> it doesn't work inside `foo` 16:35 < rxr> hm 16:40 < CIA-9> rene * r6857 /trunk/package/base/uclibc/recent-tar.patch: 16:40 < CIA-9> * removed uclibc/recent-tar.patch, it results in usr/include as shared 16:40 < CIA-9> file with 00-dirtree and is not needed on my iBook - seems to be some 16:40 < CIA-9> defect in my Athlon's tar 16:51 < mnemoc> iirc i used tar -cf - | tar -xf - just for that 16:51 < rxr> the problem was multiple arguments 16:51 < rxr> tar -cf ... include/* 16:52 < rxr> the include/* caused trouble ... 16:52 < mnemoc> yes? 16:52 < mnemoc> uhm 16:52 < mnemoc> weird 16:52 < rxr> on my athlon ... 16:52 < rxr> yes weired ... 16:53 < rxr> the 2nd tar complain some chunks have not been tar headers and so ... 16:56 < mnemoc> oh 16:56 < rxr> on my iBook it worked ... 16:56 < rxr> so I have to compare later what tar is on my athlon that does not work that well 16:57 < rxr> mnemoc: btw I just hack my own minit init stuff 16:57 < rxr> runit does not look that comatible ... 16:58 < mnemoc> take a look into the 'tail' thing, that way you can generate minit specific init scripts 16:58 < mnemoc> i'll go to shower, bbl 17:10 < mnemoc> re 17:28 < CIA-9> amery * r6858 /trunk/package/base/uclibc/parse-config-9: * added uclibc autohack for device-mapper and lvm2 17:45 < mipe> Error logs from system-2.1.0-beta4-x86-64-athlon64-desktop-expert: 17:45 < mipe> 257 builds total, 91 completed fine, 0 with errors. 17:46 < mipe> not bad,i'm curious to see if any actuall desktop related package will fail 17:47 < mnemoc> mipe: does a minimal run? 17:48 < mipe> yep,i'm building that on generic-minimal 17:49 < mnemoc> cool :D 17:49 < mipe> some problems with grub and cracklib but i'll fix those before this build,i hope... 17:50 < mnemoc> will we provide amd64-desktop ISOs for beta4 then? :) 17:50 < mipe> when is its ETA? 17:51 < mnemoc> it was scheduled for today, but.... 17:51 < mnemoc> i have to go to parents-in-law house... 17:52 < mnemoc> cu in some hours 17:52 < mipe> cu 18:13 < rxr> re 18:17 < CIA-9> rene * r6859 /trunk/package/base/minit/ (66 files in 20 dirs): * added the first bunch of init files for minit 18:19 < mipe> hmm,i get lots of kde packages failing because of conflict with libs in other packages.... 18:19 < rxr> hm 18:19 < rxr> what files exactly ? 18:21 < CIA-9> rene * r6860 /trunk/target/psion-pda/build_image.sh: 18:21 < CIA-9> * removed obsolete manually prebuild minit files injection from the 18:21 < CIA-9> psion-pda target 18:21 < mipe> kde-i18n-de 18:21 < mipe> kde-i18n-es 18:21 < mipe> kde-i18n-fr 18:21 < mipe> kdeartwork 18:21 < mipe> kdebase 18:21 < mipe> kdeedu 18:21 < mipe> kdelibs 18:21 < mipe> kdepim 18:21 < mipe> kdeutils 18:21 < mipe> koffice 18:21 < mipe> are failing because of conflict with each other 18:22 < rxr> hm - strange 18:22 < rxr> you did not had those installed before ? 18:22 < mipe> they do build fine,just the post stuff is failing 18:22 < mipe> yep, they are installed,updating them to my laptop 18:22 < rxr> ah - an 3.3 -> 3.4 update ? 18:23 < mipe> yep 18:23 < rxr> I got that too ... 