| 03:57:42 | flypiper | anyone awake? |
| 03:58:40 | flypiper | gota 3700 that I would like to install gpe on.. where would I find install docs? |
| 04:40:05 | flypiper | thanks |
| 07:24:49 | pb_ | morning |
| 07:31:50 | cmarqu_ | Morning. |
| 07:32:03 | mallum | pb_: hey |
| 07:32:26 | mallum | pb_: looks like I got a 'discontinued' email from expansys too |
| 07:32:34 | pb_ | :-( |
| 07:32:39 | mallum | pb_: that sucks :( |
| 07:33:01 | mallum | pb_: Im gonna phone them up in a minute and moan |
| 07:33:30 | mallum | pb_: and there still on there web site ! |
| 07:33:49 | pb_ | mallum: yeah, I noticed that, sigh. |
| 07:34:00 | mallum | pb_: many thanks for the exec patch too btw |
| 07:34:29 | mallum | pb_: Im gonna go up tottenham court road later, maybe they'll have one there ... |
| 07:34:52 | pb_ | ah, good plan. |
| 07:38:07 | mallum | pb_: my job has gone even wierder, the company Im contracting for are totally cowboy |
| 07:38:28 | pb_ | mallum: heh, what's going on now? |
| 07:38:36 | mallum | pb_: Im going to be writing sms micropayment things for porn sites ! |
| 07:38:41 | pb_ | ! |
| 07:38:59 | mallum | pb_: I think the agency should have told me that ... |
| 07:39:09 | pb_ | well, they say that's the only way to make money on the web... |
| 07:39:49 | mallum | yeah, its hosted in the UK, so it cant be anything too bad ( I havn't seen it /know what it is ) yet |
| 07:40:10 | mallum | and its only the one |
| 07:40:28 | mallum | the sms stuff is quite fun |
| 07:41:01 | pb_ | cool, I've never done anything with SMS. |
| 07:43:17 | mallum | they use a third party ( mobliemessagecenter.com ), which you send soap requests to and you can do all sorts of things such as set logo's, ringtones, send special messgaes etc |
| 08:10:31 | cmarqu_ | pb_: Hey, cool. Gotta update. "GPE Firestarter" is cool too :) |
| 08:10:36 | pb_ | heh |
| 08:11:00 | cmarqu_ | Now, we need a nice GPE backdrop. Too bad gpe-appmgr is in the way :) |
| 08:11:24 | pb_ | I took support for button events out of gpe-login in this version too: as far as I know there's no reason you can't get that effect by running a keylaunch from gpe-login.setup. |
| 08:11:29 | pb_ | cmarqu_: indeed :-/ |
| 08:12:01 | pb_ | maybe I should make a bootleg gpe-appmgr ipk with quitting enabled. |
| 08:13:34 | cmarqu_ | We also need a gpe-changepassword. The root pw is too well known to not be changed. OTOH, for this, the command line could be enough. |
| 08:13:58 | pb_ | busybox provides "passwd" |
| 08:14:34 | mallum | whats firestarter ? |
| 08:15:05 | pb_ | mallum: some kind of codename - mibus made a splash screen that talks about GPE 0.0 "Firestarter". |
| 08:15:24 | pb_ | mallum: ipkg install gpe-bootsplash and reboot :-} |
| 08:15:55 | cmarqu_ | Should play some Prodigy too :) |
| 08:19:41 | cmarqu_ | OT: Does anybody know of a nice renaming tool for digicam photos? |
| 08:20:03 | cmarqu_ | Sorting hundreds of photos is soo boring... |
| 08:20:55 | cmarqu_ | "[Gpe] only c ? can i use pygtk ?" Uh oh. :) |
| 08:21:15 | pb_ | :-) |
| 08:21:26 | cmarqu_ | But Erez seems to be a good coder. |
| 08:22:25 | pb_ | moray: could you check in the familiar subdirectory under keylaunch? |
