| 00:10.41 | *** join/#openembedded qip|Virus|da (~Virus@pD9524705.dip.t-dialin.net) |
| 00:14.15 | gb2 | kergoth: you want to call our glibc package libc6? |
| 00:14.24 | kergoth | OZ uses debian naming |
| 00:14.25 | gb2 | that mean uclibc should be libc0 ? |
| 00:14.33 | CosmicPenguin | libc-1 |
| 00:14.39 | gb2 | libc-1 ? |
| 00:14.40 | kergoth | heh, good question |
| 00:14.49 | gb2 | it's libc.so.0 |
| 00:14.50 | CosmicPenguin | I guess that really doesn't work, does it |
| 00:15.09 | gb2 | -number is a little confusing potentially |
| 00:15.18 | kergoth | heh |
| 00:15.19 | gb2 | oh |
| 00:15.22 | gb2 | i get it |
| 00:15.27 | gb2 | :P |
| 00:16.31 | gb2 | kergoth: imo debian's naming is wrong, but that's me. |
| 00:17.45 | kergoth | yeah i'm inclined to agree, although it _is_ a good idea to change the package name to reflect the abi version (libc6 vs libc5) |
| 00:18.10 | gb2 | well, libc5 is not glibc. |
| 00:18.44 | kergoth | the point is, we want ncurses5 and ncurses4 instead of libncurses5 and libncurses4, correct? |
| 00:18.53 | gb2 | hm |
| 00:18.58 | kergoth | since they can both easily be installed |
| 00:19.01 | gb2 | those may be correct. |
| 00:19.03 | kergoth | we want them to be able to live side by side |
| 00:19.05 | gb2 | in my view |
| 00:19.17 | kergoth | heh, you just said their naming was wrong |
| 00:19.21 | gb2 | (i just carried 4 loads of laundry to the laundrimat so i'm a little out of things) |
| 00:19.43 | gb2 | kergoth: i'm of the opinion that their glibc package being named libc6 is wrong. |
| 00:19.48 | treke | I read that as being wrong in the case of glibc |
| 00:19.56 | treke | not a general incorrect naming policy |
| 00:19.59 | kergoth | exceptions arent good policy. |
| 00:20.09 | gb2 | the glibc people call it glibc |
| 00:20.09 | kergoth | if you say a package name is the lib soname, then thats what it should be |
| 00:20.21 | kergoth | and the ncurses people call it ncurses. |
| 00:20.22 | kergoth | whats your point? |
| 00:20.28 | gb2 | hrm |
| 00:21.03 | kergoth | i see why you disagree with it, but how can we retain the same functionality with an alternative |
| 00:21.10 | kergoth | and without ugly special casing distro policy |
| 00:21.29 | gb2 | some packages contain more than one library |
| 00:22.02 | kergoth | they shouldnt. |
| 00:22.05 | kergoth | xlibs is an aberration |
| 00:22.12 | kergoth | at least, the way debian has things now they shouldnt. |
| 00:22.27 | gb2 | how would you package something like say.. krb5 |
| 00:22.39 | *** join/#openembedded treke (~ggilbert@12.107.12.130) |
| 00:23.07 | kergoth | you'd break it up by .so and name them accordingly, with proper interpackage dependency |
| 00:23.20 | gb2 | even though you'll never want a subset of them? |
| 00:23.38 | kergoth | if you use debian's naming, yep. |
| 00:23.50 | kergoth | personally i hate the way linux handles shared libs in general |
| 00:23.52 | kergoth | just a mess. |
| 00:24.43 | gb2 | linux does it the same way other unixes tend to from what i've seen |
| 00:25.07 | kergoth | yep |
| 00:25.10 | kergoth | which is a mess |
| 00:25.23 | gb2 | does any OS do it better in your opinion? |
| 00:25.31 | kergoth | heh, probably not |
| 00:25.38 | gb2 | atleast it's versioned here |
| 00:26.51 | gb2 | wrt packaging, i'm of the opinion that things should be kept into units that the user will concievably want or need ot deal in. making it more granular without need ends up making stuff more confusing and unnecessary complexity |
| 00:27.07 | *** join/#openembedded dugong (~Miranda1@utip062.cs.utwente.nl) |
| 00:27.38 | *** part/#openembedded dugong (~Miranda1@utip062.cs.utwente.nl) |
| 00:27.40 | Twiun | it would be nice if you could have packages containing other packages |
| 00:27.51 | kergoth | gb2: what the user wants is dependent on the market |
| 00:27.53 | Twiun | rather than a deps package that pulls others in |
| 00:27.57 | kergoth | embedded needs a great deal more granularity |
| 00:29.02 | gb2 | kerboth: for something like krb5, there is nothing that needs only one of those libraries, so there isn't any use in packaging them separately. |
| 00:31.06 | kergoth | in which case naming the package by soname doesnt make any sense |
| 00:32.17 | CosmicPenguin | kernel oops suck |
| 00:33.54 | gb2 | i've also found debian tends to miss library dependencies because of their packaging. |
| 00:34.03 | kergoth | k, do_package strips now |
| 00:34.26 | kergoth | gb2: how so? dh_shlibdeps adds the dependency on the soname'd packagename |
| 00:34.32 | kergoth | nothin to it |
| 00:34.33 | kergoth | heh |
| 00:34.44 | gb2 | i've seen it get things wrong |
| 00:35.03 | kergoth | well, we'll be implementing an equivalent |
| 00:35.11 | kergoth | so here's hoping we do it better |
| 00:35.15 | gb2 | heh |
| 00:35.25 | gb2 | iirc, it got versioned symbols wrong. |
| 00:35.27 | mmelbert|laptop | good luck |
| 00:35.38 | kergoth | heh, what do symbols have to do with it? |
| 00:35.51 | kergoth | just needs to call readelf and match up what we link against to a packagename |
| 00:35.55 | kergoth | that simple |
| 00:36.24 | gb2 | kergoth: libraries can have versioned symbols |
| 00:36.27 | kergoth | .. |
| 00:36.31 | gb2 | here's a redhat example: |
| 00:36.32 | gb2 | annabel% rpm -q --requires bash |
| 00:36.36 | gb2 | libc.so.6(GLIBC_2.0) |
| 00:36.36 | gb2 | libc.so.6(GLIBC_2.1) |
| 00:36.36 | gb2 | libc.so.6(GLIBC_2.2) |
| 00:36.36 | gb2 | libc.so.6(GLIBC_2.3) |
| 00:36.43 | CosmicPenguin | ARgh |
| 00:36.56 | gb2 | i've had similar things where dpkg was happy installing a package |
| 00:37.00 | gb2 | and it wouldn't dynamically link |
| 00:37.11 | CosmicPenguin | God hates me for trying to commit crimes against nature |
| 00:37.14 | kergoth | thats expected. does dpkg even hold that information? |
| 00:37.20 | kergoth | dh_shlibdeps is only as good as the info in the db. |
| 00:37.22 | CosmicPenguin | Or more specifically, trying to shoe horn in a 800x600 kenrel logo |
| 00:37.46 | gb2 | kergoth: rpm's equivalent does versioned symbol dependencies properly |
| 00:37.51 | kergoth | thats one thing i prefer about rpm, it has the info on dependencies more granular than just package name |
| 00:38.19 | gb2 | it has proper deps on libraries rather than relying on the package naming and package versioning |
| 00:38.24 | kergoth | how hard is this to understand? if the dpkg db doesnt hold the info about what versioned lib i provide, dh_shlibdeps sure as hell isnt going to allow you to exert a depenency |
| 00:38.38 | kergoth | i'm saying its probably not dh_shlibdeps fault, but other aspects of hte system |
| 00:38.55 | gb2 | kergoth: annabel% readelf -V /bin/bash |
| 00:38.56 | gb2 | ... |
| 00:38.59 | gb2 | <PROTECTED> |
| 00:38.59 | gb2 | <PROTECTED> |
| 00:38.59 | gb2 | <PROTECTED> |
| 00:39.01 | gb2 | ... |
| 00:39.03 | treke | CosmicPenguin: Make him hate you for fixing the pixil build |
| 00:39.07 | kergoth | what about it? |
| 00:39.14 | gb2 | well .. *shrug* |
| 00:39.20 | gb2 | i guess it's a limitation of the dpkg format |
| 00:39.21 | kergoth | if i dont know WHAT PACKAGE PROVIDES GLIBC_2.3, i cant very well add it depends, now can i? |
| 00:39.36 | kergoth | yeah, i like that about rpm |
| 00:39.44 | kergoth | more granular than just package dependency |
| 00:40.07 | treke | of course it sucks when you are trying to track down what it depends on :) |
| 00:40.24 | gb2 | treke: what what depends on? |
| 00:40.44 | kergoth | its easy to know what libs you depend on, its not as easy to know what package satisfies it |
| 00:40.47 | gb2 | which package corresponds to the lib depend? |
| 00:41.06 | treke | yes |
| 00:41.13 | gb2 | annabel% rpm -q --whatprovides 'libc.so.6(GLIBC_2.3)' |
| 00:41.13 | treke | or any file dependency |
| 00:41.14 | gb2 | glibc-2.3.2-101.1 |
| 00:41.15 | gb2 | that? |
| 00:41.29 | kergoth | yep, that. i dont think dpkg's db holds that information |
| 00:41.41 | mallum | treke: my hack dont work :/ |
| 00:41.50 | treke | mallum: no? |
| 00:41.53 | treke | :( |
| 00:42.04 | mallum | treke: fbset still shows wrong info :( |
| 00:42.08 | CosmicPenguin | treke: something brtoke? |
| 00:42.45 | mallum | kergoth: know much about fb code ? |
| 00:42.57 | kergoth | not much |
| 00:43.05 | treke | CosmicPenguin: Yeah. Pasted earlier. It doesnt like the microwindows.mk makefile. Missing seperator at 151, if you add a : to the end of the line it gives me an unterminated variable reference at 158 |
| 00:43.19 | Twiun | mallum: fbset should be querying the kernel's fb stuff via ioctls |
| 00:43.21 | gb2 | the other thing i ran into was things like mozilla with their own pile of shared objects weren't handled well. |
| 00:43.35 | Twiun | mallum: so if the info is wrong, that's the place to start tracing |
| 00:43.44 | mallum | Twiun: right, but the rotation code is not updating the FBIOGET_VSCREENINFO info |
| 00:44.11 | mallum | Twiun: the screen is rotating but the var stuff is not getting updated |
| 00:44.20 | Twiun | mallum: from memory, some of those info fields can be non-retrieveable altogether - hmmm |
| 00:44.20 | treke | CosmicPenguin: you use gmake? |
| 00:44.29 | Twiun | mallum: is the kernel actually updating that state internally? |
| 00:44.37 | mallum | Twiun: all I need is xres and yres |
| 00:44.59 | Twiun | mallum: it's quite possible the kernel isn't being told about the switch |
| 00:45.14 | mallum | Twiun: it wasn't. I hacked it so it should. But Im not 100% sure what call a FBIOGET_VSCREENINFO makes in the kernel code |
| 00:45.14 | Twiun | mallum: or at least not the fb code |
| 00:45.29 | mallum | Twiun: also the w100fb code is a complete abortion ;-) |
| 00:45.36 | Twiun | mallum: ah. |
| 00:45.53 | Twiun | mallum: is that code in a viewcvs repo somewhere? I've had a look at the fb code for the 5000d |
| 00:47.11 | mallum | Twiun: see http://handhelds.org/~mallum/w100fb.c |
| 00:47.17 | kergoth | gb2: hm, seems like just saying somethinig provides libc.so.6(GLIBC_2.3) is iffy. i mean, what if i installed that shared lib somewhere outside my dirs in my ld.so.conf and/or LD_LIBRARY_PATH .. doesnt very well provide it anymore.. |
| 00:47.22 | kergoth | heh |
| 00:47.22 | kergoth | ugly |
| 00:48.34 | mallum | Twiun: w100fb_ioctl() does the actual rotation |
| 00:48.43 | Twiun | 'k |
| 00:49.02 | kergoth | pushed the altered package class that strips the binaries |
| 00:49.28 | kergoth | forgot to check if the package doesnt want to be stripped though |
| 00:49.33 | kergoth | we really need a USE equivalent |
| 00:49.45 | mallum | Twiun: I just hacked it to set the current_par values |
| 00:49.54 | Twiun | yeah |
| 00:50.05 | mallum | Twiun: I thoght that should do it but it dont |
| 00:50.17 | Twiun | mallum: those should be calls to the encode/decode functions, right? |
| 00:50.44 | mmelbert|laptop | did something happen to zaurus.com?? The domain doesnt seem to be resolving for me |
| 00:50.56 | mallum | Twiun: I dunno. its either them or the get_var calls |
| 00:51.20 | treke | I'm guessing fb_dO_set_var is what you want |
| 00:52.18 | Twiun | i'm amazed at the amount of string manipulation code in there |
| 00:53.33 | treke | mallum: who did the ipaq fb code? |
| 00:54.14 | mallum | treke: I think the original stuff was probably rmk, newer kiethp |
| 00:54.27 | mallum | treke: on the 5xxx's |
| 00:54.31 | treke | kiethp might know |
| 00:54.47 | treke | he did the mediaq one right? |
| 00:55.11 | mallum | treke: right |
| 00:55.42 | kergoth | there we go, full build of stripped tarballs |
| 00:55.44 | Twiun | yeah, none of this stuff is familiar anymore - I'd need to debug it myself ;) |
| 00:55.56 | Twiun | kergoth: nice :) what's next on the list? |
| 00:56.48 | mallum | treke: I'll ask jamey tommorow, he's usually most helpful |
| 00:56.57 | kergoth | Twiun: i need to push my changes for the fetch class to obey md5sum= and size=, since i never cleaned it up and pushed it.. |
| 00:57.08 | treke | yeah. I'm sure someone in the bunch knows about the fb |
| 00:57.10 | kergoth | then.. lets see.. |
| 00:57.16 | kergoth | probably an ipk installer for oeinstall |
| 00:57.19 | mallum | treke: pb_ would probably know too, but you may have more luck asking him |
| 00:57.22 | treke | just wondering if there is any obvious code in the mq module |
| 00:59.16 | mallum | treke: yeah though Im not sure current mq does any resizing / rotation |
| 00:59.50 | treke | actually... |
| 00:59.52 | treke | it looks like it does |
| 01:00.03 | treke | MqDataToFormat |
| 01:00.33 | mallum | treke: you have a url for the code ? |
| 01:00.51 | mallum | treke: also who is the person that hacks on xine for oz/oe ? |
| 01:01.07 | kergoth | didnt simon hausmann pull that in? |
| 01:03.10 | mallum | kergoth: btw in oe how do I now do a make mrproper ( just removing stamps ) ? |
| 01:03.21 | kergoth | rm -rf tmp |
| 01:03.24 | kergoth | ;) |
| 01:03.24 | mallum | kergoth: and also how do I uninstall a package from the image ? |
| 01:03.33 | kergoth | you dont, not the way it is right now |
| 01:03.41 | kergoth | right now oeinstall just unpacks tarballs into it |
| 01:03.45 | kergoth | no ipk installation support yet |
| 01:04.00 | mallum | I see |
| 01:04.32 | kergoth | not all types will support uninstallation.. ipk obviously will.. but tarball cant since it doenst check to see if we inadvertantly replaced files from another package |
| 01:04.49 | kergoth | it doesnt check for file conflicts, so we shouldnt arbitrarily remove the files we put there |
| 01:05.22 | mallum | kergoth: I always like the image building part of oz where you could select your packages then build the image |