18:23 < rxr> just ignore them I would say ... 18:23 < mipe> ah ok :) 18:23 < rxr> they are built and installed ... 18:23 < rxr> just the post build check complains about shared files 18:23 < mipe> yep 18:23 < rxr> however 2 of them also happen for a 3.4 build from scratch 18:23 < rxr> we need to fix those ;-) 18:28 < mipe> hmm,should test if grub2 is usable... 18:45 < mipe> Error logs from system-2.1.0-beta4-x86-64-athlon64-desktop-expert: 18:45 < mipe> [5] x11/xorg [5] gnome14/glib12 18:45 < mipe> 257 builds total, 155 completed fine, 2 with errors. 18:46 < mipe> xorg failing because:/usr/lib64/gcc/x86_64-t2-linux-gnu/3.4.3/../../../../lib64/libfreetype.so: undef 18:46 < mipe> ined reference to `inflate' 18:56 < rxr> isn't grub2 very work in progress ? 19:09 < mipe> rxr:yep 19:13 < mipe> hmm,this is wierd,freetype gets sysconfdir as /usr/etc 19:27 < CIA-9> rene * r6861 /trunk/package/base/minit/ (functions minit.desc): 19:27 < CIA-9> * fixed minit file installation for rebuilds as well as sorted it behind 19:27 < CIA-9> uclibc 19:29 -!- mipe [~mipe@dsl10040.japo.fi] has quit ["tomorrow"] 20:41 < CIA-9> rene * r6862 /trunk/package/base/minit/functions: * fixed minit to not copy .svn files ;-) 20:43 < _Ragnar_> rxr: did you fix uclibc?:) 20:44 < CIA-9> rene * r6863 /trunk/target/psion-pda/build_image.sh: 20:44 < CIA-9> * improved the pda target to preserve symlinks, not symlink identical 20:44 < CIA-9> files with a size of zero and not copy .old files 20:51 < rxr> _Ragnar_: it did not fail here ... 20:51 < rxr> however I got another report from someone where it looks like your problem 20:52 < rxr> see the mailing lits 21:16 < _Ragnar_> rxr: did you propose a solution? 21:23 < _Ragnar_> yes, that's the same problem 21:23 < rxr> _Ragnar_: can you send me the .config* files so I can try to imagine why this happens ? 21:23 < rxr> it does not happen here - and not for mnemoc 21:24 < _Ragnar_> problem is: the sed scripts make from: 21:24 < _Ragnar_> # 21:24 < _Ragnar_> # Automatically generated make config: don't edit 21:24 < _Ragnar_> # 21:25 < _Ragnar_> to: 21:25 < _Ragnar_> # 21:25 < _Ragnar_> Automatically=y 21:25 < _Ragnar_> # 21:25 < rxr> ah 21:25 < rxr> I also right now got the logs from this other person 21:25 < _Ragnar_> :) 21:26 < rxr> whyn busybox finished on my iBook I review this next 21:26 < rxr> amazing how logn uclibc and busybox build altough they should be minimal ... 21:27 < _Ragnar_> ah I see the problem 21:27 < rxr> ok 21:27 < rxr> now uclibc can come again 21:28 < CIA-9> rene * r6864 /trunk/package/base/busybox/busybox.conf: * fixed busybox link installation and removed the #FIXME comment ;-) 21:28 < _Ragnar_> s,^\(# \)\?\([^ #]*\)[= ].*$,\2=y,g 21:28 < _Ragnar_> in config.sed 21:31 < CIA-9> rene * r6865 /trunk/target/psion-pda/config.in: 21:31 < CIA-9> * enabled busybox symlink creation for the pda target and removed no longer 21:31 < CIA-9> needed stand alone versions (like minised, ash, ...) 21:39 < rxr> _Ragnar_: any idea where you get that expression f from ? 