| 08:23:02 | pb_ | cnarqu_: Don't think I've ever seen any of his stuff, but I'll take your word for it. |
| 08:25:01 | cmarqu_ | Well, he is quick. Was asking for a file managaer and mailer, no real answer, so he wrote some on his own, withing a few days. |
| 08:25:19 | cmarqu_ | He also wrote that nvram kernel module. |
| 08:25:42 | cmarqu_ | He might not be the most organized person though. :) |
| 08:28:20 | pb_ | ha :-) |
| 08:29:24 | cmarqu_ | I didn't look at his code, and even if I had... I'm a VHDL guy, not a C guy. |
| 08:29:37 | pb_ | vhdl ugh |
| 08:29:38 | pb_ | ;-) |
| 08:29:47 | cmarqu_ | VHDL rules. :) |
| 08:29:54 | pb_ | verilog rulez |
| 08:30:01 | cmarqu_ | Stringly typed languages rule. IMO. |
| 08:30:12 | cmarqu_ | Verilog? You gotta be kidding! :) |
| 08:30:37 | cmarqu_ | For FPGAs, Verilog might be okay, but not for ASICs. |
| 08:30:57 | pb_ | ARM design all their cpus in verilog. |
| 08:31:37 | cmarqu_ | I can initialize a state machine with 5 bits state register with a 3 bit vector without being told. |
| 08:32:01 | cmarqu_ | I seem to remember an ARM softcore in VHDL? |
| 08:32:12 | cmarqu_ | Ha! Bug! |
| 08:32:27 | pb_ | Yeh, they do translate all their cores to VHDL too, probably wouldn't be able to sell them otherwise :-) |
| 08:32:49 | cmarqu_ | Unable to load icon "/usr/local/share/gpe/pixmaps/gpe-logo.p..." |
| 08:32:55 | pb_ | oh no |
| 08:32:59 | cmarqu_ | That's gpe-dm 0.14. |
| 08:33:11 | pb_ | probably gpe-login actually |
| 08:33:18 | cmarqu_ | Or that :) |
| 08:33:18 | mallum | what is verilog ? |
| 08:33:36 | cmarqu_ | I bet ARM simulate the VHDL to make sure their Verilog works! :) |
| 08:34:05 | cmarqu_ | mallum: VHDL and Verilog are the emacsen and vi's of the chip designer world. |
| 08:34:19 | cmarqu_ | Hardware Description languages. |
| 08:34:33 | pb_ | darn, gpe-login's makefile wasn't setting PREFIX=/usr for ipkg builds. |
| 08:34:47 | pb_ | building 0.15.1 now... :-/ |
| 08:35:03 | cmarqu_ | It still comes up, good. |
| 08:35:11 | mallum | If I do an upgrade is getty and god know what else gonna dissapear ? |
| 08:35:34 | pb_ | mallum: getty might go away, but so long as you don't reboot there is no cause for concern |
| 08:35:41 | pb_ | mallum: you can just "ipkg install getty" to get it back. |
| 08:36:13 | mallum | pb_: ok |
| 08:36:31 | mallum | I getting loads of errors from busybox ( which I din't event think I had installed ) |
| 08:36:56 | mallum | But that file is already provided by package textutils |
| 08:36:56 | mallum | ERROR: Package busybox wants to install file /usr/bin/env |
| 08:36:56 | mallum | But that file is already provided by package shellutils |
| 08:36:56 | mallum | ERROR: Package busybox wants to install file /usr/bin/logname |
| 08:36:57 | mallum | But that file is already provided by package shellutils |
| 08:36:59 | mallum | ERROR: Package busybox wants to install file /usr/bin/nc |
| 08:37:01 | mallum | But that file is already provided by package netcat |
| 08:37:03 | mallum | ERROR: Package busybox wants to install file /usr/bin/which |
| 08:37:05 | mallum | But that file is already provided by package debianutils |
| 08:37:07 | mallum | ERROR: Package busybox wants to install file /usr/bin/tty |
| 08:37:09 | mallum | and its just going in a loop :( |
