| 13:00:18 | pattieja | is back (gone 35:02:06) |
| 13:14:55 | nikos | Morning bworth |
| 13:17:23 | bworth | morning nikos, everyone |
| 13:21:03 | cworth | bworth: morning. This weekend I taught xstroke to ignore scrollbars, menus, and buttons for recognition! |
| 13:21:25 | bworth | that's really cool! |
| 13:21:30 | cworth | bworth: it works great with xterm and gtk apps. Little bug with rxvt scrollbars that I am chasing down now. |
| 13:43:35 | nikos | Morning sjohnson and cworth |
| 13:43:47 | sjohnson | Hi nikos |
| 13:58:22 | _Axu_ | hi and bye folks :9 |
| 15:20:33 | beeblebrox | is away: Away! |
| 16:29:40 | sjohnson | is away: lunch |
| 16:41:06 | nikos | is away: errand |
| 16:48:22 | pattieja | is away: lunch |
| 17:17:44 | nikos | is back |
| 17:22:43 | nikos | Morning a7r |
| 17:22:56 | a7r | hey nikos.. |
| 17:23:56 | Disconnect | pokes a7r. |
| 17:24:13 | Kero | looks at a7r and wonders... |
| 17:24:50 | Disconnect | su - a7r -c "cat ipaqstat > feeds" |
| 17:24:53 | a7r | <- busy tired man |
| 17:25:15 | nikos | a7r: Not a good state for Monday morning. |
| 17:25:58 | Disconnect | ..which is great proof of the need for a better method ;) (not that it would help this, since all new packages need intervention. but if its properly automated intervention can be taken to a minimum yes/no..) |
| 17:26:37 | a7r | Disconnect: yeah, this is kind of a half case for automation. ;) |
| 17:26:44 | a7r | believe me, I'd like it to be easier. |
| 17:27:07 | Kero | ``maintainer'' |
| 17:27:23 | Kero | Make kicking the maintainer by users a policy for familiar :) |
| 17:27:35 | Kero | (iff the package disbehaves) |
| 17:27:39 | Disconnect | a7r: you were gonna mail me and take apart the idea (as posted) so we could find somehting that worked.. |
| 17:27:52 | a7r | Disconnect: it's still on my todo list. ;) |
| 17:28:47 | Disconnect | Kero: I have horrible visions of 500 packages all going "new, incompatible {libc/libgtk/whatever}!! Update at once!" .. even worse if the new [libxxxx] is mandated by huge [security hole | bug | hardware damage causer] or whatever, making it urgent that the upgrade happen. |
| 17:29:31 | Disconnect | htinks a7r should he having that nightmare too, assuming he's got time to sleep. ;) |
| 17:29:51 | Kero | Disconnect: debian has that issue solved, apparently; how? |
| 17:30:42 | Disconnect | Kero: with a huge set of packages/upgrade methods/scripts/nightmares. it works great, for debian (which we are similar to, in many ways) but its not only way big for this, it requires way more care&feeding than the manual method does for a dist this size. |
| 17:31:47 | Disconnect | if we had the 9000-odd packages debian did, with 8500-odd maintainers (or whatever huge number it is) on a bunch of dists/etc, it would work great. for 100 packages, 30 maintainers, all we need is something that lets a7r (and whomever else becomes a feed maintainer/ass't maintainer/etc) intervene when necessary instead of all the time .. |
| 17:32:20 | pattieja | is back (gone 00:43:58) |
| 17:32:28 | Disconnect | a7r: is that a fair summation? ;) |
| 17:33:01 | Kero | Disconnect: You didn't answer my question. ``scripts'' means nothing :) |
| 17:34:09 | Kero | Plus, dependencies are in ipkg information, so failed dependencies and packages to be upgraded when a basic package is are already available. |
| 17:34:17 | Disconnect | Kero: AFAIK its all scripts. scripts means plaintext files that operate under an interpreter. and the submission/update proggies are bash scripts, the bug-filing proggies are bash scripts, dunno whats on the backend other than mail-parsing scripts. |
| 17:34:25 | Disconnect | Kero: right. but those upgraded packages have to get there somewhere. |
| 17:34:50 | a7r | Disconnect: I'm too busy to read, I'll let you know in 5. |