21:39 < _Ragnar_> functions.in I'd say 21:42 < _Ragnar_> rule="${prerule//\*/[^ #]*}" <- 21:49 < rxr> _Ragnar_: can you mail me the config.sed ? 21:51 < _Ragnar_> .oO(how can this work for you?) 21:52 < rxr> no idea ... 21:52 < rxr> it does ... 21:52 < _Ragnar_> dcc chat? 21:52 < _Ragnar_> would be easier;) 21:52 < rxr> I'm not dcc trained ... 21:52 < _Ragnar_> oh;) 21:52 < rxr> and my client is running on y remote router anyway ... 21:53 < _Ragnar_> http://66.146.166.62/~ragnar/config.sub 21:54 < rxr> hm - mine is full 21:55 < _Ragnar_> does this have any bash special meaning: ${prerule//\*/[^ #]*} 21:56 < rxr> which part do you mean ? 21:56 < _Ragnar_> all of it;) 21:56 < rxr> hm 21:57 < rxr> maybe a slash to much ? 21:57 < _Ragnar_> I would understand ${prerule}... 21:57 < rxr> ${x/abc/123} 21:58 < rxr> is substitude the content of x from abc to 123 21:58 < rxr> man bash ;-) 21:58 < rxr> can you try: 21:59 < rxr> ${prerule/\*/[^ #]*} 21:59 < _Ragnar_> okay 22:00 < rxr> what system are oyu building ont ? 22:00 < rxr> bash2 ? 22:00 < _Ragnar_> no change 22:00 < _Ragnar_> GNU bash, version 2.05b.0(1)-release 22:01 < rxr> hehe ;-) 22:01 < rxr> bash3 here ... 22:01 < rxr> non the less it should work for both ... 22:02 < rxr> can you build with XTRACE and place the log on the http server ? 22:03 < _Ragnar_> http://66.146.166.62/~ragnar/ERROR-LOG 22:04 < _Ragnar_> I don't get what the substitute should do 22:10 < rxr> I think that one works ... 22:10 < rxr> can you add a echo "action: $action" 22:10 < rxr> echo "prerule: $prerule" 22:10 < rxr> echo "value: $value" 22:11 < rxr> between the while read and the rule= ? 22:11 < rxr> and update the error log ? 22:12 < _Ragnar_> *g* 22:12 < _Ragnar_> I already did that before 22:13 < _Ragnar_> ===[apply_conffiles:15 (last $?=0)> echo 'action: X prerule: UCLIBC_HAS_GLOB value: ' 22:13 < _Ragnar_> action: X prerule: UCLIBC_HAS_GLOB value: 22:13 < _Ragnar_> ===[apply_conffiles:16 (last $?=0)> rule=[^ #]* 22:13 < _Ragnar_> ===[apply_conffiles:17 (last $?=0)> echo 'rule: [^ #]*' 22:13 < _Ragnar_> rule: [^ #]* 22:14 < rxr> prerule=TARGET_* 22:15 < rxr> echo ${prerule/\*/[^ #]*} 22:15 < rxr> echo ${prerule/\*/[^ #]*} 22:15 < _Ragnar_> *lol* 22:15 < _Ragnar_> got it 22:16 < rxr> does it work ? 22:16 < _Ragnar_> no, sth else: 22:16 < _Ragnar_> ${prerule//\\*/[^ #]*} 22:17 < _Ragnar_> try that 22:18 < rxr> what does that yield for oyu ? 22:20 < _Ragnar_> s,^\(# \)\?\(UCLIBC_HAS_FLOATS\)[= ].*$,\2=y,g 22:20 < _Ragnar_> s,^\(# \)\?\(UCLIBC_HAS_[^ #]*\)[= ].*$,\2=y,g 22:20 < _Ragnar_> (in config.sub) 22:20 < _Ragnar_> sed* 22:20 < rxr> hm 22:20 < _Ragnar_> which actually matches and changes stuff in uC*/.config 22:20 < rxr> does not work for my bash 22:20 < rxr> and it is overquoted ... 22:21 < _Ragnar_> ? 22:21 < rxr> I have a bash2 now for testing ... 22:21 < rxr> damn that it behaves so different.y ... 22:21 < _Ragnar_> what is overquoted? 22:22 < _Ragnar_> the sed expr? 22:22 < rxr> more backslashes thatn needed 22:23 < _Ragnar_> hm 22:25 < _Ragnar_> s,^(# )?(TARGET_[^ #]*)[= ].*$,# \2 is not set,g 22:25 < _Ragnar_> s,^(# )?