| 01:05.29 | kergoth | .. |
| 01:05.37 | kergoth | you have that right now with oe, via INST_OEPKGS |
| 01:05.42 | kergoth | just no gui to populate it for you yet |
| 01:06.06 | treke | mallum: OE is going to have all the pretty gui stuff, just hasnt been writte yet |
| 01:06.25 | kergoth | yeah, making it work is the focus atm |
| 01:06.54 | mallum | right, I understand. Dont think I bleeting btw. Just finding out how to do stuff without GUI |
| 01:07.25 | kergoth | echo 'INST_OEPKGS=glibc netbase ifupdown base-files busybox' >> conf/local.conf |
| 01:07.29 | kergoth | or whatever |
| 01:07.33 | gb2 | kergoth: S has a sane default now right? |
| 01:07.40 | gb2 | is it ${WORKDIR}/${PN}-${PV} ? |
| 01:07.41 | kergoth | gb2: yep, ${WORKDIR}/${P} is default |
| 01:07.44 | kergoth | yep |
| 01:07.45 | gb2 | ah |
| 01:07.46 | gb2 | P |
| 01:08.01 | mallum | kergoth: also should I use packages/conf/local.conf or oe/conf/local.conf ? |
| 01:08.07 | kergoth | mallum: neither. |
| 01:08.12 | gb2 | when does it get set? |
| 01:08.13 | kergoth | you use conf/local.conf relative to where you're buliding |
| 01:08.21 | mallum | right |
| 01:08.21 | kergoth | gb2: its in oe.conf. |
| 01:08.27 | gb2 | ok. |
| 01:08.33 | gb2 | i'm wondering about includes |
| 01:08.41 | gb2 | i'm toolchain hacking again |
| 01:08.45 | kergoth | ah |
| 01:17.34 | gb2 | kergoth: as far as you know doing FILES_libstdc++ = blah... |
| 01:17.36 | gb2 | still fails right? |
| 01:21.59 | CosmicPenguin | treke: no |
| 01:23.03 | treke | CosmicPenguin: that could be the problem |
| 01:23.25 | CosmicPenguin | treke: d'oh |
| 01:23.31 | treke | I'm using gmake |
| 01:23.41 | treke | What make do you use? |
| 01:25.07 | *** join/#openembedded Ruied_Zaurus (~Ruied_Zau@81.84.147.82) |
| 01:25.21 | CosmicPenguin | Well - I guess it is gmake |
| 01:25.23 | CosmicPenguin | shiver me timbers |
| 01:27.22 | CosmicPenguin | shiver them! |
| 01:28.11 | CosmicPenguin | whats it doing? |
| 01:28.38 | treke | ggilbert@dreamland:~/pixil/nanoxdev-1.1$ make |
| 01:28.38 | treke | /home/ggilbert/pixil/nanoxdev-1.1/make/microwin.mk:158: *** unterminated variable reference. Stop. |
| 01:28.47 | treke | wait. that is after a modification |
| 01:29.18 | treke | ggilbert@dreamland:~/pixil/nanoxdev-1.1$ make |
| 01:29.18 | treke | /home/ggilbert/pixil/nanoxdev-1.1/make/microwin.mk:151: *** missing separator. Stop. |
| 01:29.36 | CosmicPenguin | ahh, shit |
| 01:29.50 | *** join/#openembedded chewy (~chouimat@r2351064.cidc.net) |
| 01:30.02 | CosmicPenguin | toss a colon on the end of that line |
| 01:30.12 | treke | then you get the first error |
| 01:30.17 | treke | That's the modification I made :) |
| 01:30.21 | chewy | hmmm how to kill someone who have my nick? |
| 01:30.38 | treke | <PROTECTED> |
| 01:30.40 | CosmicPenguin | Always test before you submit, dumbass |
| 01:30.46 | treke | hehe |
| 01:30.56 | chewy | thanks treke |
| 01:31.22 | CosmicPenguin | Ahh - it has a bracket instead of a paren |
| 01:31.25 | CosmicPenguin | I blame OE... :) |
| 01:31.31 | *** mode/#openembedded [+o chouimat] by ChanServ |
| 01:31.31 | mallum | ah bedtime, night all |
| 01:31.33 | treke | oops |
| 01:31.39 | chouimat | better |
| 01:31.41 | treke | maybe I need glasses |
| 01:31.55 | CosmicPenguin | Thats ok - I didn't see it the first time I looked either |
| 01:34.23 | CosmicPenguin | Ahh, I see the problem now - Our fine feathered kernel friends don't allow the console to overwrite the logo |
| 01:34.34 | CosmicPenguin | So when the logo is the size of the screen, that causes pain and/or suffering |
| 01:35.34 | treke | haha |
| 01:35.49 | treke | 2.4.6 has a patch for that in buildroot, I thin |
| 01:36.41 | CosmicPenguin | I'm trying to figure out what my previous patch did that I'm not doing now |
| 01:41.54 | kergoth | gb2: it shouldnt fail, it just wont export that var into shell. it should still become FILES fine |
| 01:42.04 | kergoth | iirc i'm stripping out - and + vars |
| 01:42.05 | treke | http://www.meow.org.uk/stan/xserver/dumps.html |
| 01:42.09 | kergoth | er vars that contain - and + that is |
| 01:42.52 | CosmicPenguin | oh, god help us |
| 01:43.50 | kergoth | treke: haha |
| 01:46.55 | gb2 | kergoth: ok |
| 01:47.12 | *** join/#openembedded andersee (~andersee@codepoet.org) |
| 01:48.47 | treke | ok. next problem is that it doesnt download the nxlib source |
| 01:51.04 | CosmicPenguin | probably because of my spelling |
| 01:51.17 | CosmicPenguin | by mainly because my code sucks |
| 01:54.52 | treke | duh |
| 01:55.14 | treke | you never actually wget the source. |
| 01:55.38 | treke | oh wait. You do. |
| 01:55.49 | treke | NXLIB_SITE=ftp://ftp.microwindows.org/pub/microwindows/$(NXILB_SOURCE) |
| 01:55.54 | treke | NXILB |
| 01:57.49 | treke | but of course sourceforge is down so I cant download the pixdev package |
| 01:58.30 | *** join/#openembedded ictyl (bradford@user-0ce2ku9.cable.mindspring.com) |
| 02:02.31 | CosmicPenguin | gah |
| 02:07.50 | chouimat | CosmicPenguin: go do something else |
| 02:09.21 | Disconnect | kergoth: <ahem> fixed those gpl violations yet? |
| 02:09.39 | chouimat | Disconnect: ? |
| 02:10.36 | Disconnect | chouimat: ask for the sources used for 3.3.5 .. complete and in the "format preferred for making modifications" ... |
| 02:10.52 | treke | And what is the problem there? |
| 02:11.25 | kergoth | did mickey tag 3.3.5 yet? |
| 02:11.31 | chouimat | treke: problably the same as my watchout product ... some part is gpl but no customer asked for the source yet ... |
| 02:11.44 | Disconnect | kergoth: not sure yet. lemme check. |
| 02:11.45 | treke | beats me |
| 02:11.57 | treke | just give me the command to tag it and I can do it |
| 02:12.03 | Disconnect | nope |
| 02:12.10 | Disconnect | ibot: 3.3.5 |
| 02:12.11 | | 3.3 |
| 02:12.26 | chouimat | kergoth: QNX 6.2.1 is installing in vmware and it's easy to do |
| 02:12.29 | Disconnect | damnit. (he gave the revision # to ibot, but I don't recall what the keyword is) |
| 02:12.50 | Disconnect | treke: bk tag -r<revision> v3.3.5 would do it, similar in style to kernels (v#.#.#-rc# etc) |
| 02:14.25 | treke | fine. tagged |
| 02:15.53 | Disconnect | cool. now go try to make it build.. |
| 02:16.07 | treke | I already have. |
| 02:16.53 | Disconnect | considering it doesn't build according to the docs, and hte sources to several of those patches aren't available anymore, thats surprising. |
| 02:17.27 | treke | Maybe the docs are wrong. |
| 02:17.36 | treke | And The source to those patches are available |
| 02:17.41 | Disconnect | (the one that comes to mind most readily is the debian ext2 patches, which are no longer on any mirrors.) |
| 02:17.56 | CosmicPenguin | He's not imagining that one |
| 02:18.05 | CosmicPenguin | I've seen it too |
| 02:18.07 | Disconnect | and btw since the opie checkout is a branch, not a tag, that changes underfoot as well. |
| 02:18.21 | Disconnect | (at least it looked like a branch, I haven't confirmed that since kergoth was going to fix it all and roll 3.3.6 ASAP) |
| 02:18.24 | treke | CosmicPenguin: You just have to point it at a mirror that has the package. |
| 02:18.40 | CosmicPenguin | treke: ok. I tried a few, and I figured that the mirrors were all the same... you know, mirrors.. :) |
| 02:18.51 | Disconnect | treke: put e2fsprogs_1.33-4.diff.gz into google. it shows all the deb mirrors, which don't have it anymore. thats not "available". |
| 02:18.54 | treke | Nope. There are mirrors specifically for this purpose. |
| 02:19.36 | treke | Disconnect: archives.debian.net iirc |
| 02:19.36 | Disconnect | treke: and if you want to be strict, buildroot should have -working- mirrors in it.. (to be really strict, I should be able to just dl a snapshot with sources included) |
| 02:19.51 | treke | That would be nice :) |
| 02:20.13 | Disconnect | treke: and the toolchain? and I recall something about "I just grabbed <some library> and threw it in".. sources for that build of "some library"..? (sdl maybe?) |
| 02:20.21 | treke | your getting the source as we build it. |
| 02:20.37 | kergoth | Disconnect: sounds like stdc++, which he pulled from his chain |
| 02:20.42 | kergoth | (mickey) |
| 02:21.10 | mmelbert|laptop | hey guys...where is a good place to start learning about linux kernels, writting device driver etc. I know google......but where have you all got your start? |
| 02:21.16 | treke | buildroot doesnt build straight through, it has to be done by hand quite a bit. |
| 02:21.19 | Disconnect | mmelbert|laptop: I got started with a tivo ;) |
| 02:21.28 | *** join/#openembedded mmelbert|laptop (~mmelbert@h-141-153-101-53.ckb.meer.net) |
| 02:21.34 | treke | That is now the preferred way of making changes :) |
| 02:21.42 | chouimat | and a Dreamcast |
| 02:21.45 | treke | http://snapshot.debian.net/ is the mirror I use. |
| 02:21.59 | mmelbert|laptop | hey guys...where is a good place to start learning about linux kernels, writting device driver etc. I know google......but where have you all got your start? |
| 02:22.09 | kergoth | mmelbert|laptop: read the free linux device drivers book is a good starting point, as well as reading code |
| 02:22.42 | mmelbert|laptop | is that by O'Reilly?? |
| 02:22.47 | kergoth | yep |
| 02:23.27 | mmelbert|laptop | thanks |
| 02:23.31 | CosmicPenguin | find something important, and make it work |
| 02:23.32 | mmelbert|laptop | any other tips? |
| 02:23.43 | CosmicPenguin | there is no motivation like having a vital piece of hardware in front of you |
| 02:23.50 | mmelbert|laptop | that is true |
| 02:23.53 | kergoth | CosmicPenguin: indeed |
| 02:23.54 | Disconnect | from the gpl: The source code for a work means the preferred form of the work for making modifications to it. For an executable work, complete source code means all the source code for all modules it contains, plus any associated interface definition files, plus the scripts used to control compilation and installation of the executable. |
| 02:23.54 | chouimat | mmelbert|laptop: get the embedded linux book by o'reilly too |
| 02:24.07 | Disconnect | show me the scripts used to make 3.3.5 .. |
| 02:24.13 | treke | Disconnect: It's right there in buildroot. |
| 02:24.14 | treke | Tagged. |
| 02:24.19 | Disconnect | show me the sources for the modules it contains. |
| 02:24.47 | Disconnect | (the 'core OS components' exception is unless the binary accompanies those components.. which it does) |
| 02:24.58 | treke | fuck this. |
| 02:25.11 | kergoth | Disconnect: per discussions with the fsf folks, its valid to have a system like this that obtains the sources from upstream. the problematic piece is that we dont control any of hte upstream urls, so its possible that at a given point the sources have been removed and we're therefore violating |
| 02:25.27 | chouimat | treke: I think he's currently following the GPL to the letter |
| 02:25.31 | treke | chouimat: Well fuck him |
| 02:25.44 | chouimat | treke: go first ... I don't like guys ;) |
| 02:26.12 | kergoth | treke: he's correct. i've discussed the situation with fsf people. if we maintained a source mirror, the problem wouldnt exist. |
| 02:26.17 | treke | If he wants to request a fucking specific bit of source fine. |
| 02:26.27 | treke | Until then he can deal. |
| 02:26.29 | chouimat | treke: i was joking btw |
| 02:26.30 | gb2 | kergoth: ++ worked btw |
| 02:26.31 | treke | kergoth: And I do mirror this shit |
| 02:26.34 | kergoth | gb2: good |
| 02:26.51 | kergoth | treke: he's requested the sources. hand him the url |
| 02:26.58 | Disconnect | treke: my issue isn't the gpl violation. my issue is, simply, that 3.3.5 is not reproducable. I CANNOT GET SOURCES. that means I can't test patches. and I can't contribute. but nobody gives a fuck to "ok, someone post fucking directions on how 3.3.5 was built" .. |
| 02:27.08 | treke | Disconnect: it is reproducible |
| 02:27.48 | Disconnect | really? I follow the directions and use the scripts and I get "oh, thats a toolchain error, we used some other toolchain. and that one isn't versioned. but we copied files out of it and nobody knows when it was downloaded, so it could be any of 3 different revisions.." |
| 02:27.58 | treke | Disconnect: The directions are wrong. |
| 02:28.14 | treke | Disconnect: Download the sources it requests by hand and put them, in sources. |
| 02:28.17 | kergoth | the gpl doesnt say you have to use the project's official toolchain |
| 02:28.18 | Disconnect | fine. what day was the release toolchain downloaded from hh.o? |
| 02:28.21 | kergoth | use the one mickey used |
| 02:28.23 | chouimat | kergoth: a source mirror like *BSD/Gentoo distfiles directory? |
| 02:28.25 | kergoth | ask him |
| 02:28.28 | treke | Disconnect: We dont know. it doesnt matter. |
| 02:28.29 | kergoth | chouimat: yep |
| 02:28.42 | treke | Disconnect: We dont have to provide you with sources for every component on our systems |
| 02:28.50 | treke | wget was used to download stuff, do you want the source for that? |
| 02:28.52 | Disconnect | treke: it does matter. there are bins straight out of the toolchain. there are changes in how code compiles based on whcih chain was used. |
| 02:29.09 | chouimat | kergoth: and we use it as the primary source for buildroot and oe |
| 02:29.09 | Disconnect | treke: wget was used from my local machine, if I want those sources I talk to debian.org |
| 02:29.34 | kergoth | Disconnect: you have the sources that match our binaries and the scripts associated. since when does the GPL say the binaries used to build have to be available? |