| 08:37:13 | cmarqu_ | ipkg install busybox-update? |
| 08:37:16 | pb_ | mallum: hmm, looks like something is _trying_ to install busybox. |
| 08:37:37 | pb_ | ok gpe-login 0.15.1 is uploaded |
| 08:38:30 | cmarqu_ | http://burnalljpegs.org/ :) |
| 08:39:19 | mallum | pb_: how did you switch to busybox ? |
| 08:43:05 | pb_ | mallum: well, I don't recommend this as a good procedure, but: I did "ipkg install busybox", noted the list of packages it conflicts with, unpacked the busybox binary into a temporary directory, removed all the conflicting packages, used busybox to move itself into /bin, then did "ipkg -force-overwrite install busybox" to get the packaging right. |
| 08:43:44 | pb_ | mallum: the reason this is a bad idea is that the conflicting packages include things like the shell and fileutils, so there is some risk of getting into a situation you can't recover from when you remove those. |
| 08:44:12 | mallum | pb_: yes thats what I was worried about |
| 08:44:27 | pb_ | mallum: suggest you try the busybox-update package |
| 08:44:37 | mallum | pb_: yeah that seems to have worked |
| 08:44:41 | pb_ | cool |
| 08:46:23 | mallum | lets reboot and see what happens .... |
| 08:48:14 | mallum | pb_: hmm, I have gpe-dm but no xkbd ? |
| 08:48:39 | pb_ | hmm |
| 08:48:43 | pb_ | what gpe-login have you got? |
| 08:49:17 | mallum | 0.15 |
| 08:49:38 | pb_ | well, you really want 0.15.1, but I'm not sure it will make much difference to that. |
| 08:49:41 | pb_ | lemme try it on mine. |
| 08:49:59 | pb_ | huh, works for me. |
| 08:50:21 | pb_ | can you check with "ps", see if it started the xkbd up at all? |
| 08:51:37 | mallum | hmm, I dont have a ps ! |
| 08:51:47 | pb_ | whoa |
| 08:51:57 | pb_ | sounds like your busybox didn't install right |
| 08:52:15 | pb_ | try busybox --install -s |
| 08:53:10 | mallum | hmm, I have no busybox |
| 08:53:15 | pb_ | ! |
| 08:55:24 | mallum | ipkg sucks |
| 08:55:46 | pb_ | :-/ |
| 08:55:48 | mallum | ipkg install busybox gives; |
| 08:55:50 | mallum | pkg_hash_fetch_best_installation_candidate: apkg=busybox |
| 08:55:50 | mallum | provided by: busybox |
| 08:55:50 | mallum | Multiple providers. Not defaulting providers automatically.Please request one of the above providing packages by name.ERROR: The following packages conflict with busybox: |
| 08:55:51 | mallum | debianutils fileutils grep gzip hostname login mount net-tools netkit-ping sed shellutils tar textutils ash wget |
| 08:56:03 | mallum | but it wont let me remoce any of them packages |
| 08:56:47 | pb_ | sigh |
| 08:57:00 | mallum | I think I need to reflash it |
| 08:57:10 | pb_ | you could try talking to sandman |
| 08:57:32 | cmarqu_ | "< part> ipkg screws me over like I'm some cheap bitch" |
| 08:57:33 | mallum | I have a really old base image anyway, its more than likely that causing the problem |
| 08:57:47 | mallum | why is ipkg so verbose too |
| 08:58:03 | pb_ | ah, could be |
| 08:58:20 | mallum | does anyone know what package ps is in ? |
| 08:58:25 | pb_ | dunno, at least some of the debug seems to have been turned off again in the latest version. |
| 08:58:28 | pb_ | procps I think |
| 08:59:35 | mallum | xkbd is not running |
| 09:00:13 | pb_ | hmm |
| 09:01:43 | pb_ | hacked gpe-appmgr is in www.handhelds.org/~pb |