| 17:34:57 | Disconnect | Kero: there is no way to look at libgtk.ipk and find out what packages require rebuilding if you change it (trust me, we looked when we first messed with libgtksquash) |
| 17:34:59 | Disconnect | a7r: ok |
| 17:35:45 | Kero | You can't, unless the dependencies are strict. Versioning should be proper, too. |
| 17:36:21 | Kero | like, >=1.2 often means 1.3 should work, but 2.0 need not. |
| 17:36:31 | Disconnect | Kero: you can look at gtk-bubble-blower.ipk and tell that it depends on libgtk-version-whatever, which is great for installing packages (99% of the time, thats what ipkg/feeds/etc will be doing..) but if libgtk has to be changed in some non-binary-compatible form (such that all packages depending on it need rebuilding) you get the various maintainers sending packages to a7r. which is great. except he's gonna twitch, smother, and die under the |
| 17:36:31 | Disconnect | load. we're working on fixing that. |
| 17:37:51 | Kero | Cool; how? |
| 17:37:55 | Disconnect | put out ipaqstat-1.3 on thursday morning(ish) as an ipk. Its just now getting to feeds (well, later today even) .. if something Very Very Bad had been found over the weekend, necessitating a full rebuild of most/all of the dist, we wouldn't hit the next ALPHA for -weeks-, and we'd have quite a long time with only a couple of packages. |
| 17:38:10 | Disconnect | thats the discussion. |
| 17:38:46 | sjohnson | is back (gone 01:09:06) |
| 17:39:11 | Disconnect | check the list archives (subject 'package maintenance proposal') for thursday. probably new version out later today. |
| 17:40:57 | Kero | which list? |
| 17:41:10 | Kero | ah, familiar :) |
| 17:41:33 | Disconnect | yep |
| 17:44:44 | a7r | Disconnect/Kero: Your packages are in unstable, test them out.. shout if there are any problems. |
| 17:45:00 | cworth | a7r: not much of a telecon --- no jg and no a7r so it just sort of folded |
| 17:45:36 | Kero | a7r: thanks. |
| 17:45:43 | | Kero: sorry... |
| 17:45:43 | Kero | ibot, familiar unstable feed |
| 17:45:46 | | Kero: sorry... |
| 17:45:46 | Kero | ibot, familiar unstable |
| 17:46:45 | a7r | cworth: Yeah, I slept through it. :( |
| 17:49:35 | Kero | Disconnect: current/new-stable/unstable/critical... too many versions? |
| 17:50:02 | Kero | Plus it doesn't solve the problem of a libc upgrade (as far as I can see). |
| 17:50:49 | Kero | How does this relate to debian Policies? What are the responsibilities of a maintainer? |
| 17:53:04 | cworth | as far as feed/releases layout, I see only one big problem with the current scheme: there's no way for a user to know if upgrading from release 0.x to stable will render their iPAQ unusable. |
| 17:53:08 | pattieja | would this be a good place to discuss the upcoming release of Familiar 0.5? |
| 17:53:16 | a7r | as good as any. |
| 17:53:19 | pattieja | heh |
| 17:53:30 | pattieja | We're attempting to use the 0.5pre release |
| 17:53:34 | pattieja | and having some problems |
| 17:53:38 | a7r | nod. |
| 17:53:57 | pattieja | the first is that ipkg seems to have a dependency on the package it is installing |
| 17:54:14 | cworth | so, the name "stable" should likely change although that feed still has a useful purpose. |
| 17:54:14 | pattieja | like if you attempt to install a package, it will have a dependency of itself |
| 17:54:38 | Kero | cworth: it's called stable; I don't know it for debian either. |
| 17:54:59 | pattieja | another thing is, we were trying to get X to install, which had dependencies on xstroke and xlibs. xlibs in turn had a dependency on xstroke. |
| 17:55:14 | pattieja | when we attempted to install xstroke, it failed with a 404 error, not found. |
| 17:55:24 | a7r | pattieja: when did this happen? |
| 17:55:29 | pattieja | just now |