| 02:29.38 | kergoth | oh thats right, it doesnt. |
| 02:29.42 | treke | Disconnect: and the toolchain used was from your local machine also. |
| 02:31.00 | Disconnect | now look, chanserv stroked out. |
| 02:31.45 | mmelbert|laptop | chouimat: is it Building Embedded Linux Systems |
| 02:31.45 | mmelbert|laptop | by Karim Yaghmour ?? |
| 02:31.55 | chouimat | wait |
| 02:31.57 | Disconnect | treke: toolchain came from oz/official/ .. I'm not even gonna wonder if the sources for that are there. I'm guessing either in OE or buildroot, but I could give a fsck about that. |
| 02:32.05 | CosmicPenguin | I just don't understand how any of this is OE's problem. Source code available, scripts available. End of story. |
| 02:32.34 | kergoth | Disconnect: once again, i'm not obligated to release the source to the gcc i used to build binaries. it isnt a derived work, therefore availability of hh.o's toolchain as of that date isnt my responsibility. |
| 02:32.34 | Disconnect | CosmicPenguin: source available, if it didn't change. at a minimum it needs to be documented where it is, since some of the mirrors provided have gone away. |
| 02:32.45 | chouimat | mmelbert|laptop: yes ... |
| 02:33.05 | Disconnect | kergoth: you are obligated to release the source to the gcc you distributed in official/toolchain. and again, I don't care about gcc source. I'm going on the assumption its there. |
| 02:33.10 | kergoth | if you have buildroot as of the correct point, and the sources that went with it, whether you can get it to build is irrelevent. |
| 02:33.30 | treke | just download your fucking source at http://home.treke.net/sources-3.3.5.tgz |
| 02:33.38 | Disconnect | kergoth: fine. where are the sources? I was told bk pull etc. (btw you need to make a snapshot available as of that point, plus the configs that he added later. but thats not at issue, I have bk.) |
| 02:33.38 | kergoth | my point is, the opie failure to copmile with the chain you're using isnt my responsibility. |
| 02:33.49 | chouimat | treke: calm down |
| 02:33.52 | treke | chouimat: blow me |
| 02:34.00 | Disconnect | treke: put that in the 3.3.5 docs and its done. |
| 02:34.00 | kergoth | no, i dont. if the sources that match the binaries are available, how i distribute them is my choice |
| 02:34.12 | chouimat | treke: go out and find a nice lady for doing this |
| 02:34.33 | Disconnect | kergoth: bk is not "customarily used to distribute software". because not everyone can use it. |
| 02:34.58 | kergoth | oh i see what you meant, a 'bk export' from the tag |
| 02:35.00 | chouimat | kergoth: you just need to distribute them if the version is not available any more on the master site |
| 02:35.16 | Disconnect | right. plus that one config, which falls under scripts but was not present at the rev that was tagged. |
| 02:35.17 | kergoth | treke: can you bk export -rv3.3.5 a buildroot and toss it up in ~bk? |
| 02:35.46 | kergoth | along with altering the 3.3.5 def-configs to pull v1_0_3 opie tag |
| 02:36.16 | gb2 | kergoth: what's the difference between tmp/base/${P}/image and install? |
| 02:36.22 | kergoth | hm, windows machines dont answer icmp echo requests on 224.0.0.1, thats odd |
| 02:36.37 | kergoth | gb2: image is do_install output, install is split by package (result of the package task) |
| 02:36.46 | gb2 | ok |
| 02:36.47 | kergoth | gb2: unintuitive, if you come up with better names, let me know |
| 02:36.51 | gb2 | done by mv'ing files? |
| 02:37.01 | kergoth | yep |
| 02:37.06 | kergoth | though we should probably be copying, not moving |
| 02:37.12 | kergoth | since it should be possible to have two subpackages own the same file |
| 02:37.19 | gb2 | or linking |
| 02:37.19 | Disconnect | kergoth: s/image/target/ and s/install/contents/? (or package-contents?) |
| 02:37.45 | gb2 | kergoth: though.. i can easily do "find image" to find stuff i forgot to package |
| 02:37.58 | kergoth | gb2: yeah, i've been doing that.. its hand |
| 02:38.00 | kergoth | y |
| 02:38.07 | kergoth | find tmp/base/*/image -type f |
| 02:38.08 | kergoth | heh |
| 02:39.16 | CIA | repository=03buildroot user=11kergoth@direwolf.ppp.ti.com |
| 02:39.16 | CIA | 07ChangeSet@1.1376 The 3.3.5 def-configs should be using hte v1_0_3 opie tag, not the tip of the branch. |
| 02:41.12 | *** join/#openembedded ChanServ (ChanServ@services.) |
| 02:41.12 | *** mode/#openembedded [+o ChanServ] by tolkien.freenode.net |
| 02:42.34 | gb2 | kergoth: odd.. i found some that are in both image and install |
| 02:43.12 | kergoth | hmm weird |
| 02:43.27 | Disconnect | kergoth: afaict (without checking) the sources and a tgz with the bk tree from that tag (plus configs and that last revision) solve the gpl issue. (that wasn't all that hard ;) ..) the only remaining issue is if libgcc (etc) was shipped, in which case the toolchain used to build it does matter. |
| 02:44.00 | kergoth | yep, indeed |
| 02:44.13 | Disconnect | um. libgcc_s.so.1 .. |
| 02:45.12 | kergoth | which 3.3.5 images are gcc3 and which are 2? |
| 02:46.43 | Disconnect | looks like everything except 5000/5500 |
| 02:46.48 | Disconnect | (is gcc3 I mean) |
| 02:47.15 | kergoth | hm, so the other images dont care about compatibility.. since sharp already threw compatibility out the window due to their extensive changes |
| 02:47.20 | kergoth | bleh |
| 02:47.20 | mmelbert|laptop | well....I just downloaded the linux device driver book |
| 02:47.51 | mmelbert|laptop | now to see if the building linux embedded is out there |
| 02:48.43 | chouimat | mmelbert|laptop: no I bought it |
| 02:49.07 | gb2 | hrm. |
| 02:49.10 | gb2 | how annoying |
| 02:49.24 | gb2 | binutils uses CC_FOR_BUILD |
| 02:49.27 | gb2 | with CFLAGS |
| 02:49.29 | chouimat | hmmmm Nanoxdev 1.1 is now ready to terrorize the streets and scare small children |
| 02:49.36 | kergoth | gb2: ugh, i hate it when packages do that |
| 02:49.38 | gb2 | which we add -I/path/to/our/c/headers |
| 02:49.51 | gb2 | kergoth: CFLAGS_FOR_BUILD ? |
| 02:49.57 | CosmicPenguin | chouimat: and terrorize and scare it did |
| 02:50.01 | kergoth | check the makefile rule |
| 02:50.06 | CosmicPenguin | is sourceforge back up? |
| 02:50.08 | chouimat | CosmicPenguin: hehe |
| 02:50.13 | chouimat | CosmicPenguin: seems |
| 02:50.23 | gb2 | kergoth: wondering if that was roughtly how you fixed it |
| 02:50.30 | kergoth | iirc yes |
| 02:50.58 | kergoth | sometimes the buildsystem is complex, and i dont want to add the var to the configure.in and/or Makefile.* |
| 02:51.08 | kergoth | in which case i cheat and use BUILD_CC to copmile the binary before even calling the normal oe_runamke |
| 02:51.12 | kergoth | :) |
| 02:51.24 | chouimat | CosmicPenguin: nanoxdev-1.1.tar.gz is only 5KB? |
| 02:52.46 | mmelbert|laptop | chouimat: yup....looks like I have to buy that one |
| 02:53.45 | CosmicPenguin | chouimat: when did you get it? |
| 02:53.54 | CosmicPenguin | chouimat: I just uploaded the working version..... |
| 02:53.57 | CosmicPenguin | thanks to treke |
| 02:55.08 | chouimat | CosmicPenguin: about 3 minutes ago |
| 02:56.42 | gb2 | kergoth: heh |
| 02:57.04 | gb2 | ERROR: PARSER: gcc-uclibc_3.3.2.oe:21: unparsed line |
| 02:57.07 | gb2 | yay, my error |
| 02:57.22 | gb2 | uhoh |
| 02:57.24 | gb2 | it's a comment |
| 02:57.35 | kergoth | heh |
| 02:57.53 | gb2 | comments don't increment line count |
| 02:58.04 | gb2 | the error was the number of comment lines previous further down |
| 02:58.25 | Disconnect | kergoth: this toolchain issue, btw, is why I was pleased (and quiet) after you said you'd just fix it by making a 3.3.6 that would build (or even mostly build) straight out of buildroot with a (possibly-new) official toolchain. |
| 02:59.00 | kergoth | yeah, figured |
| 02:59.31 | kergoth | working on it, just no time to do so |
| 02:59.44 | kergoth | particularly with this deadline tuesday at work |
| 02:59.46 | Disconnect | (worst case at that point is you pull 3.3.5 downloads and say "look, sorry, we can't nail down the sources. its a bummer. but we released 3.3.6 to make up for that." or some such .. if 3.3.6 is there and complete, nobody is gonna have much of a fit over it) |
| 03:00.53 | chouimat | CosmicPenguin: where is pixildev-1.0? |
| 03:01.07 | CosmicPenguin | chouimat: pixdev-stable-1.0 is on SF |
| 03:01.15 | CosmicPenguin | 0.1 |
| 03:01.36 | CosmicPenguin | The new pixdev-stable is currently experiencing some technical difficulties |
| 03:01.53 | chouimat | CosmicPenguin: ok because 1.0 is announced on freshmeat |
| 03:04.03 | CosmicPenguin | dammit |
| 03:04.07 | CosmicPenguin | I hate freshmeat |
| 03:04.53 | CosmicPenguin | Dam |
| 03:04.55 | CosmicPenguin | damn damn |
| 03:05.08 | CosmicPenguin | I should've done two seperate packages - I tried to get all cute with it, and it ended up pissing me off |
| 03:05.34 | chouimat | CosmicPenguin: did thing too fast? |
| 03:06.23 | gb2 | kergoth: is it possible to inhibit package building entirely? |
| 03:06.34 | gb2 | setting PACKAGES= works, but it still tries |
| 03:07.14 | gb2 | NOTE: Not creating empty archive for gcc-uclibc-cross-3.3.2-r0 |
| 03:07.34 | kergoth | oh, skip the package task? |
| 03:07.38 | kergoth | dont inherit package_anything |
| 03:07.43 | gb2 | i get things like that which are ok for now, but it might be nice to make that a fatal error |
| 03:07.45 | kergoth | hehe |
| 03:09.36 | CosmicPenguin | chouimat: yeah, well its done now. All I can do is apologize, and release the next version as 1.1 to set us back on track |
| 03:09.39 | CosmicPenguin | Or not |
| 03:11.57 | CosmicPenguin | And my frickin' PHP code sucks too |
| 03:12.15 | kergoth | heh |
| 03:13.22 | chouimat | CosmicPenguin: look go see your wife for the rest of the day ... that will chaneg your mind |
| 03:13.36 | CosmicPenguin | chouimat: she has the flu |
| 03:13.59 | chouimat | CosmicPenguin: hmmm ok ... |
| 03:15.31 | chouimat | ok qnx is now patched to work A1 in vmware ... next OS/2 warp4 and BeOS |
| 03:22.30 | gb2 | ok.. i want to turn a .am into a .in |
| 03:22.35 | gb2 | i want to run.. automake? |
| 03:24.48 | treke|home | Ok. Vons employees can all go to hell and die. |
| 03:24.52 | kergoth | gb2: yep |
| 03:25.20 | gb2 | any flags that are good? |
| 03:25.31 | kergoth | umm, read autotools.oeclass? |
| 03:25.32 | kergoth | heh |
| 03:25.36 | kergoth | dont recall offhand |
| 03:25.37 | CosmicPenguin | Gah - I'm in a room on efnet, and they're all watching the Video Game Awards on Spike TV and thinking they are all geeky |
| 03:26.13 | CosmicPenguin | So I asked if they have given the lifetime achivement award to Bresenham yet. Nobody got it |
| 03:26.41 | treke|home | now to commence with the getting drunk. |
| 03:28.28 | chouimat | CosmicPenguin: here on spike it's chicks wrestling |
| 03:28.40 | CosmicPenguin | chouimat: thats the same show |
| 03:28.43 | CosmicPenguin | :) |
| 03:29.01 | chouimat | CosmicPenguin: shit! I have wrestling ... it's all fake |
| 03:29.01 | kergoth | god i hate building glibc |
| 03:29.06 | chouimat | s/have/hate |
| 03:29.45 | treke|home | CosmicPenguin: did you ever commit the gcc3 fixes to pixdev cvs? |
| 03:30.17 | gb2 | kergoth: know any way of turning some ${path} into the right number of ../..'s ? |
| 03:30.26 | CosmicPenguin | treke|home: funny story. the Sourceforge CVS was down.... :) |
| 03:30.34 | treke|home | hehe. Typical. |
| 03:30.50 | kergoth | gb2: what do ya mean? make it absolute and ditch the .. references? |
| 03:31.02 | treke|home | kernel, opie, x, or bug tracker. |
| 03:31.06 | gb2 | kergoth: i want to generate a relative symlink |
| 03:31.21 | CosmicPenguin | treke|home: if you can wait just a bit longer, I'll release a tarball |
| 03:31.29 | treke|home | ping me when you do. |
| 03:31.37 | CosmicPenguin | Of course |
| 03:31.52 | kergoth | gb2: just write a shell function to compare the number of '/'s between two paths, and output that many ..'s |
| 03:32.00 | kergoth | s/paths/absolute paths/ |
| 03:32.13 | gb2 | ah |
| 03:32.53 | kergoth | /usr/bin -> /bin = ../../bin/ .. so its difference of / + 1 |
| 03:33.04 | kergoth | s/of \//of \/s/ |
| 03:33.08 | gb2 | well.. i might have //'s in there |
| 03:33.21 | kergoth | well, ditch those. and kill trailing |
| 03:33.52 | gb2 | fog% echo 'foo/bar//baz///quux' | tr -s / |
| 03:33.52 | gb2 | foo/bar/baz/quux |
| 03:33.52 | gb2 | score |
| 03:34.02 | kergoth | hehe |
| 03:35.17 | gb2 | fog% echo 'foo/bar//baz///quux' | tr -s / |sed -e 's,^/,,' -e 's,[^/]*,..,g' |
| 03:35.17 | gb2 | ../../../.. |
| 03:35.19 | gb2 | this seems good |
| 03:37.24 | treke|home | cool. |
| 03:37.35 | treke|home | oreilly lets you download chapters for a fee now |
| 03:37.45 | chouimat | treke|home: which book? |
| 03:37.52 | treke|home | chouimat: Their safari service. |
| 03:38.25 | gb2 | kergoth: that was actually surprisingly simple |
| 03:39.13 | kergoth | heh, yeah |
| 03:40.13 | chouimat | treke|home: how much is the subscription? |
| 03:40.18 | gb2 | are you doing glibc toolchain hacking? |
| 03:40.25 | kergoth | yeah |
| 03:40.30 | kergoth | if the damn thing will ever finish building |
| 03:40.33 | gb2 | what're you making it do? |
| 03:40.41 | kergoth | work :) |
| 03:40.52 | gb2 | i've got a whole bunch of gcc fixes |
| 03:41.15 | kergoth | ah, they upstream yet? |
| 03:41.16 | gb2 | if this build finishes i'll be effectively done with it |
| 03:41.19 | kergoth | cool |
| 03:41.20 | gb2 | i'm testing |
| 03:41.48 | gb2 | i was going to do a test trying to build a glibc and uclibc toolchain together |
| 03:45.02 | gb2 | kergoth: i nuked a few of the random virtual/* packages |
| 03:45.07 | gb2 | since they weren't really virtual |