| 09:02:04 | pb_ | featuring: alpha blended tab and tray icons, plus reenabled "close" icon |
| 09:02:39 | mallum | hmm, I get an error when gpe-dm starts up complaining of no icon |
| 09:02:47 | pb_ | mallum: yeah, fixed in 0.15.2 |
| 09:03:04 | pb_ | mallum: you might wanna upgrade, should be in unstable very shortly. |
| 09:03:37 | mallum | pb_: still no xkbd though |
| 09:03:37 | pb_ | mallum: I'll let you know when it's available |
| 09:03:51 | pb_ | can you strace gpe-login starting up and mail me the results? |
| 09:04:04 | mallum | pb_: sure, how do I stop it ? |
| 09:04:14 | mallum | pb_: and no strace |
| 09:05:09 | pb_ | mallum: you'll have to kill gpe-dm and run gpe-login by hand from the console: |
| 09:05:18 | pb_ | mallum: strace -f -o strace.log gpe-login |
| 09:06:05 | mallum | pb_: I have no gpe-dm running, just the login |
| 09:06:17 | pb_ | ah |
| 09:06:26 | mallum | pb_: and when I kill it, it starts up again |
| 09:06:42 | pb_ | bizarre, where's it getting started from? |
| 09:06:55 | mallum | pb_: I dont know :( |
| 09:06:59 | pb_ | hell! |
| 09:07:12 | mallum | I dont think gpe-dm was installed |
| 09:07:24 | mallum | I've rebooted after installing it |
| 09:07:35 | pb_ | weird, that's the only thing that _should_ be spawning gpe-login. |
| 09:07:49 | pb_ | grep for gpe-login in /etc I suppose. |
| 09:08:04 | pb_ | gpe-login 0.15.2 is in unstable now, probably worth upgrading to that. |
| 09:09:32 | mallum | X11/Xsession-gpe.d/50autolock:gpe-login --autolock |
| 09:10:00 | pb_ | that doesn't get run until you actually log in |
| 09:10:54 | mallum | pb_: gpe-login laods xkbd if I run after killing it |
| 09:11:06 | pb_ | hmm |
| 09:11:22 | mallum | its probably a modmap issue |
| 09:11:31 | pb_ | could be |
| 09:11:38 | pb_ | that's supposed to be taken care of in gpe-login.setup |
| 09:11:44 | mallum | the server is getting reset after the xmodmap ( 0 clients on it ) |
| 09:11:48 | pb_ | ah |
| 09:11:54 | mallum | and xkbd is refusing to run |
| 09:12:01 | pb_ | okay, you ought to install gpe-dm |
| 09:12:24 | mallum | yes I have 0.14-1 installed |
| 09:12:55 | pb_ | if you let gpe-dm run gpe-login for you, it should curtail the server reset behaviour. |
| 09:13:38 | mallum | running gpw-dm, just flashes up then dissapears :( |
| 09:13:59 | pb_ | hmm |
| 09:14:32 | mallum | I have to go out now |
| 09:14:35 | pb_ | ok |
| 09:14:37 | mallum | will play with it later |
| 09:14:42 | pb_ | righto |
| 09:14:51 | mallum | Im sure its something simple, that I'be broken messing with stuff |
| 09:15:05 | mallum | thanks for helping me out |
| 09:19:07 | cmarqu_ | pb_: FWIW, gpe-login and gpe-dm are working nicely for me now, with locking and all. |
| 09:19:14 | cmarqu_ | Thx very much. |
| 09:19:16 | pb_ | cmarqu_: cool |
| 09:19:51 | cmarqu_ | BBL |
| 11:14:46 | flypiper | morning? |
| 11:14:57 | pb_ | afternoon :-) |
| 11:15:02 | flypiper | ahhhh |
| 11:15:10 | flypiper | afternoon:) |
| 11:15:50 | pb_ | heh :-) |
| 11:15:50 | flypiper | would like to install gpe on my 3700....whattta I need ? |
| 11:16:03 | pb_ | it's in the unstable feed: ipkg install gpe |
| 11:16:16 | flypiper | OK |
| 11:17:07 | flypiper | Thanks :) |