| 17:55:34 | a7r | update your packages file. |
| 17:55:40 | cworth | Also, it would also be nice to have a simple way to point to the latest point-release upgrade from a major release. Maybe just a symlink named releases/0.4.latest |
| 17:55:42 | a7r | I just added a new xstroke to the feed. |
| 17:55:51 | pattieja | as in the last 2 minutes? |
| 17:55:54 | a7r | yes. |
| 17:55:57 | pattieja | coolness |
| 17:56:39 | pattieja | still not working |
| 17:56:45 | pattieja | what's the deal with fstroke? |
| 17:56:53 | a7r | fstroke is stale. |
| 17:56:58 | pattieja | ok |
| 17:57:03 | a7r | pattieja: What is the URL it's trying to fetch? |
| 17:57:59 | pattieja | Downloading http://familiar.handhelds.org/familiar/feeds/unstable/packages/armv4l/./xstroke_0.5-6_arm.ipk ... |
| 17:58:36 | a7r | pattieja: yeah, run ipkg update again.. |
| 17:58:43 | pattieja | again? |
| 17:59:09 | Kero | cworth: well, for libc upgrades, you would switch an entire version; one version can only evolve so far. |
| 17:59:12 | pattieja | now I'm getting a 'getcwd() failed: No such file or directory' |
| 17:59:20 | a7r | pattieja: cd /tmp ; ipkg update |
| 17:59:21 | pattieja | when trying to run 'ikpg update' |
| 17:59:24 | pattieja | ok |
| 17:59:55 | Disconnect | was away, back now.. |
| 17:59:56 | Kero | cworth: What's the point in differentiating between 0.4.1 and 0.4.2? Why not just 0.4 that evolves? |
| 18:00:09 | pattieja | just a little spam a7r: |
| 18:00:12 | pattieja | # ipkg install xstroke |
| 18:00:12 | pattieja | The following package were previously requested but have not been installed: |
| 18:00:12 | pattieja | xstroke |
| 18:00:12 | pattieja | Install them now [Y/n] ? |
| 18:00:12 | pattieja | Downloading http://familiar.handhelds.org/familiar/feeds/unstable/packages/armv4l/./xstroke_0.5-6_arm.ipk ... |
| 18:00:14 | pattieja | --00:05:23-- http://familiar.handhelds.org:80/familiar/feeds/unstable/packages/armv4l/xstroke_0.5-6_arm.ipk |
| 18:00:16 | pattieja | => `//tmp/ipkg/xstroke_0.5-6_arm.ipk' |
| 18:00:18 | pattieja | Connecting to familiar.handhelds.org:80... connected! |
| 18:00:20 | pattieja | HTTP request sent, awaiting response... 404 Not Found |
| 18:00:22 | pattieja | 00:05:23 ERROR 404: Not Found. |
| 18:00:24 | pattieja | |
| 18:00:26 | pattieja | ipkg_download: ERROR: Failed to retrieve http://familiar.handhelds.org/familiar/feeds/unstable/packages/armv4l/./xstroke_0.5-6_arm.ipk, returning |
| 18:00:29 | pattieja | ipkg_get_install: Perhaps you need to run `ipkg update'? |
| 18:00:31 | pattieja | |
| 18:00:33 | pattieja | Downloading http://familiar.handhelds.org/familiar/feeds/unstable/packages/armv4l/./xstroke_0.5-6_arm.ipk ... |
| 18:00:34 | a7r | pattieja: yeah.. that's because the package in the feed is now 0.5-9 |
| 18:00:36 | pattieja | --00:05:23-- http://familiar.handhelds.org:80/familiar/feeds/unstable/packages/armv4l/xstroke_0.5-6_arm.ipk |
| 18:00:38 | pattieja | => `//tmp/ipkg/xstroke_0.5-6_arm.ipk' |
| 18:00:40 | pattieja | Connecting to familiar.handhelds.org:80... connected! |
| 18:00:42 | pattieja | HTTP request sent, awaiting response... 404 Not Found |
| 18:00:43 | nikos | pattieja: ipkg update |
| 18:00:44 | pattieja | 00:05:23 ERROR 404: Not Found. |
| 18:00:46 | pattieja | |
| 18:00:47 | cworth | Kero: right. but 0.4-latest would freeze at some point, (which it should). Then users would have to move to 0.5-latest manually. |
| 18:00:48 | pattieja | ipkg_download: ERROR: Failed to retrieve http://familiar.handhelds.org/familiar/feeds/unstable/packages/armv4l/./xstroke_0.5-6_arm.ipk, returning |
| 18:00:51 | pattieja | ipkg_get_install: Perhaps you need to run `ipkg update'? |
| 18:00:53 | pattieja | # |
| 18:00:57 | pattieja | ok |
| 18:00:59 | pattieja | I've done ipkg update about 3 times now |