(CONFIG_[^ #]*)[= ].*$,# \2 is not set,g 22:25 < _Ragnar_> this does NOT work 22:25 < _Ragnar_> (removing the \) 22:27 < rxr> it works on bash3 and is what the expression should be ... 22:27 < rxr> I see on a bash2 myself that bash2 is buggy ... 22:27 < _Ragnar_> this is sed, not bash talkin 22:28 < rxr> alone the first double / 22:28 < rxr> the syntax is ${var/pattern/subst} 22:28 < rxr> not // ... 22:29 < _Ragnar_> huh? 22:29 < _Ragnar_> no! 22:29 < _Ragnar_> ${parameter/pattern/string} 22:29 < _Ragnar_> ${parameter//pattern/string} 22:29 < _Ragnar_> In the first form, only the first match is replaced. The second form causes all matches of pattern to be replaced with string. 22:30 * _Ragnar_ quotes man bash 22:30 < _Ragnar_> :D 22:30 < rxr> ok - yeah - well 22:31 < rxr> forgot about that 22:31 < rxr> but that is nto what makes the change anyway 22:31 < rxr> the changes between bash2 and 3 is the \\ before the * 22:31 < _Ragnar_> well pattern is interpreted as with pathname globbing 22:32 < _Ragnar_> and it seems bash2 'strips' out one \ for the " " string encapsulation 22:32 < _Ragnar_> lemme test that theory 22:33 < rxr> ok - somehow this is too fragile lets use s.th. else 22:33 < rxr> like sed for this substitution 22:35 < _Ragnar_> good idea 22:36 < rxr> echo "$(echo $prerule | sed 's/\*/[^ #]*/')" 22:37 < rxr> works for bash2 and 3 22:37 < rxr> ;-) 22:37 < rxr> can you test and I commit then ... 22:37 < _Ragnar_> *hrhr* okay 22:39 < _Ragnar_> looks good 22:39 < _Ragnar_> I used: rule=$(echo $prerule | sed 's/\*/[^ #]*/') <-- correct? 22:39 < rxr> why hr hr ? 22:39 < _Ragnar_> just felt like it 22:41 < rxr> ok 22:41 < rxr> dump a note when it built ... 22:42 < _Ragnar_> tar -chf - include/ | tar -xf - -C /export/t2-trunk/build/a32-2.1.0-beta4-x86-athlon-xp-generic-expert/usr/i386-linux-uclibc/usr/; 22:42 < _Ragnar_> tar: /export/t2-trunk/build/a32-2.1.0-beta4-x86-athlon-xp-generic-expert/usr/i386-linux-uclibc/usr: Cannot chdir: No such file or directory 22:42 < _Ragnar_> next error 22:43 < rxr> oehm 22:43 < rxr> can you svn up 22:43 < rxr> the first error is gone now ? 22:45 < _Ragnar_> yeah 22:46 < _Ragnar_> seems sed really really needs those backslashes;) 22:47 < _Ragnar_> okay 0-uclibc built 22:49 < rxr> - rule="${prerule//\*/[^ #]*}" 22:49 < rxr> + rule="$(echo $prerule | sed 's/\*/[^ #]*/')" 22:50 < rxr> can I commit that ? 22:58 < _Ragnar_> yea 22:58 < _Ragnar_> okay later on: 22:58 < _Ragnar_> In file included from ../../../include/locale.h:30, from ctype.c:38: 22:58 < _Ragnar_> ../../../include/bits/uClibc_locale.h:77:37: bits/uClibc_locale_data.h: No such 22:58 < _Ragnar_> file or directory 22:59 < rxr> hm 22:59 < rxr> I do not get that ... 23:00 < _Ragnar_> you have locale off? 23:05 < _Ragnar_> okay I removed the test 'X UCLIBC_HAS-*' from my config 23:05 < _Ragnar_> with that it compiles:) sorry 23:11 < rxr> do you have some add on config ? or do you mean the on in package/base/uclibc/... ? 