| 11:27:44 | flypiper | hmmm do I need to install task-x first.. giving me a depends error with xserver?? |
| 11:30:26 | pb_ | oh, yeah, probably |
| 11:30:35 | flypiper | :) |
| 11:30:37 | pb_ | you need at least xserver-tiny-h3600 |
| 11:31:01 | flypiper | OK,,this is on a new install of 5.3 |
| 11:31:12 | pb_ | make sure you get the version from unstable, not 0.5.3 or anything like that. |
| 11:31:22 | flypiper | k |
| 11:37:10 | flypiper | had to -force-depends (gpe) after installing xserver-tiny-h3600.. but looks like it's installing:-) |
| 11:37:28 | pb_ | ok, cool |
| 11:37:44 | pb_ | which package was it having dependency problems with? |
| 11:38:46 | flypiper | gives me xserver again and x-display-manager and gtksolitaire |
| 11:39:15 | pb_ | I think maybe you need the ipkg from unstable. Never mind. |
| 11:40:01 | flypiper | that's what my src is pointed to |
| 11:43:02 | pb_ | Yeah, but have you installed the latest version of ipkg itself? |
| 11:43:43 | flypiper | prolly not.. just flashed 5.3 bootstrap..:( |
| 11:43:48 | pb_ | ah |
| 11:44:01 | pb_ | well, it probably isn't that important. |
| 11:44:23 | flypiper | I'll see how it turns out.. not at all afraid to reflash:-) |
| 11:45:41 | pb_ | okay, cool |
| 11:46:01 | pb_ | if there are 0.6 prerelease jffs images available, you might want to reflash with one of those. |
| 11:46:04 | flypiper | had opee on this ipaq for my daughter to play games with on a trip to Miami last week.... couldn't wait to get it off :-\ |
| 11:46:09 | pb_ | heh :-) |
| 11:46:24 | flypiper | hmmm OK |
| 11:46:51 | flypiper | have to look |
| 11:51:18 | flypiper | finished install |
| 11:52:36 | flypiper | reboot... looks pretty broken .. kernel panic: Attempted to kill init! |
| 11:53:02 | flypiper | bin/init: no such file or directory |
| 11:55:28 | pb_ | crumbs |
| 11:55:37 | pb_ | I wonder how that happened. |
| 11:56:02 | flypiper | dunno.. let me see if I can use my 8M dual boot CF card and mount it to look at it |
| 12:03:48 | flypiper | found a task-busyboot.jffs2 in 6.0-pre1 ....lemme try that out:) |
| 12:05:00 | pb_ | cool |
| 12:05:26 | pb_ | I think gpe needs a load of stuff from 0.6/unstable, and upgrading a live system from 0.5.x seems to be quite tricky. |
| 12:07:07 | cmarqu_ | pb_: Got a few minutes? I want to make the "Login incorrect" text to be red in gpe-login. The docs about how to get the style from a gtkrc file are at http://www106.pair.com/rhp/gtk-colors.html, but I cannot figure it out. Once I know how it works, I can go and make the rest of the apps nicer. |
| 12:08:42 | cmarqu_ | Ouch. Wait, I might have gotten the wrong function. |
| 12:12:18 | pb_ | cmarqu_: okay, let me know if you are still having trouble. |
| 12:13:32 | pb_ | BTW: I checked in a load of changes to ipaq-sleep; it ought to support the sleep-on-low-battery thing we talked about the other day, though there isn't any facility for an external GUI yet. |
| 12:14:35 | cmarqu_ | Cool. |
| 12:16:07 | cmarqu_ | Damn. It's not working :) |
| 12:16:24 | cmarqu_ | (my gtkrc stuff) |
| 12:16:48 | pb_ | Hell. |
| 12:16:52 | cmarqu_ | :) |
| 12:17:21 | cmarqu_ | First, I have to give the widget a name. |
| 12:18:01 | cmarqu_ | gtk_widget_set_name (label_result, widget); (in enter_callback), isn't that right? |