| 18:01:15 | pattieja | nikos: a7r: do I need to change some configuration somewhere? |
| 18:01:18 | Disconnect | Kero: the many-many feeds (4 or so?) was basically as so: stable -> immutable. changes in bulk. unstable - changes to stable (0.4) .. new-unstable: 0.5 .. critical: 99.99% of the maintainers will never use it. means something is broken that breaks the stable dist and made it through QA anyway, needs to be upgraded immediately. |
| 18:01:18 | a7r | it might be caching at the webserver.. |
| 18:01:23 | pattieja | to make it look at the new repository? |
| 18:01:30 | a7r | no.. |
| 18:01:34 | Disconnect | lemme try that again ;) |
| 18:01:34 | pattieja | we are running squid on our firewall |
| 18:01:46 | a7r | pattieja: oh, that might be it.. |
| 18:01:49 | pattieja | hmm |
| 18:01:52 | pattieja | how to get around that? |
| 18:01:53 | Kero | It need not freeze; if stable is a symlink that switches version, no manual intervention is needed. |
| 18:01:59 | Disconnect | stable -> immutable. 0.4.0 .. will change all at once to 0.4.1. unstable - 0.4.1pre. 'drifts' changes. |
| 18:02:24 | Kero | stable != immutable in English; why try to pretend :) |
| 18:02:57 | a7r | stable means ``it's going to work''. |
| 18:03:35 | Disconnect | Kero: its basically right out of debian.. 0.4.0 doesn't change, but stable does occasionally change, from one revision to the next. if we both build ipaqs with 0.4.0 as stable, but at different times, we should have the same packages. |
| 18:03:55 | cworth | Kero: but in cases like the current 0.4 -> 0.5 transition, it is impossible to do a smooth upgrade. So forcing manual intervention is a Good Thing, (much better than an attempted upgraded which totally hoses the system). |
| 18:04:13 | Kero | 0.5 is not stable yet -> no auto-upgrade. |
| 18:04:19 | Disconnect | whereas if we both build from unstable, at diff't times we get diff't package versions. (with unstable as 0.4.1pre, rather than the current 0.5pre) |
| 18:04:55 | cworth | Kero: even when 0.5 is declared stable, we won't have a 0.4 -> 0.5 migration path without a reflash though. |
| 18:05:23 | Disconnect | Kero: as far as maintainer responsibilities, thats something that has to be worked out AFAIK. |
| 18:05:30 | Kero | The debian security fixes are not fixed for a stable release. |
| 18:05:47 | Kero | cworth: what does familiar know about my partitions? |
| 18:07:14 | Disconnect | debian security is a good example of 'critical' (although security on the ipaq isn't necessarily critical, its a good example of something that may have to change immediately, w/o waiting for the next release) |
| 18:08:29 | Disconnect | (and submissions to the 2 unstable feeds can be automatic by any validated maintainer.. nobody touches stable, and a7r deals with critical and deciding when to make a new stable, as well as creating/changing maintainers) |
| 18:08:33 | Kero | It may be practical to check out a version w.r.t. a date or tag, instead of a fixed-number version. Like CVS offers. |
| 18:09:36 | Disconnect | Kero: the 'pool' concept (instead of feeds) wherein multiple versions sit around and you grab the package info based on release/feed might work better. |
| 18:09:39 | pattieja | why when installing xstroke it installs dillo? |
| 18:09:52 | Disconnect | pattieja: packaging bug - its for the help. |
| 18:10:00 | pattieja | ok? |
| 18:10:16 | cworth | pattieja: I'm gonna remove that dependency. it's for xstroke-help. |
| 18:10:22 | Disconnect | it should recommend it (or be a separate xstroke-help package) but only hard dependencies are working right now - recommends/suggests/etc don't work |
| 18:10:32 | pattieja | it also wanted to install libgtk, lipng2, and xlib6g (exact spellings) but couldn't find them |