23:11 < _Ragnar_> the latter 23:12 < _Ragnar_> I put a config with a * in it to test whether the subst really works 23:12 < rxr> mysterious 23:13 < CIA-9> rene * r6866 /trunk/package/base/uclibc/functions.in: 23:13 < CIA-9> * fixed uclibc for building with bash2 - the substitution does not 23:13 < CIA-9> yield the same result compared for bash2 vs. bash3 ... 23:31 < rxr> arm uclibc ssh client! 23:31 < rxr> 106816 Feb 27 20:08 dbclient 23:35 < _Ragnar_> :D 23:36 < CIA-9> rene * r6867 /trunk/package/base/busybox/busybox.conf: 23:36 < CIA-9> * fixed busybox installation to not add usr to the PREFIX for symlink 23:36 < CIA-9> installation - busybox distributes them itself over {,/usr}/{bin,sbin} 23:38 < CIA-9> rene * r6868 /trunk/target/psion-pda/config.in: 23:38 < CIA-9> * removed more "just for cross testing" packages from the pda target 23:38 < CIA-9> (no spare space on the Revo ;-) 23:45 < nosfod> is there any x86-64 iso to build t2? 23:45 < rxr> not yet - mipe is working on that ... 23:45 < nosfod> on debian x86-64 does not work 23:45 < rxr> you can however either build from a 32bit t2 - cross building ... 23:45 < rxr> and some other diststributions should work as base ... 23:45 < nosfod> hmm thats not documented - right? 23:46 < nosfod> i simply choose cross build? 23:47 < rxr> cross build should work nearly ever 23:47 < nosfod> ok i will try it 23:47 < rxr> when you are on a x86-64 box you shoudl be able to just build normally (that is not cross) 23:48 < nosfod> that stopped with an can not exe bin/bash in stage 2 23:48 -!- mtr__ [~michael@H99b6.h.pppool.de] has quit [Remote closed the connection] 23:48 < rxr> nosfod: that is not host system related then 23:48 < rxr> stage2 is entering the chroot 23:48 < rxr> so it is some problem with the freshly built t2 ... 23:49 < rxr> mipe maybe be a help there ... 23:49 < rxr> he had such an situation already a few times ... 23:49 < nosfod> but it should work with an x86 t2? 23:50 < nosfod> on an x86-64 23:50 < rxr> nosfod: youre build already boostraped the basic system it would continue the build in 23:50 < rxr> so that is all fine 23:51 < rxr> nope you have some dynamic linker or so problem mipe also had a few times 23:51 < rxr> guess he already know the fix or is working on it ... 23:51 < nosfod> rxr: ok i will ask later 23:52 < rxr> that would be great 23:52 < rxr> you might already do this: 23:54 < rxr> cd build/ 23:54 < CIA-9> rene * r6869 /trunk/package/graphic/libpng/libpng.conf: * cleanup of libpng.conf and fixed it for cross builds 23:54 < rxr> and tell me the content of etc/ld.so.conf 23:56 < nosfod> rxr: "lib lib64 usr/lib usr/lib64 usr/X11/lib usr/local/lib opt/*/lib" 23:57 < rxr> file bin/bash 23:57 < rxr> tells you it is a 64bit executable ? 23:58 < nosfod> yes 23:58 < nosfod> but the chroot in stage 2 fails 23:58 < rxr> strings /bin/bash | grep lib64/ 23:59 < rxr> sorry 23:59 < rxr> strings bin/bash | grep lib64/ 23:59 < rxr> nosfod: the error is s.th. like bash no such file or directory or so ? 23:59 < rxr> ah - can not exec you wrote ;-) 23:59 < rxr> what does the strings | grep yield ? --- Log closed Mon Feb 28 00:00:09 2005