| 12:19:11 | cmarqu_ | The whole thing is: |
| 12:19:16 | cmarqu_ | else { |
| 12:19:16 | cmarqu_ | gtk_label_set_text (GTK_LABEL (label_result), _("Login incorrect")); |
| 12:19:16 | cmarqu_ | gtk_widget_set_name (label_result, widget); |
| 12:19:17 | cmarqu_ | gtk_rc_parse("~/.gtkrc-gpe"); |
| 12:19:17 | pb_ | That's the right function, but I don't think the second argument is right. |
| 12:19:17 | cmarqu_ | gtk_widget_modify_style (label_result, "gpe_bold"); |
| 12:19:18 | cmarqu_ | } |
| 12:20:31 | pb_ | I _think_ all that's required is for you to put "gtk_widget_set_name (label_result, "result-widget");" in the code somewhere, |
| 12:20:56 | pb_ | and then: widget "label-result-widget" style "red-text-style" in your gtkrc. |
| 12:21:30 | pb_ | Adding a .gtkrc-rc seems like a reasonable thing to do, might even be worth putting that code in libgpewidget. |
| 12:21:40 | pb_ | (.gtkrc-gpe duh) |
| 12:22:16 | cmarqu_ | Ah. The thing is that I don't want to define the widget in the gtkrc file, but just the style. Then I want to apply the style name to the widget. That's what I did in storm. |
| 12:22:56 | cmarqu_ | I was parsing some files for all the styles first a global one, then a per-user file. |
| 12:24:29 | pb_ | Oh, I see. |
| 12:24:39 | Veggen | Hmm...is gpe nowadays in a "playable" status? |
| 12:24:45 | pb_ | In that case, I don't think you need to give the widget a name at all, so gtk_widget_set_name is not required. |
| 12:24:50 | Veggen | i.e. how do I set up a GPE environment? |
| 12:24:59 | pb_ | Veggen: ipkg install gpe |
| 12:25:10 | Veggen | pb: ok, did. but it won't start. |
| 12:25:19 | pb_ | cmarqu_: I broached the idea of making gpe-appmgr quittable with mibus. |
| 12:25:24 | pb_ | Veggen: what happens? |
| 12:25:26 | Veggen | specifically, it doesn't start X. |
| 12:25:34 | cmarqu_ | pb_: What does he say? |
| 12:25:35 | Veggen | before trying to start gpe-dm |
| 12:25:47 | pb_ | cmarqu_: "WHAT?!? gpe-appmgr is designed to be run all the time!" |
| 12:26:09 | pb_ | Veggen: it shouldn't, gpe-dm starts X itself. |
| 12:26:20 | Veggen | pb: oh...then why doesn't it? ;) |
| 12:26:45 | pb_ | Veggen: well, that's a good question. Try starting gpe-dm from the console, see if it prints any debug messages. |
| 12:26:53 | pb_ | Veggen: also try running /etc/X11/Xserver by hand, make sure that works. |
| 12:27:22 | Veggen | There's no /etc/X11/Xserver. |
| 12:27:30 | pb_ | cmarqu_: he does make the valid point that gpe-appmgr takes ages to start up, so you probably don't want to run it anew every time you need to launch a program. |
| 12:27:42 | pb_ | Veggen: ah, do you have the gpe-session-scripts package installed? |
| 12:28:05 | Veggen | no. |
| 12:28:19 | Veggen | ipkg install gpe didn't install it. |
| 12:28:51 | pb_ | Veggen: bizarre, gpe-apps depends on it |
| 12:29:03 | pb_ | Veggen: well, try installing it by hand. |
| 12:29:54 | Veggen | ok. that seemed to do the trick. trying a reboot. |
| 12:31:06 | Veggen | how do I write in the password-field? ;) |
| 12:31:22 | pb_ | you should have an xkbd, maybe that didn't get installed either. |
| 12:32:05 | Veggen | it didn't. |
| 12:32:21 | pb_ | Veggen: might be worth trying "ipkg install gpe-apps" once more |