| 18:11:03 | cworth | pattieja: I should have used "Recommends:" but we don't have anything to deal with that yet, (pikpak and friends is where that should be happening). |
| 18:11:18 | pattieja | hmm |
| 18:11:27 | pattieja | do these packages need to be installed? |
| 18:11:51 | pattieja | and does xstroke work inside the Xvnc server? |
| 18:11:56 | cworth | Disconnect: yep. But ipkg need not even change for Suggests/Recommends, (although Provides is another matter -- ipkg needs some work for that). |
| 18:13:16 | cworth | pattieja: the other stuff must be Depends in dillo. They are not in xstroke. |
| 18:13:24 | pattieja | ok |
| 18:13:30 | pattieja | xstroke in Xvnc? |
| 18:14:52 | cworth | pattieja: never tried xstroke in Xvnc. I doubt it provides the Render extension, so current xstroke won't know how to draw anything. |
| 18:15:08 | Disconnect | whats a good icewm theme for ipaq? |
| 18:15:09 | pattieja | aww :( |
| 18:15:25 | pattieja | that would've been the ultimate |
| 18:15:48 | pattieja | does it need the RENDER extension? |
| 18:20:09 | mandrake | pattieja: yes. |
| 18:20:36 | pattieja | does it need RENDER just to draw on the screen or for the mouse capture functionality as well? |
| 18:20:38 | cworth | pattieja: fixing xstroke to be backwards-compatible with non-Render servers woulb be quite easy for any C hacker. Source is in CVS. |
| 18:20:55 | pattieja | ^sheepish^ |
| 18:21:15 | mandrake | render is just for display |
| 18:21:18 | cworth | pattieja: just for the translucent drawing. Old xstrokes didn't need Render at all, so the code is even all there already. |
| 18:22:27 | pattieja | which older version of xstroke would be the best to use in that scenario? |
| 18:27:17 | cworth | pattieja: I think 0.4. IIRC, 0.5 was the first to require RENDER. But, what you should really do is just grab the drawing code out of brush.c and move it into the latest brush.c. There are many improvements in recognition with recent versions. |
| 18:27:51 | Disconnect | can't get icewm running on familiar (xstroke won't dock either, so I don't feel too bad about it.) |
| 18:27:59 | Disconnect | got it running, but no docking and no xstroke. |
| 18:38:03 | pattieja | what's the difference among dir, vdir, and ls? |
| 18:38:15 | pattieja | (concerning Familiar) |
| 19:10:01 | dneuer | hi |
| 19:11:10 | pattieja | how big is the ipkg-build command? |
| 19:12:31 | mpakula | hi all |
| 19:14:19 | pattieja | nevermind |
| 19:15:37 | mpakula | guys is familiar suporting 100 base PCMCIA cards or it is not dependend on distro |
| 19:22:16 | nikos | mpakula: Not dependent on distro, just need supporting driver. |
| 19:22:33 | nikos | mpakula: There are at least some 100baseT Ne2K cards that work supposedly. |
| 19:24:00 | mpakula | nikos: thanks |
| 19:33:51 | pattieja | we just created an ipkg of fileutils using BusyBox, and reduced the fileutils footprint from 600KB to 150KB. |
| 19:34:00 | pattieja | will be sending it shortly |
| 19:34:21 | pattieja | unfortunately, this is the first ipkg, not sure everything is 100% |
| 19:36:25 | cworth | pattieja: fantastic. Have you done lots of testing? busybox sed/tar have been known to break ipkg. |
| 19:41:43 | pattieja | no, we're doing that now |
| 19:41:59 | pattieja | it's kind of funny when ipkg no longer works when you remove fileutils :) |
| 19:42:21 | pattieja | but we were able to do a 'busybox ln -s busybox ln' |
| 19:42:25 | pattieja | whew! |
| 19:42:27 | pattieja | etc. |
| 19:46:38 | dneuer | hi a7r |
| 19:46:45 | pattieja | hello a7r |
| 19:46:45 | a7r | hey.. |
| 19:46:47 | cworth | pattieja: OK. Also, your package should use update-alternatives, (with a low priority), to create all the busybox links. That way, the "real" package could also use update-alternatives with a higher priority and users could easily switch back and forth. |