| 03:45.15 | kergoth | cool |
| 03:45.31 | kergoth | blah-headers vs blah-initial is silly |
| 03:45.34 | kergoth | heheh |
| 03:46.23 | gb2 | i just made it depend on the -initial package directly |
| 03:46.34 | kergoth | good |
| 03:46.37 | gb2 | i left virtual/libc since that is proper |
| 03:46.42 | kergoth | yea |
| 03:46.42 | gb2 | nothing uses it |
| 03:46.45 | gb2 | it's just a provide |
| 03:46.49 | kergoth | good |
| 03:46.55 | kergoth | so oemake doesnt shit itself |
| 03:47.05 | gb2 | yeah, it's a reminder though |
| 03:47.11 | kergoth | yep |
| 03:47.33 | kergoth | i'd like to get the cvs fetching outputting properly named tarballs for archival purposes before i go back to messing with oemake |
| 03:50.00 | chouimat | CosmicPenguin: I saw porn movie that had better quality than this award show on spike |
| 03:53.03 | gb2 | i wonder what creates ${CROSS_DIR}/{lib,usr/lib}/gcc-lib |
| 03:53.18 | gb2 | the proper one is in ${CROSS_DIR}/${TARGET_SYS}/... |
| 03:53.26 | kergoth | yep |
| 03:53.29 | kergoth | good question |
| 03:53.37 | gb2 | there are empty gcc-lib dirs |
| 03:53.58 | CosmicPenguin | chouimat: I didn't say it would be pretty |
| 03:53.58 | gb2 | the one other thing i need to do is make binutils vanish its ldscripts into the target sys one |
| 03:54.02 | gb2 | s/one/dir/ |
| 03:54.08 | *** part/#openembedded ictyl (bradford@user-0ce2ku9.cable.mindspring.com) |
| 03:54.35 | chouimat | CosmicPenguin: I think I will what cartoon network instead |
| 03:54.43 | CosmicPenguin | chouimat: I'm watching scrubs |
| 03:54.51 | kergoth | yea, we dont want any includes or libs outside of the target_sys area.. bad policy, would break having chains for multiple archs in the same prefix |
| 03:54.54 | kergoth | heh |
| 03:55.06 | kergoth | which should be possible |
| 03:55.08 | gb2 | hrm.. does anything even use the ldscripts? |
| 03:55.19 | gb2 | i went to look at redhat's packaging and they don't even include them |
| 03:55.22 | chouimat | CosmicPenguin: currently checking what is on pay tv tomorrow ... 12h15: 2001: Une parodie de l'espace |
| 03:58.10 | chouimat | night |
| 03:58.14 | CosmicPenguin | nite |
| 03:58.32 | gb2 | kergoth: oh god. binutils-cross sucks |
| 03:58.41 | gb2 | D=${CROSS_DIR} |
| 03:58.44 | gb2 | and it uses do_install |
| 03:58.47 | kergoth | heh, how bad is it |
| 03:58.54 | gb2 | instead of a do_stage |
| 03:58.55 | kergoth | ah |
| 03:59.04 | kergoth | wtf, the cross class takes care of all that |
| 03:59.12 | kergoth | wonder if we didnt ahve a cross class when it was done |
| 03:59.13 | kergoth | :) |
| 04:00.48 | gb2 | gar.. my running automake didn't do the right thing |
| 04:01.07 | *** join/#openembedded noda (~noda@modemcable013.119-201-24.mc.videotron.ca) |
| 04:01.08 | noda | Moo! |
| 04:01.12 | kergoth | heh, yeah, just modify both the .am and .in |
| 04:01.15 | kergoth | hey noda |
| 04:02.01 | noda | So, any reason the downloads weren't set up "the oz.org way?" Do you want me to put symlinks in all the right places and stuff? :) |
| 04:02.38 | noda | Oh, and am I correct in assuming that the 2.4.18 kernel works with SD? :) |
| 04:03.15 | kergoth | noda: yeah, please do |
| 04:03.16 | kergoth | yep |
| 04:03.18 | kergoth | its a sharp tree |
| 04:03.23 | noda | Ah. |
| 04:03.31 | kergoth | so their drivers, as shitty as they are, including their buzzer and stuff |
| 04:03.40 | kergoth | its just a holdover to give us time to get it done right |
| 04:03.53 | noda | Okay. Probably much better than 2.4.6 though, right? |
| 04:04.08 | kergoth | yep |
| 04:04.52 | noda | kergoth could you please mkdir /home/groups/o/op/openzaurus/htdocs/official/3.3.5 and chmod g+w it? :) |
| 04:05.21 | noda | 3.3.5 *is* testing, right? |
| 04:05.42 | kergoth | yep |
| 04:06.03 | gb2 | kergoth: globs in FILES_* .. what sort of globs can i use? |
| 04:06.23 | kergoth | noda: done |
| 04:06.27 | noda | kergoth: Thanks. |
| 04:06.54 | noda | btw, anyone who uses epiphany: I made a site-specific popup-allower plugin for epiphany this morning :). |
| 04:06.58 | noda | Anyone want? :) |
| 04:07.05 | kergoth | gb2: anything the python glob.glob will handle, so anything the c lib glob() function can |
| 04:09.26 | *** join/#openembedded chinande (~chinande@adsl-68-78-141-202.dsl.emhril.ameritech.net) |
| 04:11.19 | gb2 | hrm |
| 04:11.27 | noda | Ack, hotplug doesn't always work on the new kernel? |
| 04:11.42 | noda | Evil :) |
| 04:16.04 | noda | Oh, what's the release name for this one? :) |
| 04:16.17 | kergoth | ask mickey, its his release |
| 04:16.26 | noda | Bah, he's not here... |
| 04:16.35 | kergoth | heh, guess it doesnt have a name |
| 04:16.42 | noda | "Out with the old?" -- then the first OE one could be "In with the new" |
| 04:17.01 | treke|home | "I'm not dead yeat" |
| 04:17.07 | noda | Done |
| 04:18.12 | CosmicPenguin | "Thanks for the memories"? |
| 04:18.28 | CosmicPenguin | "Good night, Cleveland!" |
| 04:18.48 | treke|home | "You go to hell and die" |
| 04:19.55 | gb2 | hrm |
| 04:19.59 | gb2 | looks like a nor'easter |
| 04:20.09 | kergoth | hm, these 18 stalled torrent downloads are sucking up all my upstream |
| 04:20.12 | kergoth | gr |
| 04:22.45 | kergoth | gb2: planning on adding ${HOST_PREFIX}gcc adn g++ to the DEPENDS in base.oeclass to get this all to happen transparently? |
| 04:24.33 | gb2 | kergoth: once it's all working, yes. |
| 04:25.02 | kergoth | k |
| 04:25.36 | gb2 | i'm nearly done getting this working. |
| 04:25.43 | gb2 | keep finding small packaging bugs |
| 04:25.47 | kergoth | heh, k |
| 04:27.55 | gb2 | hm. |
| 04:39.42 | noda | Okay, the website has proper links to files now :) |
| 04:43.05 | noda | Back to Advanced Calculus studying :) |
| 04:48.48 | CIA | repository=03packages user=11kergoth@direwolf.ppp.ti.com |
| 04:48.48 | CIA | 07ChangeSet@1.354 Get apache starting on boot properly. |
| 04:49.53 | CIA | repository=03packages user=11kergoth@direwolf.ppp.ti.com |
| 04:49.53 | CIA | 07ChangeSet@1.355 Helps to have an /sbin/init.. |
| 04:51.59 | CIA | repository=03packages user=11kergoth@direwolf.ppp.ti.com |
| 04:51.59 | CIA | 07ChangeSet@1.356 glibc: for cross_dir, dont bother with the asm -> asm-ARCH symlink, just drop it into asm. |
| 04:55.27 | kergoth | heh, this glibc toolchain build didnt like me last week, but is happy this week, and i didnt change anything in that tree |
| 04:55.34 | kergoth | oh well, i wont complain |
| 04:56.54 | gb2 | kergoth: do we use "install" to install stuff? |
| 04:56.59 | gb2 | in scripts |
| 04:57.04 | gb2 | or should i just cp stuff |
| 04:57.12 | kergoth | generally yeah, since its easy to get the perms right |
| 04:57.14 | kergoth | -m 0755 |
| 04:57.15 | kergoth | whatever |
| 04:57.19 | gb2 | yeah |
| 04:57.25 | gb2 | that's what i was about to use it for |
| 04:57.47 | kergoth | relying on the umask just seems like bad policy |
| 04:57.49 | kergoth | heh |
| 05:06.11 | *** join/#openembedded Shdwdrgn (Picard@sourpuss.net) |
| 05:10.41 | kergoth | gb2: looks like bug1 may be willing to contribute to the new version of oe.. he has a metadata handling lib thatd work for one piece of what we need |
| 05:11.52 | *** join/#openembedded treke|laptop (~ggilbert@evrtwa1-ar13-4-62-145-063.elnk.dsl.gtei.net) |
| 05:12.48 | gb2 | cool |
| 05:13.05 | kergoth | indeed.. then we'd just need fetching, and a task/job execution scheduler with dependency handling |
| 05:13.15 | kergoth | then itd be time to adapt the tools |
| 05:14.56 | cyn | kergoth: so we had a little design session (I seriously need to get this dude ON irc, it would motivate both of us) and we're going to implement just the fuzzy search stuff - because really there are tools for all 'the other shit' once your id3s are correct |
| 05:15.30 | cyn | you can make playlists and reorganize directories and rename files etc. in all kinds of programs once the id3 tags are correct, so wtf should we reimplement it for |
| 05:15.34 | kergoth | yeah, makes sense. thats what i was thinking as well |
| 05:15.39 | kergoth | yep, precisely |
| 05:15.42 | cyn | it took some convincing |
| 05:15.48 | kergoth | i'm always using tools like that once i tag my shit right |
| 05:16.12 | cyn | and getting that agreed upon was a big help in getting me further interested with this |
| 05:16.20 | cyn | because before it was just this huge project that I would get like 0 benefit from heh |
| 05:16.21 | kergoth | hehe |
| 05:16.35 | cyn | now it's just a neat small-medium one I'll get 0 benefit from |
| 05:17.06 | kergoth | yeah. i've always found searching and relevency and the like interesting when doing engines in teh past |
| 05:17.14 | kergoth | so i'd be willing just for the novelty of it if nothing else |
| 05:17.28 | cyn | we also decided that once we get some kinda beta working we'll probably make up a little query engine so ppl don't have to download freedbs database to try it and give us feedback |
| 05:17.37 | cyn | yeah, that's pretty much why I'm willing |
| 05:17.41 | cyn | learn python + novelty |
| 05:17.48 | cyn | data structures are just damned sexy |
| 05:18.03 | kergoth | hehe, yea |
| 05:19.03 | *** join/#openembedded stigger (~jeff@p012-n1-gos.nlc.net.au) |
| 05:37.21 | gb2 | so very nearly done |
| 05:38.01 | kergoth | http://www.betondisco.de/php/showtext.php3?id=40&lang=en .. hmm, ya gotta love a song written from the perspective of a fetus. |
| 05:38.05 | kergoth | gb2: nice |
| 05:39.36 | gb2 | i keep having to do whole builds to test that it works properly |
| 05:39.48 | gb2 | god i'm glad i have ccache |
| 05:39.53 | kergoth | hehe, no kidding |
| 05:46.05 | gb2 | i keep ending up with usr/* in my cross_dir |
| 05:52.11 | treke|laptop | man. this kernel shit is scary. |
| 05:52.17 | kergoth | heh |
| 05:52.21 | kergoth | s/scary/fun/ |
| 05:52.25 | treke|laptop | hehe |
| 05:52.38 | treke|laptop | been readin the oreilly book. |
| 05:54.00 | treke|laptop | and forgetting this stuff about memory segmentation as soon as my eye hits the next paragraph |
| 05:54.31 | kergoth | hehe |
| 05:54.46 | CosmicPenguin | bah |
| 05:54.47 | kergoth | half of it you never need if you only deal in drivers anyway |
| 05:54.53 | kergoth | just use it as a reference and write some modules |
| 05:54.57 | CosmicPenguin | only losers use memory segmentation |
| 05:54.58 | treke|laptop | hehe |
| 05:55.06 | kergoth | hehe |
| 05:55.12 | CosmicPenguin | mmus are lame |
| 05:55.13 | CosmicPenguin | :) |
| 05:55.31 | treke|laptop | kergoth: figured I'd at least skim over it |
| 05:56.00 | gb2 | drivers are like hello world compared to stuff like virtual memory |
| 05:58.11 | kergoth | hm, wonder why our e2fsprogs build is statically linking e2fsck |
| 05:58.41 | kergoth | oh |
| 05:58.48 | kergoth | we need --enable-dynamic-e2fsck |
| 05:58.50 | kergoth | hehe |
| 05:58.50 | gb2 | that's probably the usual case for it, need to manually make it shared |
| 05:58.54 | kergoth | yep |
| 05:59.19 | kergoth | shits like 600k static |
| 05:59.29 | kergoth | compiling anything static using glibc is just bad |
| 05:59.31 | kergoth | heh |
| 06:00.29 | treke|laptop | hmmm. the only difference between bzImage and zImage is where in ram the kernel gets loaded? |
| 06:00.51 | gb2 | i think so |
| 06:01.01 | treke|laptop | nifty |
| 06:01.05 | gb2 | bzImage can be larger than zImage because of it |
| 06:01.14 | gb2 | it's not bzip2'd like people commonly think |
| 06:01.20 | treke|laptop | I always assumed it was more complicated |
| 06:04.07 | kergoth | gb2: well, got ldscripts out of the wrong path.. but who knows if itll have other consequences |
| 06:04.12 | kergoth | i added to cross.oeclass: |
| 06:04.15 | gb2 | i rm'd it |
| 06:04.20 | kergoth | libdir=${prefix}/${TARGET_SYS}/lib |
| 06:04.24 | gb2 | in do_install/do_stage |
| 06:04.27 | kergoth | includedir=${prefix}/${TARGET_SYS}/include |
| 06:04.34 | kergoth | sanity checking the build now |
| 06:04.39 | gb2 | cool. |
| 06:05.08 | kergoth | i wish there were more standard granular install targets for autotools/libtool |
| 06:05.24 | kergoth | to get just what we want, you know |
| 06:05.25 | kergoth | heh |
| 06:06.46 | kergoth | gah, ccache really doenst help glibc builds all that much, it still takes a long fucking time |
| 06:07.03 | gb2 | glibc has tons of autogenerated stuff |
| 06:07.08 | gb2 | NOTE: build 200312050045 completed |
| 06:07.09 | gb2 | score |
| 06:07.25 | kergoth | nice |
| 06:08.16 | kergoth | heh, um |
| 06:08.22 | kergoth | changing that path.. made the world explode |
| 06:11.04 | gb2 | toolchain build is 1.1gig |
| 06:11.07 | gb2 | just toolchain |
| 06:11.11 | gb2 | just uclibc toolchain |
| 06:11.20 | gb2 | since it builds it so many times |
| 06:11.33 | kergoth | hah |
| 06:11.48 | gb2 | fog% time rm -rf tmp |
| 06:11.48 | gb2 | rm -rf tmp 0.05s user 3.29s system 20% cpu 16.616 total |
| 06:12.13 | gb2 | one more make to verify.. then i'll check it in |
| 06:13.50 | kergoth | cool |
| 06:14.32 | gb2 | the 3.3.1 gcc stuff is now horribly out of date |
| 06:14.39 | gb2 | should i just nuke it while i'm at it? |
| 06:14.46 | kergoth | yeah, go for it |
| 06:14.52 | gb2 | cool |
| 06:14.57 | kergoth | if somebody wants it back, they can copy it from the 3.3.2 shit |
| 06:15.06 | gb2 | time to learn more bk commands |
| 06:17.17 | gb2 | so i'll "bk new filename" to add stuff, "bk rm filename" and "bk rmdir dirname" then use "bk ci" to check in all my changed files, then bk commit all of the whole lot into a changeset, correct? |
| 06:17.45 | kergoth | yup |
| 06:17.59 | kergoth | use bk sfiles -c|bk diffs -u -; to see the changes |
| 06:18.09 | kergoth | and bk pending or bk sfiles -pC to see deltas waiting to go into a cset |
| 06:18.24 | kergoth | s/changes/changes not yet checked in with ci/ |
| 06:19.41 | kergoth | bk sfiles -x to see files in the repos that havent yet been bk new or bk add'd |
| 06:19.55 | gb2 | i'm reading the bk sfiles manpage now |
| 06:20.09 | kergoth | useful stuff, most of the commands can accept its output |
| 06:20.18 | gb2 | sfiles is an entirely non-desctructive command right? |
| 06:20.22 | kergoth | so e.g. i can only commit some of what i ci'd |
| 06:20.33 | kergoth | bk sfiles -pC|grep glibc| bk commit -y"glibc modifications" - |
| 06:20.36 | kergoth | or what have you |
| 06:20.37 | kergoth | yep |
| 06:20.39 | gb2 | cool |
| 06:21.42 | gb2 | what's -pC do? |
| 06:21.46 | gb2 | oh |
| 06:21.48 | gb2 | found it |
| 06:21.53 | kergoth | just the tip of each set of deltas |
| 06:21.58 | kergoth | iirc a shows all of them |
| 06:22.20 | kergoth | i use C because commit only needs that, a will make it unhappy |
| 06:22.41 | gb2 | ah |
| 06:23.32 | gb2 | i'm very near the end of my build |
| 06:23.36 | kergoth | i've been collecting useful combinations of bk commands in shell functions, should probably put it all somewhere |
| 06:23.51 | gb2 | that'd be cool |
| 06:28.26 | *** join/#openembedded mrwoody (~paolo@pool-141-150-39-41.mad.east.verizon.net) |
| 06:33.14 | gb2 | NOTE: build 200312050112 completed |
| 06:33.15 | gb2 | yay |
| 06:36.20 | gb2 | hrm |
| 06:36.28 | gb2 | what dir does do_install execute in? |
| 06:37.17 | kergoth | iirc ${B} or ${S} |
| 06:37.30 | kergoth | so you can install relative paths from the build area into ${D}, saves typing |
| 06:37.45 | gb2 | i need to pull something out of the build tree and install it. |
| 06:37.50 | gb2 | i guess i can use ${S} :) |
| 06:38.08 | kergoth | you mean the source tree? yeah |
| 06:38.12 | gb2 | yeah |
| 06:38.17 | gb2 | that's what i meant |
| 06:38.31 | kergoth | in fact, i'd bet a number of packages assume that we're in ${S}, when we could be in ${B} |
| 06:38.39 | kergoth | course those packages dont _use_ a build dir, but even so |
| 06:38.39 | kergoth | heh |
| 06:40.35 | gb2 | yeah |
| 06:42.50 | gb2 | hrm. |
| 06:43.03 | gb2 | does the package_tar class refuse to clobber existing tar files? |
| 06:43.08 | kergoth | yes |
| 06:43.12 | gb2 | ok |
| 06:43.15 | kergoth | annoying, but necessary sine we dont have stamps for it |
| 06:43.20 | gb2 | that explains why it did nothing then |
| 06:43.23 | kergoth | yeah |
| 06:43.28 | kergoth | if you have a better way to handle it, let me know |
| 06:43.54 | gb2 | *shrug* yeah |
| 06:43.57 | kergoth | er, we do have a stamp for it, but we shouldnt |
| 06:44.00 | kergoth | heh |
| 06:44.00 | W8TVI | ethereal is so dam cool |
| 06:44.02 | treke|laptop | hmmm. this was probably a bad mood |
| 06:44.06 | gb2 | there's a stamp for package_tar |
| 06:44.13 | gb2 | oh |
| 06:44.16 | gb2 | you just said that |
| 06:44.18 | gb2 | god it's late |
| 06:44.21 | kergoth | gb2: yeah, thats silly. either we should obey the stamp, or obey the files, not both :) |
| 06:44.38 | kergoth | probably the stamp, since the install dir contents can change |
| 06:44.52 | gb2 | yay, finished. |
| 06:44.54 | gb2 | now to play with bk |
| 06:44.59 | kergoth | nice |
| 06:45.14 | treke|laptop | I've been splitting this patch and applying in pieces. Except I forgot to keep track of what has been applied |
| 06:45.24 | kergoth | hah |
| 06:45.40 | kergoth | use diffsplit and patcher |
| 06:46.28 | treke|laptop | patcher should be in debian. |
| 06:46.37 | treke|laptop | since it rocks. |
| 06:46.50 | kergoth | yeah, definately |
| 06:46.55 | kergoth | schurig should get it into distros |
| 06:47.03 | gb2 | hrm |
| 06:47.22 | gb2 | bk sfiles -c doesn't include bk new'd files |
| 06:47.24 | kergoth | treke|laptop: you know what pisses me off about patcher |
| 06:47.38 | kergoth | gb2: -c is unci'd modifications |
| 06:47.53 | gb2 | oh.. bk new was like ci'ing the file |
| 06:47.55 | kergoth | -p shows pending deltas (add/new/ci) |
| 06:47.58 | kergoth | yup |
| 06:48.08 | kergoth | a delta is like a patch. an individual change |
| 06:48.13 | kergoth | add and ci both create them |
| 06:48.33 | treke|laptop | brilliant |
| 06:48.37 | gb2 | is bk rmdir recursive? |
| 06:48.45 | kergoth | gb2: yep |
| 06:48.53 | treke|laptop | split this into 500 patches |
| 06:49.00 | kergoth | it also has mvdir, and cp |
| 06:49.10 | kergoth | bk cp copies the file including all history, every delta to date |
| 06:49.26 | kergoth | which is kinda cool |
| 06:50.41 | gb2 | bk sfiles -c | bk ci -y'foo bar' - |
| 06:50.44 | gb2 | does that do what i think it does? |
| 06:50.56 | kergoth | should |
| 06:51.07 | kergoth | can also bk -r ci -y"stuff" |
| 06:51.12 | kergoth | -r operates on every file in the repos |
| 06:51.16 | kergoth | for files that arent -c, it does nothing useful |
| 06:51.25 | gb2 | that'll work. |
| 06:51.33 | kergoth | yep, i use that on occasion |
| 06:51.48 | kergoth | note that bk -r diffs -u is what bk recommends to ssee your changes as a patch |
| 06:51.57 | kergoth | its like 300 times slower than bk sfiles -c | bk diffs -u - |
| 06:52.47 | gb2 | is there a global dry-run? |
| 06:52.49 | gb2 | -n ? |
| 06:53.10 | gb2 | doesn't look like it |
| 06:53.47 | kergoth | i need to add the "build environment" depends and provides |
| 06:54.00 | kergoth | so if i dont want to build a local chain i can say my environment provides arm-linux-gcc and g++ |
| 06:54.25 | kergoth | or something similar |
| 06:54.47 | CIA | repository=03packages user=11gbritton@doomcom.org |
| 06:54.49 | gb2 | hrm |
| 06:55.22 | kergoth | hehe |
| 06:55.25 | kergoth | i should fix that trigger |
| 06:55.34 | gb2 | figures.. the one that doesn't work makes it here |
| 06:55.38 | gb2 | and the one that does doesn't |
| 06:55.43 | kergoth | yeah, not sure why that is |
| 06:56.23 | gb2 | ok, now time for sleep at nearly 2am |
| 06:57.15 | kergoth | night |
| 06:57.34 | gb2 | i -think- glibc and uclibc toolchains should coexist |
| 06:57.36 | gb2 | and both build ok |
| 06:57.40 | gb2 | but i haven't tested that |
| 06:57.48 | gb2 | i just tested a uclibc only build |
| 06:58.25 | kergoth | cool, so i can probably just use HOST_VENDOR to select between them for the apps, eh? |
| 06:58.44 | gb2 | oh. |
| 06:58.53 | gb2 | my gcc .oe files force TARGET_VENDOR |
| 06:59.08 | gb2 | that's how they ensure that they'll build that way. |
| 06:59.11 | kergoth | good |
| 06:59.53 | gb2 | i haven't tested someone building with something like TARGET_VENDOR=-bogus |
| 07:00.06 | kergoth | .oe overrides .conf data |
| 07:00.10 | kergoth | so shouldnt be an issue |
| 07:00.21 | gb2 | my thought is that dependencies would endup making it unhappy |
| 07:00.25 | kergoth | if i override HOST_VENDOR, i shoudl be able to use either uclibc or glibc for hte apps on command |
| 07:00.28 | kergoth | ill test |
| 07:00.33 | gb2 | ok, cool. |
| 07:00.44 | gb2 | it should work if you set it to empty or "-uclibc" |
| 07:00.50 | kergoth | yea |
| 07:02.23 | gb2 | kergoth: i'm wondering if HOST_* would potentially get set to what the .oe overrides |
| 07:03.06 | gb2 | but i think it should do the right thing. |
| 07:03.09 | kergoth | HOST_PREFIX probably would |
| 07:03.13 | kergoth | the rest would be fine |
| 07:03.23 | kergoth | so i may have to override prefix directly, or vendor and prefix, to keep it happy |
| 07:03.27 | kergoth | i'll play around |
| 07:04.10 | gb2 | i override it after the inherit |
| 07:04.45 | *** join/#openembedded BadCode (~sumwid@dsl-082-082-166-107.arcor-ip.net) |
| 07:04.48 | kergoth | doesnt matter, HOST_PREFIX is always TARGET_PREFIX, which is SYS, which obeys VENDOR |
| 07:05.01 | kergoth | remember, _PREFIX obey one another, not the individual components. |
| 07:05.05 | gb2 | i override target_prefix too |
| 07:05.12 | kergoth | if you fuck with HOST_ components, you have to modify its prefix manually |
| 07:05.28 | gb2 | i only touch target |
| 07:05.33 | kergoth | i know. |
| 07:05.50 | kergoth | if i modify HOST_VENDOR, that wont modify HOST_PREFIX. just saying i need to set both |
| 07:05.51 | gb2 | my thought is.. what I want is HOST to be the target of OE, but this .oe to only work for the target it specifies. |
| 07:06.22 | kergoth | yep, that sounds correct |
| 07:10.01 | gb2 | yeah, i override vendor/sys/prefix |
| 07:10.17 | gb2 | mostly i touch sys and prefix to ensure they get updated |
| 07:11.05 | gb2 | ok, bed for real now |
| 07:27.29 | treke|laptop | hmmmm |
| 07:27.45 | treke|laptop | arch/arm/mach-sa1100 has a whole lot more crap in it than mach-pxa :) |
| 07:39.04 | *** join/#openembedded chinande_ (~chinande@adsl-68-78-132-18.dsl.emhril.ameritech.net) |
| 07:41.59 | CIA | repository=03packages user=11kergoth@direwolf.ppp.ti.com |
| 07:42.00 | CIA | 07ChangeSet@1.359 Correct trigger syntax error |
| 07:42.17 | kergoth | there, pushing without pulling first, shouldnt send anything to CIA anymore |
| 07:43.38 | treke|laptop | hmmm |
| 07:43.47 | treke|laptop | wonder what the difference between husky and husky-j is |
| 07:45.07 | kergoth | japanese.. the bootup logo please wait thing |
| 07:45.46 | treke|laptop | I guess development didnt toss out the idea of a non japanese version :) |
| 07:58.10 | treke|laptop | I'm amazed how much of this patch applies |
| 07:58.26 | treke|laptop | though applying and compiling are two different things :p |
| 08:45.22 | *** join/#openembedded lsmith (~vandal@p213.54.249.107.tisdip.tiscali.de) |
| 08:45.34 | *** join/#openembedded LinuxDoc (~chrissy@pD9E271E6.dip.t-dialin.net) |
| 09:06.59 | *** join/#openembedded mk501 (~Snak@216.112.15.59) |
| 10:06.34 | *** join/#openembedded vandal (~vandal@p213.54.249.107.tisdip.tiscali.de) |
| 11:10.05 | *** join/#openembedded lsmith (~vandal@p213.54.249.107.tisdip.tiscali.de) |
| 11:10.35 | *** join/#openembedded pb_ (~pb@dsl-62-3-66-204.zen.co.uk) |
| 11:13.29 | *** join/#openembedded darkschneider (~gab@213.140.22.64) |
| 11:43.57 | Twiun | hiya darkschneider |
| 11:44.08 | darkschneider | lo Twiun :) |
| 11:44.33 | darkschneider | me working on triton (arm pxa250 board) |
| 11:44.40 | Twiun | sorry to say I had to go with the Sharp laptop in the end - the weight is the main factor |
| 11:45.35 | darkschneider | Twiun: yeah, say me who it is |
| 11:45.41 | darkschneider | wheigth counts a lot |
| 11:45.54 | darkschneider | wheigth/weigth |
| 11:47.12 | Twiun | tried ordering from sharp themselves, but they were out of stock for the foreseeable future!? |
| 11:47.17 | Twiun | ah well |
| 11:47.28 | Twiun | I should receive it around the 20th |
| 11:48.27 | darkschneider | that's not much |
| 11:48.58 | darkschneider | normaly here i have to wait at least 10 days for out of stocks things, but more usualy a month |
| 11:49.45 | Twiun | I ended up ordering from another supplier - the time constraint couldn't be worked around |
| 11:50.35 | darkschneider | eheh |
| 12:06.59 | Twiun | mickey|lunch: give that domain a try again when you're back please, I think dns has finally propagated |
| 12:11.22 | *** join/#openembedded pizzathief (jason@dsl-203-113-221-3.VIC.netspace.net.au) [NETSPLIT VICTIM] |
| 12:11.27 | darkschneider | later all, going to study |
| 12:12.57 | Twiun | see yas |
| 12:50.08 | Twiun | I think it's time I played with another distro... mdk just isn't sturdy enough for upgrading stuff |
| 12:50.45 | Twiun | any suggestions? debian? lfs? (not gentoo, can't afford the time to build it) |
| 12:55.19 | killefiz | Twiun: lfs doesn't build any faster ;) |
| 12:55.37 | Twiun | killefiz: no, it's true - no idea why I mentionned it at all :) |
| 12:55.54 | Twiun | killefiz: this is for a headless machine, pure server/compilation box |
| 12:56.30 | killefiz | debian |
| 12:58.06 | Twiun | version? |
| 12:58.27 | Twiun | I was looking at that jigo stuff on the debian site |
| 12:58.30 | mithro | debian |
| 12:58.54 | mickey|lunch | Twiun: ah sorry i lost the address & login - could you /query this please |
| 12:59.07 | Twiun | mickeyl: sure |
| 12:59.10 | killefiz | hmm - i'd go with unstable if you want current compilers and stuff |
| 13:02.25 | Twiun | 'k, unstable it is |
| 13:04.12 | mickeyl | k, i'm off for some hours... installing Mandrake 9.2 here. hopefully bbl ;) |
| 13:04.22 | Twiun | hrmph. now which sid jigdo file do I want out of here? ftp://ftp.fsn.hu/pub/CDROM-Images/debian-unofficial/sid/jigdo/ or should I just get the iso? |
| 13:04.25 | Twiun | mickeyl: ok |
| 13:06.50 | Twiun | or is that one jigdo per cd?? |
| 13:09.56 | Twiun | ... actually looks like 12 iso anyway, so that sounds about right |
| 13:14.32 | *** join/#openembedded lsmith (~vandal@p213.54.249.107.tisdip.tiscali.de) |
| 13:15.09 | Twiun | hmmm, 4 hours to go |
| 13:50.36 | *** join/#openembedded leffe (~leffe@as5-6-1.h.s.bonet.se) |
| 14:09.08 | *** join/#openembedded andreas__ (~andreas@p508C5878.dip.t-dialin.net) |
| 14:10.18 | andreas__ | hi. which compiler do i need to compile software for simpad/zaurus? |
| 14:10.19 | andreas__ | a normal gcc with --target=arm? |
| 14:16.26 | TheMasterMind1 | you'd need arm-linux-gcc |
| 14:16.30 | TheMasterMind1 | for zaurus anyway |
| 14:16.38 | TheMasterMind1 | ibot oz toolchain |
| 14:16.40 | | it has been said that oz toolchain is http://www.openzaurus.org/official/toolchain/, or mkdir /usr/local/arm and extract them in there. |
| 14:17.59 | andreas__ | can i just build my own compiler & binutiles, or is there something "special" about it? |
| 14:21.40 | *** join/#openembedded pb_ (~pb@dsl-62-3-66-204.zen.co.uk) |