| 19:47:00 | cworth | a7r: greetings. |
| 19:47:24 | pattieja | cworth: what would that mean? if you installed busybox, it wouldn't overwrite previously installed versions of the utilities? |
| 19:48:11 | cworth | pattieja: uhm, yes. but when the other utilities were removed, then busybox would automatically reclaim the links, (preventing the problem you've already hit). |
| 19:49:03 | pattieja | ok |
| 19:49:04 | cworth | pattieja: familiar already has packages that use the update-alternatives support to mange /bin/sh, /bin/vi, and /usr/bin/x-window-manager for exmaple. See any of the packages which provide one of those for an example. |
| 19:50:36 | cworth | pattieja: therefore, when you submit busybox, please also submit an update fileutils, etc. For example, the new fieutils would have a binary named ln-complete or some such and use update-alternatives to create /bin/ln. Same for all other affected binaries. |
| 19:51:06 | cworth | pattieja: BTW, I really appreciate you looking at this. I've been meaning to try it for quite some time, (especially so after I added the update-alternatives support). |
| 19:52:30 | pattieja | it'll be interesting, cause there's alot of overlap that busybox provides |
| 19:52:52 | pattieja | just installed task-x for 0.5pre |
| 19:52:57 | pattieja | how does one do anything? |
| 19:54:19 | Disconnect | pattieja: ipkg install menu-tiny gtkmenu |
| 19:54:27 | pattieja | thanks Disconnect |
| 19:54:49 | Disconnect | np |
| 19:55:53 | pattieja | which package provides 'ps'? |
| 19:56:04 | cworth | pattieja: procps |
| 19:56:48 | pattieja | cworth: wasn't installed, and yet the /etc/init.d/x script needs it to work |
| 19:57:03 | pattieja | guess, task-x or whatever needs dependencies updated? |
| 19:57:25 | cworth | pattieja: yes, this will be a bit of a challenge. But it probably makes sense to do as much as possible in a single busybox package this way. Then, most users won't touch it. Some will install -complete versions of a couple of packages and will still win on space. A few might install -complete versions of many things and decide to drop busybox altogether. |
| 19:57:51 | nikos | acutally, /etc/init.d/x needs to be fixed to use pidof rather than ps. |
| 19:57:55 | cworth | pattieja: sounds like it. please file a bug in Familiar's GNATs. |
| 19:58:36 | cworth | pattieja: using update-alternatives will allow all the flexibility I just described above, (and all with simple ipkg install/remove commands, no manual symlinkage). |
| 19:59:03 | pattieja | where's GNATs? |
| 19:59:20 | | i guess gnats is at http://familiar.handhelds.org/cgi-bin/wwwgnats.pl |
| 19:59:20 | nikos | ibot: gnats |
| 19:59:21 | pattieja | do you have documentation (wiki?) of the update-alternatives configuration? |
| 19:59:57 | pattieja | nikos: cworth: one bug or two? |
| 20:00:05 | nikos | Aren't all the busybox entries symlinks to a single binary? |
| 20:00:14 | pattieja | yes |
| 20:00:28 | cworth | nikos: yes, but update-alternatives still makes a lot of sense here. |
| 20:01:06 | nikos | link to a link to a binary? OK. |
| 20:01:10 | pattieja | which severity level for the bug? I guess critical since it stops it from working? |
| 20:01:19 | pattieja | nikos: yep |
| 20:01:27 | pattieja | that does take more inode space doesn't it? |
| 20:01:35 | nikos | pattieja: It doesn't stop working actually. |
| 20:01:43 | nikos | pattieja: I would just call it Serious. |
| 20:01:44 | cworth | pattieja: I've never written anything up. If you have a Debian system around, then consult docs. there, (minus the slave support which I've never written). Otherwise, just look in postinst/prerm for the ion package for example. |
| 20:01:55 | pattieja | really? seems like it bombed to me |
| 20:02:18 | nikos | pattieja: It's used for stop, not start. |