| 14:33.16 | mickeyl | andreas__: depends. if you want to compile for arm you have to do patches or else the produced code will not work |
| 14:33.42 | andreas__ | why that? |
| 14:34.00 | mickeyl | because the official gcc produces buggy code for arm without patches |
| 14:34.26 | mickeyl | so... better use a precompiled toolchain |
| 14:34.27 | andreas__ | and these patches are not yet in the current gcc snapshot? |
| 14:35.00 | mickeyl | no idea - i don't follow compiler snapshots... my software is unstable enough :D |
| 14:35.04 | andreas__ | http://www.openzaurus.org/official/toolchain/README |
| 14:35.22 | andreas__ | gcc 3.3.1 final seems to be compiled without patches |
| 14:36.09 | andreas__ | ok, then i will try the precompiled toolchain... 2.95 or 3.3.1? |
| 14:38.10 | mickeyl | depends |
| 14:38.10 | mickeyl | for what device? |
| 14:38.30 | andreas__ | simpad |
| 14:38.45 | mickeyl | 2.95 if you want to be compatible with the rest of the world, 3.3.2 if you want more speed. |
| 14:40.33 | andreas__ | i never had problems with gcc3.3-compiled software & the old gcc-2.95 debian stuff... |
| 14:41.07 | andreas__ | but i will take 2.95, seems to be the "official" zaurus/simpad toolchain |
| 14:42.32 | andreas__ | btw, is the simpad CPU really called "SA1100"? I can only find the SA-1110 datasheets on intel.com |
| 14:43.45 | andreas__ | ok, found them |
| 14:49.31 | *** join/#openembedded mattaw (~mas01r@arm.ecs.soton.ac.uk) |
| 14:50.31 | mickeyl | oh and btw., i just got confirmation the C860 is a "Shepherd" |
| 14:50.46 | mickeyl | ibot: shepherd |
| 14:50.47 | | methinks shepherd is Sharp Zaurus SL-C750 or a dog |
| 14:50.52 | mickeyl | ibot: forget shepherd |
| 14:50.53 | | i forgot shepherd, mickeyl |
| 14:51.11 | mickeyl | ibot: shepherd is Sharp Zaurus SL-C750, or SL-C860, or a dog. |
| 14:51.12 | | okay, mickeyl |
| 14:51.18 | pb_ | those crazy sharp dudes. |
| 14:51.22 | mickeyl | yeah |
| 14:51.28 | mickeyl | husky would have been more reasonable |
| 14:51.45 | gb2 | my c760 identifies itself as a shepherd in places |
| 14:51.55 | pb_ | your c760 is all mixed up. |
| 14:51.56 | gb2 | the only real difference is the size of the ram/falsh |
| 14:51.58 | gb2 | flash |
| 14:52.08 | gb2 | pb: the sharp firmware says "shepherd" |
| 14:52.16 | mickeyl | firmware? |
| 14:52.19 | gb2 | erm |
| 14:52.19 | mickeyl | what firmware? |
| 14:52.27 | gb2 | i haven't had all my coffee yet |
| 14:52.31 | gb2 | s/firmware/image/ |
| 14:52.40 | gb2 | s/image/kernel/ perhaps |
| 14:53.53 | gb2 | other than ram size and flash size, i don't see anything being different for the c750 and c760 |
| 14:54.45 | mickeyl | well since you can flash what you want on every model, it's more a question of how the hardware identifies itself, e.g. what the diagnostics menu tells |
| 14:56.54 | gb2 | the diagnostics menu says "c760' i think |
| 14:57.15 | gb2 | i don't remember it actually saying a named thing |
| 14:57.52 | gb2 | if i were building platforms i would probably consider the c750 c760 and c860 to be the same board since i'll bet they are |
| 14:58.00 | pb_ | yeah, I think it's ony the kernels that have this dog fixation. |
| 14:58.07 | pb_ | s/ony/only/ |
| 14:58.08 | gb2 | just with differing ram/flash chips populated |
| 14:58.08 | mickeyl | yeah |
| 14:58.34 | mickeyl | which is why calling the c750 and c760 differently was dumb anyway. the kernel can identify the amount of flash, no need to tell it @ compile time. hehe |
| 14:58.34 | mickeyl | s/e/ |
| 14:58.46 | pb_ | right. |
| 15:04.04 | gb2 | yeah |
| 15:04.22 | *** join/#openembedded TimRiker (~timr@proxyle02.ext.ti.com) |
| 15:04.34 | *** join/#openembedded bipolar (~bflong@dhcp-103-630-1.msns.sm.ptd.net) |
| 15:16.43 | chouimat | morning |
| 15:20.12 | *** join/#openembedded BadCode (~sumwid@dsl-082-082-166-107.arcor-ip.net) |
| 15:23.40 | *** join/#openembedded TheMasterMind1 (~aman@h-66-167-239-50.MCLNVA23.dynamic.covad.net) |
| 15:23.40 | *** mode/#openembedded [+o TheMasterMind1] by ChanServ |
| 15:31.43 | *** join/#openembedded DyB|Vandal (~vandal@p213.54.249.107.tisdip.tiscali.de) |
| 15:57.26 | *** join/#openembedded lsmith (~vandal@p213.54.240.220.tisdip.tiscali.de) |
| 15:58.33 | andreas__ | do i need the openzaurus buildroot if i want to compile libSDL myself? |
| 16:04.21 | *** join/#openembedded ciaolinux (~paolo@pool-141-150-141-200.mad.east.verizon.net) |
| 16:17.07 | *** join/#openembedded Disconnect (~dis@arent.you.special.you.gotontheinter.net) |
| 16:17.54 | *** join/#openembedded andersee (~andersee@codepoet.org) |
| 16:21.31 | CosmicPenguin | Whats the current accepted pratice for the X touchscreen calibration? Do you need to restart the server after calibration? |
| 16:23.35 | gb2 | hrm. glibc and uclibc staging clobber eachother. i think that's the only real clobbering bug left in teh two toolchains. |
| 16:26.42 | treke|home | mornin |
| 16:28.13 | gb2 | kergoth's split staging ought to fix that, but i wonder if it would confuse autoconf |
| 16:36.53 | *** join/#openembedded Pendalar (Pendalar@h130.73.40.69.ip.alltel.net) |
| 16:40.47 | *** join/#openembedded andreas__ (~andreas@p508C5878.dip.t-dialin.net) |
| 16:41.01 | *** join/#openembedded mrwoody (~paolo@pool-141-150-41-12.mad.east.verizon.net) |
| 16:50.23 | *** join/#openembedded Ruied_Zaurus (~Ruied_Zau@81.84.147.82) |
| 16:51.06 | *** join/#openembedded Ruied_Zaurus (~Ruied_Zau@81.84.147.82) |
| 16:51.18 | *** join/#openembedded Ruied_Zaurus (~Ruied_Zau@81.84.147.82) |
| 17:26.26 | *** join/#openembedded BadCode (~sumwid@dsl-082-082-166-107.arcor-ip.net) |
| 17:55.01 | *** join/#openembedded _Psycho_ (~Bob@HSE-Montreal-ppp3472052.sympatico.ca) |
| 18:05.27 | chouimat | http://www.despair.com/dreams.html |
| 18:17.04 | kergoth | chouimat: hehe, i like htat one |
| 18:20.09 | gb2 | hey kergoth. |
| 18:20.24 | gb2 | i fired off a dual glibc/uclibc toolchain build |
| 18:20.50 | gb2 | the do_stage of glibc clobbered the uclibc build so subsequent stuff didn't work. but it otherwise looked mostly ok |
| 18:21.08 | kergoth | hehe, figured thatd happen |
| 18:21.09 | kergoth | cool |
| 18:24.32 | *** join/#openembedded tomasito (~tomasito@p508C3E66.dip.t-dialin.net) |
| 18:34.07 | TheMasterMind1 | kergoth: whats new |
| 18:34.30 | chouimat | kergoth: http://www.despair.com/overconfidence.html |
| 18:35.07 | *** join/#openembedded leffe (~leffe@as5-6-1.h.s.bonet.se) |
| 18:49.22 | *** join/#openembedded AndersB (~anders@15.80-202-52.nextgentel.com) |
| 18:51.24 | kergoth | TheMasterMind1: swamped with shit at work, have a deadline wednesday.. bleh |
| 18:53.29 | kergoth | oh, hordes of the underdark released eh.. gonna hve to check that out |
| 18:58.33 | *** join/#openembedded Virusmaster (~Virus@pD9524705.dip.t-dialin.net) |
| 18:59.06 | Disconnect | kergoth: got pushed back a day? (and how bad a deadline is it, if you are on here the whole time? ;) ;) ..) |
| 18:59.09 | treke | hehe. CosmicPenguin announced pixdev 1.1 on freshmeant and forgot the code should be done :) |
| 19:04.00 | kergoth | Disconnect: nah, i just had the date wrong. was wednesday the whole time. :) |
| 19:16.08 | CosmicPenguin | What? It went up to 1.1? |
| 19:16.14 | CosmicPenguin | It was 1.0 yesterday, and that was wrong too |
| 19:17.44 | treke | oh that was nanoxdev that went to 1.1 |
| 19:22.20 | andreas__ | is there a picture viewer for opie? i would like to browse through the digicam pictures on a CF card |
| 19:22.46 | andreas__ | would be nice on the 800x600x16 simpad screen |
| 19:24.33 | *** join/#openembedded mickeyl (mickey@deneb.tm.informatik.uni-frankfurt.de) |
| 19:24.40 | treke | opie-showimage |
| 19:24.41 | treke | tximage |
| 19:26.32 | mickeyl | did you built a kernel with the w100 patch yet? any other useful things from the larger patch? where does it come from anyway? |
| 19:26.41 | mickeyl | treke: |
| 19:27.07 | treke | I've built kernes with the patch |
| 19:27.11 | treke | mickeyl: Some japanese site |
| 19:27.37 | treke | It works. Qte needs to use it now. |
| 19:27.42 | mickeyl | treke: is the default rotation supposed to be landscape or portrait with this patch applied? |
| 19:27.44 | treke | I tried fixing qte, got lost in the code. |
| 19:28.16 | treke | It doesnt change the default rotation |
| 19:28.17 | andreas__ | treke: thanks. i didn't find opie-showimage, but tximage looks good |
| 19:28.59 | andreas__ | unfortunately community.zaurus.com seems to be down |
| 19:32.19 | treke | use the simpad feed |
| 19:32.33 | CosmicPenguin | The instructions on my flu medicine says to take one tablet by mouth |
| 19:33.08 | CosmicPenguin | Which is good, because before I read that, I was investigating which holes would be most suitable |
| 19:33.23 | CosmicPenguin | I'm glad I didn't go with my first choice.... :) |
| 19:34.05 | treke | well it does say by mouth |
| 19:34.15 | treke | so it might mean nose |
| 19:34.49 | andreas__ | treke: it isn't in the simpad feed |
| 19:35.55 | treke | yes |
| 19:46.50 | mickeyl | ibot: ozscripts |
| 19:46.52 | | [ozscripts] function ozclean() { rm -rf $stamps/.$1.*; make SUBDIRS=packages/$1 clean-sources; make SUBDIRS=packages/$1 clean-sourcetrees; }, or function ozmake() { make SUBDIRS=packages/$1; }, or function oznew() { ozclean $1 && ozmake $1; } |
| 19:47.43 | kergoth | gb2: hmm, it may be a good idea to let the cross packages have a do_install task |
| 19:48.03 | kergoth | e.g. if you put CROSS_DIR somewhere external to the build area, and want to deploy the toolchains to others |
| 19:49.07 | *** join/#openembedded pb_ (~pb@cpc2-cmbg4-3-0-cust239.cmbg.cable.ntl.com) |
| 19:52.47 | mickeyl | aah |
| 19:52.50 | Disconnect | mickeyl: btw I wasn't on crack re the battery. it sat at 100% for a long time, finally started going down, seemed reasonable at 50% .. 2 minutes later it went critical.. |
| 19:53.09 | Disconnect | (under the same load that took it from 75% to 50% in 15-20 mins) |
| 19:53.12 | mickeyl | the w100 patch greatly improves fb console speed |
| 19:53.15 | kergoth | hehe, sharp's battery driver is shit |
| 19:53.27 | treke | mickeyl: The w100 patch I was using? |
| 19:53.28 | kergoth | fuck, i'm out of space on / again |
| 19:53.43 | mickeyl | treke: yeah. it contains fb buffering |
| 19:53.58 | Disconnect | kergoth: me too. that root/home split is killing me. got >20M on home that I can't do anything with (package-wise) |
| 19:53.59 | treke | mickeyl: If you want to make it improve even more, change the rotation of the console :) |
| 19:54.17 | mickeyl | treke|home: don't dare to put kernel workload on me. |
| 19:54.18 | treke | mickeyl: one of the ioctls handles it |
| 19:54.31 | treke | mickeyl: hehe. I'm working with mallum on it |
| 19:54.34 | mickeyl | good |
| 19:54.35 | mickeyl | :D |
| 19:54.45 | treke | you can try the qte part if you want :p |
| 19:54.47 | Disconnect | s/20M/50M now that I look at it again |
| 19:54.50 | mickeyl | gonna play with your example script now |
| 19:54.52 | *** join/#openembedded LinuxDoc (~chrissy@pD9E271E6.dip.t-dialin.net) |
| 19:55.00 | kergoth | hm, i have a feeling i'm going to be stuck doing work stuff this weekend |
| 19:55.07 | mickeyl | sucks |
| 19:55.09 | mickeyl | your deadline? |
| 19:55.29 | kergoth | yeah. isnt too much left, really, just takes time |
| 19:56.12 | treke | mickeyl: Probably should toss this in permentantly |
| 19:56.47 | kergoth | ugh, why do we still have some default variable sets in the handlers.. that belongs in oe.conf |
| 19:57.12 | mickeyl | treke: yeah, definitly. it doesn't hurt as long as noone rotates :) |
| 19:57.20 | treke | hehe |
| 19:57.37 | mickeyl | treke: any idea what the rest of the patch is about? |
| 19:57.37 | kergoth | hmm |
| 19:57.49 | treke | mickeyl: Nope. Looks like some ts pressure stuff |
| 19:57.49 | mickeyl | treke: lots of apm stuff and clock stuff |
| 19:58.05 | treke | a big part of it is fonts |
| 19:58.16 | mickeyl | treke: does it apply completely? |
| 19:58.27 | treke | Havent tried applying the whole patch |
| 19:58.31 | treke | probably does. |
| 19:59.02 | mickeyl | what i don't understand is that this stuff looks pretty much like it has been left out in the official release kernel patches to 2.4.18 |
| 19:59.11 | mickeyl | (a part of it though) |
| 19:59.36 | treke | ? |
| 19:59.57 | mickeyl | the w100 diff, the apm and clock diff |
| 20:00.25 | treke | I dont understand what you are saying |
| 20:00.29 | mickeyl | hmm |
| 20:00.34 | mickeyl | ok, rephrase |
| 20:01.33 | mickeyl | a part of this looks like it should have been in the official released patch, i.e. the w100 stuff looks like only someone with intimate knowledge of the w100 chip could be the author |
| 20:02.08 | treke | There was w100 support in the official patch |
| 20:02.48 | treke | it could also be a patch from someone at sharp |
| 20:03.00 | mickeyl | yes i know. but it was not complete. i still have kinds of doubt that the kernel patch is exactly what we have with the kernel on the sharp rom |
| 20:03.19 | treke | Any particular reason? |
| 20:04.09 | mickeyl | not really. a feeling. and the fact that the sharp rom is faster. i don't know if that's only the ATI support in libsl though |
| 20:04.26 | treke | I think it's through libsl and libqte |
| 20:04.58 | mickeyl | if only sharp would start to release qte patches... starting with the damn german keyboard patch for the SL5500G up to the newer things. |
| 20:05.13 | kergoth | sharps sources *are* almost always behind their rom releases by a version or two.. |
| 20:05.18 | kergoth | heh, we dont need that anymore |
| 20:05.24 | treke | Someone said that sharp does have a w100 qte driver |