| 20:02:26 | pattieja | ahh. yep |
| 20:02:30 | cworth | nikos: huh? no extra links due to update-alternatives. |
| 20:02:34 | pattieja | so it's only half critical |
| 20:02:52 | cworth | nikos: I did use one fewer layer of indirection than in Debian's implementation. |
| 20:03:41 | nikos | cworth: The busybox derived packages will have no files. I probably missed part of the discussion. |
| 20:04:26 | nikos | cworth: wait, /bin/ln -> /etc/alternative/ln-busybox -> /bin/busybox |
| 20:04:40 | nikos | cworth: That's a minimum necessary setup isn't it? |
| 20:05:36 | cworth | nikos: no, that's Debian's. Mine will do /bin/ln -> /bin/busybox. |
| 20:06:22 | nikos | cworth: Hmm. Somehow that seems dangerous. But maybe it's just my imagination. |
| 20:07:02 | nikos | cworth: I guess my problem with it is that relinking it completely removes the information about the alternative. |
| 20:07:03 | cworth | nikos: my implementation breaks the rule that "All user-visible configuration belongs in /etc". But I don't think users will manually muck with alternatives too much anyway, (more likely just remove the offending package and save space at the same time). |
| 20:07:25 | cworth | nikos: it's cached in a file somewhere. |
| 20:07:25 | pattieja | bug reported |
| 20:07:32 | cworth | pattieja: thanks. |
| 20:09:07 | derrell | hi, all. i have a pcmcia card that requires no special drivers on winblows, but i get no response on the console when plugged in to the ipaq. what tools are available to poll pcmcia to see what's on the bus? |
| 20:09:42 | derrell | (linux, of course!) |
| 20:10:12 | nikos | derrell: cardctl ident |
| 20:10:51 | derrell | nikos: thanks. and if it says "no product info available" where do I go from there? :-( |
| 20:11:17 | nikos | derrell: pull it out and try reinserting it. Are you saying it worked with Windows CE on this ipaq? |
| 20:11:34 | nikos | derrell: Have you used any card on Linux with this sleeve? |
| 20:12:28 | derrell | orinoco card is working great in this dual sleeve, and works from either socket. other cards work in socket 1 (orinoco is in socket 0 right now). no i don't know if it works on ce on the ipaq. |
| 20:13:14 | jtm | this is JT working with pattieja: I want to run my original thoughts by you regarding busybox. |
| 20:13:16 | derrell | it doesn't come with custom drivers for winblows; it's just supposed to be detected and work. (this will be REALLY cool if i can get it working; it's a dual usb host card!) |
| 20:13:59 | jtm | I was going to create a busybox exe with all the functionality that I wanted, then create individual ipkgs for the sets of programs |
| 20:14:26 | jtm | fileutils, vi, textutils, grep, gzip, tar ... |
| 20:14:54 | nikos | derrell: Be sure it is not a CardBUS card, these will not work with the ipaq. |
| 20:15:13 | jtm | the busybox exe could be changed out by the user if they wanted to add or delete functionality w/o affecting the others too much |
| 20:16:05 | jtm | unfortunately, i dont quite understand the alternative implentations you were talking about, where can i find more info? |
| 20:16:30 | derrell | nikos: oh, crap. i think it is. what stops cardbus cards from working (i.e. how do they differ from standard pcmcia)? |
| 20:17:58 | nikos | derrell: CardBUS is PCI in a card, PCMCIA is ISA in a card. Very different beasts. CardBUS cannot be made to work. |
| 20:19:04 | cworth | jtm: well, the advantage of using update-alternatives is twofold: 1) all the symlinks are handled by the system, not the user. and 2) the full-functionality of the current busybox package is exposed and available to the user, and 3) users can override busybox functions by simply installing "full" versions of the desired program. |