| 20:05.54 | kergoth | k, pushed an oe cset that adds ${HOME}/.oe to OEPATH and includes site.conf, so you can put a site.conf in ~/.oe/conf to set global things like DL_DIR for all your builds |
| 20:06.31 | mickeyl | kergoth: oh and btw... something i wanted to ask you since long ago. there is exactly one part in the kernel where SL5500_G is evaluated. i wonder if you know what it does and why its there... |
| 20:06.50 | kergoth | whose kernel? |
| 20:07.04 | mickeyl | embedix |
| 20:07.12 | mickeyl | 2.4.6 and 2.4.18 |
| 20:07.24 | mickeyl | a second... i'll get you the source file and linenumber |
| 20:07.25 | kergoth | where is it? in teh source tree |
| 20:07.26 | kergoth | k |
| 20:09.17 | treke | mickeyl: that patch seems to fuck calibration |
| 20:09.22 | treke | for the touchscreen |
| 20:09.34 | treke | wait. nm. It still is doing it |
| 20:11.30 | mickeyl | hmm... d'oh. it's no longer in 2.4.18. it was only in 2.4.6 |
| 20:13.50 | *** join/#openembedded KingRobot (~Mathew@dsl093-185-092.sea1.dsl.speakeasy.net) |
| 20:14.56 | mickeyl | ah no, also in 2.4.18-embedix: File: include/asm-arm/arch-sa1100/time.h. Starts with #ifdef CONFIG_COLLIE_G in line 134. It's only two lines... I'll paste: |
| 20:15.11 | mickeyl | ifdef CONFIG_COLLIE_G |
| 20:15.11 | mickeyl | <PROTECTED> |
| 20:15.11 | mickeyl | <PROTECTED> |
| 20:15.17 | mickeyl | #else |
| 20:15.18 | mickeyl | <PROTECTED> |
| 20:15.18 | mickeyl | <PROTECTED> |
| 20:15.19 | mickeyl | that's all |
| 20:15.34 | mickeyl | any idea what's that about? |
| 20:18.26 | *** join/#openembedded ictyl (~bradford@diploid.med.unc.edu) |
| 20:24.17 | kergoth | mickeyl: heh, looks like the default date/time setting |
| 20:24.21 | kergoth | probably when coming out ouf full reset |
| 20:24.29 | kergoth | dont have the code in front of me, but regardless its not something critical |
| 20:25.18 | mickeyl | okay, thanks. dunno why that has to be differenciated for G and non-G... but hey... that's SHARP ;) |
| 20:31.33 | kergoth | what the.. |
| 20:31.34 | kergoth | hmm |
| 20:32.27 | kergoth | aha, found the bug in the glibc chains |
| 20:35.14 | gb2 | bug? |
| 20:35.17 | gb2 | "the" bug? |
| 20:36.00 | mickeyl | hmm |
| 20:36.09 | kergoth | hehe |
| 20:36.09 | kergoth | well, a big one anywya :) |
| 20:36.29 | gb2 | which bug? |
| 20:37.02 | kergoth | gb2: glibc stage wasnt installing bits/* |
| 20:37.10 | kergoth | those come from install-others, not install-headers |
| 20:37.10 | mickeyl | hmm |
| 20:37.16 | kergoth | which makes no sense to me, but hey |
| 20:39.53 | gb2 | ah |
| 20:39.58 | gb2 | weird |
| 20:40.16 | kergoth | apparently its because they're generated |
| 20:40.19 | kergoth | oh well |
| 20:40.42 | gb2 | what do you want to do about the header staging clobbering? |
| 20:40.56 | kergoth | heh, dunno, maybe its time to implement splitting staging |
| 20:41.10 | gb2 | i'm wondering if that will piss off autotools tests |
| 20:41.32 | gb2 | i suppose it should obey the passed in CPPFLAGS |
| 20:41.34 | kergoth | nah, but we'll have to change a bunch of --with-blah-libs= and -includes= |
| 20:41.50 | kergoth | it should, but some macros look for the .so directly, instead of testing a link |
| 20:41.54 | kergoth | which pisses me off, but i digress |
| 20:42.06 | gb2 | heh |
| 20:42.18 | kergoth | anywya, most of those are passed in the paths |
| 20:42.23 | kergoth | so we'd have to change all those occurances |
| 20:42.28 | kergoth | beyond that, it should be fairly clean |
| 20:42.49 | kergoth | the thing is, we need our multiple provides handling code first.. guess that should be done now. |
| 20:51.15 | *** join/#openembedded lsmith (~vandal@p213.54.240.220.tisdip.tiscali.de) |
| 21:12.29 | Disconnect | kergoth: any interest in a draft release procedure? |
| 21:12.51 | Disconnect | (I know you said you were gonna do one, but it'd prolly go faster if we put one together and then submitted it for changes/acceptance ;) ..) |
| 21:13.02 | kergoth | indeed, feel free |
| 21:13.12 | kergoth | we need a number of policies and procedures documented |
| 21:13.26 | kergoth | CM plan, OZ package naming conventions, deployment/release, blah blah |
| 21:16.49 | chouimat | CosmicPenguin: nice apology |
| 21:19.19 | *** part/#openembedded ictyl (~bradford@diploid.med.unc.edu) |
| 21:23.29 | *** join/#openembedded chouimat (~chouimat@r2351064.cidc.net) |
| 21:33.19 | *** join/#openembedded chouimat (~chouimat@r2351064.cidc.net) |
| 21:33.19 | *** mode/#openembedded [+o chouimat] by ChanServ |
| 21:36.38 | *** join/#openembedded W8TVI (~W8TVI@userd90.torchlake.com) |
| 21:48.51 | kergoth | brb, coffee run |
| 21:54.10 | chouimat | <PROTECTED> |
| 21:54.39 | gb2 | chouimat: you pasted that at the same time someone else did on another network |
| 21:54.44 | gb2 | heh |
| 21:54.49 | chouimat | hehe |
| 21:55.24 | treke | kergoth: permissions on /dev/null are wrong :) |
| 22:08.38 | *** join/#openembedded stigger (~jeff@p012-n1-gos.nlc.net.au) |
| 22:14.33 | *** part/#openembedded KingRobot (~Mathew@dsl093-185-092.sea1.dsl.speakeasy.net) |
| 22:18.21 | gb2 | cool, blizzard warning for some places |
| 22:18.46 | *** join/#openembedded drw|bbl (~drw@c-24-0-141-17.client.comcast.net) |
| 22:23.16 | chouimat | gb2: currently had some fun with a survey about some products ... can I have you salary please? yes 100 000+ |
| 22:23.38 | CosmicPenguin | What chouimat didn't say is thats in canadian dollars... :) |
| 22:24.06 | chouimat | CosmicPenguin: i know the average salary here is 32000cad |
| 22:24.46 | CosmicPenguin | chouimat: among engineers, or among everybody? |
| 22:24.52 | chouimat | CosmicPenguin: everybody |
| 22:25.19 | chouimat | CosmicPenguin: I can hope of doing max 42000cad after 7 years |
| 22:25.31 | gb2 | that sounds horribly low |
| 22:25.35 | gb2 | for an engineer |
| 22:25.42 | CosmicPenguin | Thats about right, I think... isn't the US average salary somewhere between 16 and 20? |
| 22:25.52 | _Psycho | engineer are usually 50k+ here |
| 22:25.54 | CosmicPenguin | But the engineer average is much higher |
| 22:26.07 | chouimat | _Psycho: quebec city it's 38 000 |
| 22:26.16 | _Psycho | chouimat, im at 43k and i just have a DEC |
| 22:26.31 | _Psycho | You get 50k+ easy as consultant |
| 22:26.31 | chouimat | CosmicPenguin: the Quebec average is 25K |
| 22:27.04 | gb2 | psycho: cad? |
| 22:27.23 | _Psycho | yes cad |
| 22:27.47 | chouimat | _Psycho: I have a BScA es Informatique de Genie and the max they offered me in Quebec City was 32000k ... I do a little less by myself but I'm tired of fighting to get contract ... I will probably do 23K this year ... |
| 22:27.48 | _Psycho | i have 2 interview coming |
| 22:27.51 | gb2 | which is like $38k US? |
| 22:27.53 | _Psycho | i might get a lots more soon i hope |
| 22:27.57 | _Psycho | yeah something like that gb2 |
| 22:28.04 | chouimat | ibot change 50000cad to usd |
| 22:28.12 | chouimat | ibot change 50000 cad to usd |
| 22:28.21 | gb2 | heh |
| 22:28.22 | chouimat | gb2: yup |
| 22:28.23 | gb2 | pretty close |
| 22:28.44 | _Psycho | Im waiting on a consultant contrat, would be 68k |
| 22:28.54 | _Psycho | else i might get an admin job at like 50k+ depend |
| 22:29.02 | _Psycho | else well i stay at my job =p |
| 22:29.04 | chouimat | _Psycho: for the same 50k in Ontario you will have 7000cad more in your pocket |
| 22:29.13 | gb2 | still strikes me as low, but i live in an expensive city to live in |
| 22:29.21 | gb2 | i'm in boston |
| 22:29.49 | _Psycho | yeah well indeed, the cost of "life", is lower a bit here |
| 22:30.00 | _Psycho | well no idea montreal vs boston thought |
| 22:30.13 | chouimat | gb2: in quebec and montreal you can have a 4 1/2 for 700cad (and it's expensive here) |
| 22:30.13 | gb2 | boston is one of the most expensive in the US |
| 22:30.21 | _Psycho | oh hehe |
| 22:30.35 | treke | hmm. |
| 22:30.37 | gb2 | i'm paying 1600USD/mo for a 700sqft apartment with my girlfriend |
| 22:30.53 | _Psycho | I see, im paying 425 a month, probably for the same |
| 22:30.56 | treke | gb2: My rent seems much more reasonable suddenly :) |
| 22:31.07 | CosmicPenguin | Holyl moley |
| 22:31.10 | gb2 | psycho: 425 canadian even? |
| 22:31.20 | treke | sugar what is 425CAD in USD |
| 22:31.22 | Sugar | Can't get US$ from CAD. |
| 22:31.23 | CosmicPenguin | My loan on a two bedroom 900 is 700 buck |
| 22:31.35 | treke | CosmicPenguin: lemme guess. Live in the midwest? |
| 22:31.38 | chouimat | ibot change 425 cad to usd |
| 22:31.45 | CosmicPenguin | treke: Beautiful Salt Lake City Utah |
| 22:31.49 | treke | CosmicPenguin: Figures. |
| 22:31.53 | CosmicPenguin | Land of a thousand wifes |
| 22:32.11 | chouimat | CosmicPenguin: that the problem they are all taken .... |
| 22:32.12 | CosmicPenguin | And I live in the city too... out in the boonies, its much cheaper |
| 22:32.16 | gb2 | i wouldn't balk at spending $325 on an impulse buy |
| 22:32.21 | treke | Am I the only person who thinks more than one wife would be more trouble? :p |
| 22:32.29 | gb2 | it'd be stupid still ofcourse |
| 22:32.40 | chouimat | gb2 I pay 200 cad for a huge house |
| 22:33.04 | CosmicPenguin | Anyone know right off the bat what uid_t is a typedef of? |
| 22:33.29 | chouimat | CosmicPenguin: int iirc |
| 22:33.39 | mickeyl | treke: any idea why kernel-ipaq tries to build with gcc3 ? |
| 22:33.48 | mickeyl | treke: the other kernels automatically use gcc2 if I enable both |
| 22:33.56 | mickeyl | (this is why the kernel doesn't build, btw.) |
| 22:34.00 | CosmicPenguin | chouimat: sure its an int and not an unsigned int? |
| 22:34.00 | treke | mickeyl: It should |
| 22:34.15 | treke | gcc3 is the preferred compiler for it iirc. |
| 22:34.22 | chouimat | CosmicPenguin: not sure |
| 22:34.39 | treke | pb_: Or should we be using gcc2 for the kernel? |
| 22:34.55 | mickeyl | treke: hmm. it chokes in h3600_backpaq_algo.c:17: "pasting DO_YUV_BPIXEL_REVERSE" and "c" does not give a valid preprocessing token |
| 22:35.01 | mickeyl | and ~ 1000 of similar messages |
| 22:35.21 | mickeyl | s/c/( |
| 22:35.29 | CosmicPenguin | chouimat: it is indeed unsigned |
| 22:35.48 | chouimat | CosmicPenguin: i was not sure |
| 22:36.21 | chouimat | anyone know how to lower the dpi on X? |
| 22:43.30 | treke | mickeyl: ping |
| 22:43.41 | mickeyl | pong |
| 22:44.06 | treke | mickeyl: Thinking we might want to include the patch and switch the normal rotation used until we can make qte use the w100 for all of it's rotations |
| 22:45.05 | mickeyl | treke: yes. i added it to buildroot. will commit later, want to test the big patch tomorrow. |
| 22:45.32 | treke | mickeyl: did you add the whole patch? |
| 22:46.03 | mickeyl | treke: no. atm only w100. but i want to test the whole thing - seems to include interesting patches |
| 22:46.18 | treke | yeah. |
| 22:46.36 | treke | I pulled most of the stuff out so I didnt have to worry about bugs in the other stuff |
| 22:46.41 | kergoth | treke: if the tree has the necessary patches applied, itll work fine with gcc3 |
| 22:47.01 | kergoth | (iirc the patches are still in the patch system on www.arm.linux.org.uk, as well as 2.6) |
| 22:47.02 | treke | kergoth: I'm pretty sure it does. It's 2.4.19, and I could have sworn I'd used gcc3 before |
| 22:47.15 | mickeyl | treke: ah no. kernel-ipaq uses 2.4.18 |
| 22:47.23 | mickeyl | at least that's what you did in the makefile |
| 22:47.25 | treke | mickeyl: erm. It shouldnt :) |
| 22:47.30 | kergoth | treke: there is a patch for gcc 3.3 specifically. |
| 22:47.40 | treke | ah |
| 22:47.46 | treke | I never used 3.3 |
| 22:47.50 | treke | I'd been using 3.2 |
| 22:47.56 | treke | mickeyl: That isnt right |
| 22:48.04 | treke | 2.4.18-hh19 is definitly way older |
| 22:48.16 | mickeyl | treke: well... does it work? |
| 22:48.44 | treke | "maybe" :p |
| 22:48.55 | mickeyl | treke: *sigh* k, which is the recommended kernel to build for h3600 atm? |
| 22:49.00 | treke | not sure how that happened. My builds used 2.4.19-4 |
| 22:49.12 | mickeyl | hmm... uncommited local changes, heh? |
| 22:49.18 | treke | Could be |
| 22:50.08 | treke | uhm. I'm looking at bkbits, and I pull directly from cvs |
| 22:50.20 | treke | That is 2.4.19, not 2.4.18 :p |
| 22:50.44 | mickeyl | no |
| 22:50.49 | mickeyl | look at cvsTag |
| 22:50.50 | treke | someone, /me looks at jamey, tagged cvs wrong |
| 22:51.03 | mickeyl | K2-4-18-rmk6-pxa1-hh19 is 2.4.19 ? |
| 22:51.07 | treke | yup |
| 22:51.11 | mickeyl | hmm |
| 22:51.17 | mickeyl | okay, then 2.4.19 doesn't work with 3.3.2 |
| 22:51.20 | kergoth | did they ever stop using -rmk-pxa in their version string? |
| 22:51.42 | treke | bump cvstag to hh30 |
| 22:51.46 | treke | see if that helps |
| 22:51.50 | mickeyl | k, checking... |
| 22:51.50 | treke | might as well use a recent kernel |
| 22:51.54 | mickeyl | sure |
| 22:53.33 | treke | kergoth: nope they still do |
| 22:53.48 | kergoth | treke: thats silly. |
| 22:54.00 | kergoth | just becuase you include a billion patches doesnt mean they need to be in your uname |
| 22:54.04 | kergoth | heh |
| 22:54.06 | kergoth | anyway |
| 22:54.07 | treke | EXTRAVERSION = -test10-rmk1-hh0 |
| 22:54.39 | pb_ | treke: you need gcc3 for the ipaqpxa kernels, and I think it's OK for ipaqsa as well. |
| 22:54.54 | pb_ | at one time gcc3 didn't work properly with the h3600_micro driver, but I think that's fixed now. |
| 22:55.10 | treke | ah. Well we're building an ancient build. mickeyl is going to try something more recent |
| 22:55.40 | *** join/#openembedded chouimat (~chouimat@r2351064.cidc.net) |
| 22:56.40 | pb_ | kergoth: the 2.4 hh.org kernels do all include -rmk-pxa in EXTRAVERSION. we're trying to kick that habit in 2.6. |
| 22:56.54 | kergoth | cool |
| 22:57.02 | treke | apparently not succesfully :p |
| 22:57.10 | kergoth | hehe |
| 22:57.16 | pb_ | though I seem to remember that Russell King used to get upset at people removing the "rmk". oh well. |
| 22:57.45 | mickeyl | hmmm |
| 22:57.46 | mickeyl | no such tag |
| 22:57.51 | pb_ | the "pxa" one is the main problem, because it makes strongarm users freak out. |
| 22:58.07 | *** join/#openembedded chouimat (~chouimat@r2351064.cidc.net) |
| 22:58.07 | kergoth | treke: when you have something like an hh.o tree, which has a *LOT* of patches applied, its no longer sane to even try to keep track of what came from where |
| 22:58.17 | kergoth | particularly not in a version string |
| 22:58.17 | kergoth | heh |
| 22:58.30 | pb_ | yeah, very true. |
| 22:58.51 | treke | mickeyl: you could try 2.6 :p |
| 22:59.06 | pb_ | not really |
| 22:59.10 | Disconnect | kergoth: what about adding a /proc/patches entry? |
| 22:59.13 | mickeyl | no, i want to have something working, not playing around |
| 22:59.40 | pb_ | 2.6 is close to usable on 3900. not really usable on 5000 or 3600. probably doesn't compile on 3800. |
| 22:59.51 | treke | K2-4-19-rmk6-pxa1-hh30 |
| 22:59.54 | mickeyl | treke: i want to have something i can point to people when they're acting like familiar is the only distribution for ipaq out there :D |
| 22:59.57 | treke | pb_: It been a major chore porting? |
| 23:00.30 | kergoth | Disconnect: well, regardless of how you show it, its difficult to even keep track.. particularly with active development happening on a bk or cvs tree.. |
| 23:00.43 | pb_ | treke: not a huge chore. it's been more time-consuming than you might expect because we are rewriting almost all the ipaq drivers at the same time. |
| 23:00.53 | kergoth | Disconnect: i tried teaching patcher's patchset management to integrate with bk, keeping track of what deltas in the repo belong to which patch.. |
| 23:00.55 | treke | ah |
| 23:01.20 | treke | pb_: so similar to kergoths 2.4.22 work |
| 23:01.27 | pb_ | yup, I guess so. |
| 23:01.44 | *** join/#openembedded ljp_laptop (~ljp@218.185.68.28) |
| 23:01.44 | pb_ | the old h3600_hal has gone away completely, replaced by jamey's new soc-device infrastructure. |
| 23:02.06 | treke | ljp_laptop: yo. found any volunteers yet for the rotation code? :p |
| 23:02.22 | ljp_laptop | havent tried.. still on vaca |
| 23:02.39 | treke | ljp_laptop: I poked about for a couple hours bfore giving in |
| 23:02.43 | ljp_laptop | sitting here in the mariott, looking at the beach |
| 23:02.46 | mickeyl | treke: btw.... one way or the other you should commit the rotation code to odevice. i'd like to have the auto rotation... |
| 23:02.53 | treke | yeah yeah. |
| 23:02.56 | Disconnect | kergoth: yah. but that at least gives a baseline for bug reports and such |
| 23:03.06 | treke | mickeyl: I've been on a coding streak all week, so this might happen |
| 23:03.09 | ljp_laptop | man I wish I was 20 again |
| 23:03.15 | mickeyl | treke: k, fine. |
| 23:03.59 | treke | I'll get the backend in at a minimum this week |
| 23:04.14 | kergoth | Disconnect: thats true. but then again, you can look at an annotated version of the source in question to see what checkin/changeset corresponded to it, and go from there as well |
| 23:04.17 | Disconnect | treke: interested in a 7x0 brightness patch? |
| 23:04.31 | treke | sure. why not. |
| 23:04.34 | ljp_laptop | brightness? |
| 23:04.53 | Disconnect | kergoth: easier for a user to just include /proc/patches and .config with bug reports ;) |
| 23:05.26 | Disconnect | ljp_laptop: at least 750/760, the hardware goes up a decent amount higher than OE does. (0x0111 in hardware, vs 0x0105 in L&P) |
| 23:05.41 | *** join/#openembedded chouimat (~chouimat@r2351064.cidc.net) |
| 23:05.44 | ljp_laptop | brighter? |
| 23:05.55 | kergoth | well, my point is, knowing what patches exist in the tree isnt much more useful than knowing what version they're on, if you have cvs or bk, you already know the state of the world |
| 23:05.59 | Disconnect | ljp_laptop: and its got actual granularity, instead of just the 4-step plan. but I dunno how much that'll affect. |
| 23:06.27 | Disconnect | kergoth: it leads into a nice setup where you can drop patches into packages/kernel/patches/ or whatever and have them automatically applied by OE.. |
| 23:06.29 | mickeyl | i can do that in odevice... just an interim solution of course, because we will have to rewrite the whole Zaurus class once 2.4.2x-cl is finished |
| 23:06.42 | mickeyl | (brightness, that is) |
| 23:06.58 | kergoth | Disconnect: eek, encouraging users to use non stock kernels is just asking for trouble |
| 23:07.01 | kergoth | heh |
| 23:07.05 | Disconnect | ljp_laptop: yah. L&P goes to a nice cloudy-day viewing level. hardware goes to "readable in direct blazing sunlight with a welding mask" ;) |
| 23:07.14 | mickeyl | hehe |
| 23:07.22 | ljp_laptop | hmm |
| 23:07.30 | Disconnect | well not quite as bright as all that. but it does go to direct-sunlight mode, whch L&P doesn't. |
| 23:07.59 | kergoth | oops, forgot to kill the default libpthread.so in staging.. |
| 23:08.06 | Disconnect | anyway. I'm headed out to the firestation to do some volunteering, bbl. |
| 23:08.35 | Disconnect | mickeyl: lemme know if you put a patch in before I get to it (or just put it in the commit log; I'll check there before spending effort on it) |
| 23:08.36 | chouimat | hmmm still not 75x75dpi but 87x82 is better than 126x126 imho |
| 23:09.13 | mickeyl | Disconnect: i take care of it. i will make odevice use /proc/driver/corgi_bl for c7x0 instead of the ioctl interface. |
| 23:09.21 | kergoth | hmm, wonder if our opie/qtopia versions of qte should be seperate .oe's |
| 23:09.27 | kergoth | qte, qte-opie, qte-qtopia |
| 23:09.34 | mickeyl | kergoth: YES! |
| 23:09.35 | *** join/#openembedded Pendalar (Pendalar@h130.73.40.69.ip.alltel.net) |
| 23:09.36 | kergoth | hehe |
| 23:09.36 | mickeyl | ;) |
| 23:09.43 | kergoth | well then, i'm going to add qte-opie and get opie in. |
| 23:09.56 | chouimat | kergoth: i think it will make more senses |
| 23:10.23 | kergoth | sense. senses sounds like it'll smell or see more |
| 23:10.24 | kergoth | :P |
| 23:10.31 | kergoth | yeah, it will be cleanest that way |
| 23:10.41 | Disconnect | mickeyl: excellent :) lemme know when it needs testing. |
| 23:10.55 | treke | kergoth: It makes sense. create a base qte that builds with not patches. Then a qte-opie |
| 23:11.01 | treke | ljp will probably just use plain qte |
| 23:11.04 | kergoth | yea |
| 23:11.06 | chouimat | mickeyl: what you're finaly porting opie to qt3 |
| 23:11.11 | kergoth | just i nclude the base adn add more shit to SRC_URI |
| 23:11.12 | treke | instead of -qtopia |
| 23:11.28 | mickeyl | chouimat: no plans at the moment. seems we agreed to wait for the first beta versions of qt4 |
| 23:11.40 | treke | kergoth: And remember, you can check opie patches out of CVS on their own, you don't need to hack it to download all of opie :p |
| 23:11.47 | chouimat | mickeyl: i hope it will be out soon |
| 23:11.56 | kergoth | treke: yeah |
| 23:11.58 | mickeyl | chouimat: me, too. |
| 23:12.02 | kergoth | cvs:// is spiffy |
| 23:12.31 | mickeyl | yes, btw., we need to write down a definitive guide what goes into opie cvs and what goes into the distributors rcs |
| 23:12.38 | mickeyl | atm. it's a mess |
| 23:12.41 | mickeyl | horribly |
| 23:12.42 | kergoth | agreed |
| 23:12.51 | kergoth | guess the CM plan should go into that |
| 23:12.54 | kergoth | what goes in and what doesnt |
| 23:12.57 | mickeyl | yeah |
| 23:13.00 | kergoth | when you check in, when you dont, when you clone |
| 23:13.06 | kergoth | now, to find time to write it.. |
| 23:13.07 | kergoth | heh |
| 23:13.25 | treke | :p |
| 23:13.37 | kergoth | heheh |
| 23:13.45 | mickeyl | aw... cvs.hh.org is slow as molasses |
| 23:13.53 | kergoth | i have coding/projects/user/oe/, in there is gw, oe-dev, oe-whateverelseimworkingon |
| 23:14.07 | kergoth | then coding/projects/user/oe-metadata/, with a number of empty build directories and cloned packages repositories |
| 23:14.08 | treke | I also need to hack up apm |
| 23:15.43 | pb_ | ah, what're you doing with apm? |
| 23:16.38 | treke | pb_: Hacking up suspend/resume scripts so that opie can pretend like it is handled asynchronously :) |
| 23:16.46 | mickeyl | hehe |
| 23:16.53 | pb_ | heh |
| 23:17.11 | kergoth | heh, you mean synchronously? opie expects it to block on the suspend |
| 23:17.17 | treke | yeah |
| 23:17.22 | kergoth | you should fix the apm_bios bug instead |
| 23:17.25 | treke | as it is, opie has a sleep statement |
| 23:17.28 | kergoth | so we dont need the hack |
| 23:17.37 | treke | that would also be nice :) |
| 23:17.41 | mickeyl | yeah... better fix than workaround |
| 23:17.46 | pb_ | why does opie need it to block on the suspend? |
| 23:17.58 | kergoth | pb_: if you suspend via apm_bios while apmd also has the device open |
| 23:18.03 | treke | pb_: It lets you execute code prior to and after resume |
| 23:18.06 | kergoth | pb_: itll never wake up from suspend |
| 23:18.18 | pb_ | treke: the right way to do that is to listen for events on /dev/apm_bios. |
| 23:18.19 | kergoth | pb_: so we cheat and send a signal to apmd, letting *it* suspend, which doenst see the bug |
| 23:18.32 | treke | pb_: In our case we need to check the status of the screens rotation after resuming |
| 23:18.48 | pb_ | kergoth: aha. |
| 23:19.03 | treke | as it is, that code is being executed before the suspend actually occurs :p |
| 23:19.05 | kergoth | i'd rather just fix the fucking apm implementation to fix it :) |
| 23:19.10 | kergoth | stuipd bug |
| 23:19.15 | kergoth | its in all embedix Z kernels |
| 23:19.26 | treke | wonder if that is in this patch |
| 23:19.37 | treke | mickeyl: test :p |
| 23:20.13 | mickeyl | failed. |
| 23:20.17 | mickeyl | ? |
| 23:20.19 | mickeyl | ah ok |
| 23:20.22 | mickeyl | yeah will do |
| 23:20.27 | mickeyl | _tomorrow_ |
| 23:20.28 | mickeyl | ;) |
| 23:20.30 | treke | :p |
| 23:20.31 | pb_ | amusingly, in GPE we had the opposite problem. it turned out that with synchronous suspend, you could get a deadlock condition during wakeup, so I had to change the kernel to make it asynchronous. |
| 23:20.35 | treke | maybe I'll try it tonight. |
| 23:20.48 | kergoth | hehe |
| 23:21.01 | kergoth | thats funny. guess opie shouldnt assume its synchronous, whether we fix the apm bug or not |
| 23:21.18 | treke | the 15 second sleep() in opies code is fun when you remove usb-modules package and it takes 3 seconds to suspend |
| 23:21.28 | kergoth | hah |
| 23:21.30 | mickeyl | :) |
| 23:21.44 | kergoth | damnit, i hate it when a source tree doesnt have a LICENSE or COPYING file, and i have to sit and peruse the fucking code |
| 23:24.10 | *** join/#openembedded chouimat (~chouimat@r2351064.cidc.net) |
| 23:25.18 | *** mode/#openembedded [+o chouimat] by ChanServ |
| 23:32.11 | treke | only good news is each article in the paper gets closer to scraping the bottom of the barrel |
| 23:33.03 | kergoth | wonder if we should distinguish between BSD w/ advert and w/o in the LICENSE field of the .oe files |
| 23:33.13 | treke | probably |
| 23:33.32 | kergoth | heh, what to call it.. i prefer we have one word, since some packages have multiple |
| 23:33.38 | treke | BSDAD |
| 23:33.38 | treke | ? |
| 23:33.40 | kergoth | e.g. apmd's libapm is LGPL, but apmd is GPL |
| 23:33.46 | treke | FREEBSD |
| 23:33.47 | kergoth | so i have LICENSE=LGPL GPL |
| 23:33.50 | kergoth | hmm |
| 23:34.10 | *** join/#openembedded bipolar (~bflong@ctt187223.ceinetworks.com) |
| 23:45.54 | *** join/#openembedded icefox (~ben@ool-18b87f53.dyn.optonline.net) |
| 23:49.00 | *** part/#openembedded AndersB (~anders@15.80-202-52.nextgentel.com) |
| 23:49.01 | treke | mickeyl: Its sad. I dont feel nearly the excitement as everyone else flashing 3.3.5, since its nothing new :( |
| 23:49.44 | mickeyl | treke: right. the excitement on PXA is nowhere near the excitement on SA, where we've gone to new grounds |
| 23:50.04 | treke | it's "cool. there is a bigger number than last week" |
| 23:50.24 | mickeyl | hehe, better than nothing. |
| 23:50.41 | kergoth | heh, 3.3.5 wasnt supposed to amaze and entertain |
| 23:50.46 | kergoth | it was supposed to hold people over |
| 23:50.54 | Twiun | mickeyl: did you check the site out? |
| 23:51.17 | mickeyl | Twiun: yeah, when we talked earlier that day |
| 23:51.40 | mickeyl | didn't read my comments? |
| 23:51.42 | Twiun | mickeyl: I meant today :) see what gallery looks like integrated |
| 23:51.47 | mickeyl | oh |
| 23:52.14 | Twiun | phpwiki will take a few minutes more, gotta setup the db |
| 23:52.55 | mickeyl | nice :) |
| 23:53.03 | mickeyl | esp. the screeny ;) |
| 23:53.15 | kergoth | find tmp/base/ -name COPYING -o -name LICENSE|for i in `cat`; do if (grep -q 'GNU General Public License' $i); then j=`echo $i|cut -d/ -f3|cut -d- -f1`; ls $j*/$j*.oe; fi; done|grep -E -v '\-cross|\-native'|sort -u|for k in `cat`; do mv $k $k.safe; echo 'LICENSE = GPL' > $k; echo >> $k; cat $k.safe >> $k; rm -f $k.safe; done |
| 23:53.16 | Twiun | first thing I had to hand :) |
| 23:53.20 | kergoth | heh |
| 23:53.22 | kergoth | i love scripting |
| 23:53.24 | Twiun | My Eyes! |
| 23:53.34 | mickeyl | is that binary or perl? |
| 23:53.35 | mickeyl | hehe |
| 23:53.48 | kergoth | just checks if its gpl, and if so adds the LICENSE as such |
| 23:53.50 | kergoth | heh |
| 23:53.51 | kergoth | :) |