| 20:19:39 | cworth | jtm: users still may want to make a custom busybox, (and the custom one should also be an ipkg doing its own update-alternatives stuff). |
| 20:20:09 | derrell | nikos: bummer! ok, time to return this product. thanks. |
| 20:20:14 | nikos | derrell: Sure |
| 20:20:14 | jtm | ok, where do i find more info on setting it up? |
| 20:21:05 | cworth | jtm: honestly, the best thing would be to rip apart a recent bash, vim, or ion package, (just use tar xzf), and look in control.tar.gz for the postinst/prerm scripts. |
| 20:21:26 | cworth | jtm: or you could dig up Debian docs, (ie. "man update-alternatives" on any Debian machine). |
| 20:22:53 | jtm | cworth: i grabbed the vi package from unstable on hh.org. vim doesnt seem to exist there and it looks pretty plain. I dont think it is what i'm looking for. NO pre and post scripts int the file 'control' |
| 20:23:19 | cworth | jtm: they should be files next to control. |
| 20:23:34 | jtm | oh yeah...checking... |
| 20:24:38 | cworth | jtm: unfortunately, I've got to run now. Please send email to cworth@handhelds.org if you need more help. |
| 20:24:41 | jtm | cworth: nothing. the package i am using is iv_0.0_arm.ipk |
| 20:24:57 | cworth | jtm: bad package. look at bash or ion. |
| 20:25:05 | jtm | cworthAWAY: ok |
| 20:25:47 | cworthAWAY | jtm: basic idea is you request a symlink for an absolute path using a numeric priority. Highest priority wins. |
| 20:28:01 | jtm | cworthAWAY: thanx, i'll catch you later this week to work on it some more. ion and bash have no files other than 'control' |
| 20:30:15 | nikos | jtm: What URL are you getting files from? |
| 20:30:45 | jtm | ftp://ftp.handhelds.org/pub/linux/dists/familiar/releases/latest/packages/armv4l |
| 20:30:56 | nikos | jtm: That is a frozen release. |
| 20:31:03 | nikos | jtm: look at the packages in the unstable feed |
| 20:31:07 | | i heard unstable feed was at http://familiar.handhelds.org/familiar/feeds/unstable/packages/armv4l/ |
| 20:31:07 | nikos | ibot: unstable feed |
| 20:31:38 | nikos | jtm: latest is still v0.4 and REAL old. That is pre-update alternatives. |
| 20:31:51 | jtm | gotcha....thankyou |
| 20:36:41 | dneuer | a7r: it looks like once -hh7 comes out, the only difference between the root fs images for mono & color should be in /boot/params and the eraseblock size, and the units can use the same kernel |
| 20:37:14 | a7r | dneuer: kick ass. |
| 20:37:28 | nikos | dneuer: That's excellent. |
| 20:39:37 | dneuer | a7r, nikos: don't thank me ;-) |
| 20:40:37 | nikos | dneuer: Then who should be laud? The usual list? |
| 20:40:42 | nikos | s/be/we/ |
| 20:44:22 | dneuer | nikos: andy c and jamey, I think |
| 20:45:29 | nikos | dneuer: Cool. |
| 20:52:41 | dneuer | a7r: there's a task-bootstrap.tar.gz for your latest .5pre root fs, I presume? |
| 20:53:28 | volkam | dneuer: are you working on putting .5 on the mono |
| 20:54:09 | dneuer | volkam: sort of. Most of the work has already been done by others; all that's left is to label & package things up correctly |
| 20:54:26 | a7r | dneuer: yup, right next to the jffs2.l |
| 20:57:01 | dneuer | a7r: I may build one for mono to experiment and in anticipation of the "official" one |
| 20:57:44 | a7r | cool. |
| 20:59:59 | volkam | dneuer: put it up tonight! |
| 21:00:40 | dneuer | volkam: will try |
| 21:00:53 | dneuer_DINNER | is away: busy |
| 21:01:06 | volkam | ok |
| 21:40:25 | DragonEagle | is away: I need to go lie down for a bit |
| 21:44:15 | idle | cf-card 64 plugged in, how to create a filesys, now? |
| 21:45:55 | idle | mkfs -t ext2 /dev/flashX does not work, hints? |
| 21:49:32 | volkam | heh |
| 21:49:34 | volkam | HEH |
| 21:49:39 | volkam | I would HOPE not |
| 21:49:44 | volkam | its |
| 21:49:50 | volkam | mkfs -t ext2 /dev/hda1 |
| 21:59:50 | idle | thanks |
| 22:06:44 | idle | works ) |