| 12:32:39 | pb_ | Veggen: ipkg seems to have left out all kinds of dependencies. |
| 12:35:08 | Veggen | pb: it says it's up to date, and -force-reinstall doesn't do anything. |
| 12:35:21 | Veggen | well, I'll check out the deps manually. |
| 12:36:24 | pb_ | weird |
| 12:36:28 | pb_ | what version of ipkg are you using |
| 12:36:29 | pb_ | ? |
| 12:38:11 | Veggen | 0.99.23 |
| 12:38:34 | pb_ | hmm, same as me. |
| 12:38:46 | Veggen | some more that didn't get installed. |
| 12:40:17 | Veggen | a whole bunch, actually. |
| 12:40:23 | Veggen | but I installed manually.... |
| 12:40:28 | pb_ | sigh |
| 12:41:11 | Veggen | could actually be all that gpe-apps depended on, I had a few of them installed before. |
| 12:51:26 | cmarqu_ | Damn, I don't get that style stuff. |
| 12:51:44 | pb_ | :-( |
| 12:53:37 | pb_ | cmarqu_: can you show me the code you used in storm? |
| 12:56:00 | cmarqu_ | Sure, |
| 12:56:28 | cmarqu_ | frame.set_name("today_frame") |
| 12:56:28 | cmarqu_ | # make the label of the whole frame bold |
| 12:56:28 | cmarqu_ | rc_parse_string('widget "*today_frame*" style "storm_today_title"') |
| 12:57:05 | cmarqu_ | storm_today_title is defined in stormgtkrc |
| 12:57:18 | cmarqu_ | Ah. That gives me an idea. |
| 12:57:21 | pb_ | ah, I see |
| 12:57:35 | pb_ | so you just move the widget "name" style "name" into the program, rather than having it in the rc file. |
| 12:57:45 | cmarqu_ | Yeah, apparently. |
| 12:58:16 | pb_ | Okay, well, that will translate directly into C. Do: |
| 12:58:30 | pb_ | gtk_widget_set_name (label_result, "login_result_label"); |
| 12:58:32 | Veggen | well, gpe runs. |
| 12:58:33 | pb_ | to set the name |
| 12:58:48 | pb_ | and then the same rc_parse_string() stuff to get it associated with a style. |
| 12:58:54 | pb_ | Veggen: ok, cool. |
| 12:58:54 | Veggen | but, is appmgr meant to hide the non-gpe apps? |
| 12:59:14 | pb_ | Veggen: It hides things that don't have icons, I think. |
| 13:00:32 | Veggen | hmm...but it uses the standard familiar menu-methods-things? |
| 13:00:47 | pb_ | yup |
| 13:01:18 | Veggen | well. Then I want a gpe-appmgr option that says "not hide apps without an icon". |
| 13:01:38 | pb_ | yeah, file a report in bugzilla. |
| 13:01:46 | Veggen | on my way. |
| 13:02:22 | pb_ | meantime, I think the bootleg gpe-appmgr in www.handhelds.org/~pb will show things even if they don't have icons. |
| 13:03:15 | flypiper | well this task-busyboot.jffs2 won't boot.... has all kinda INIT errors and looks like /sbin/getty is forking like crazy:( |
| 13:03:17 | Veggen | is it the same bugzilla-account as for familiar? |
| 13:03:31 | Veggen | or do I need to register a new account? |
| 13:03:44 | pb_ | no, you need to register a new one, the two bugzillas are separate. |
| 13:03:51 | Veggen | ok. |
| 13:03:57 | pb_ | flypiper: ;-( |
| 13:08:23 | cmarqu_ | gtk_rc_parse_string("widget '*login_result_label*' style 'gpe_bold'"); |
| 13:08:29 | cmarqu_ | -:1: error: invalid string constant "gpe_bold", expected valid string constant |
| 13:08:44 | cmarqu_ | But gpe_bold is there in the gtkrc file. |
| 13:10:55 | cmarqu_ | Argh. it didnn't find the iule apparently. |
| 13:11:03 | cmarqu_ | s/iule/file/ |
| 13:17:44 | cmarqu_ | YEESSSSS! |
| 13:17:51 | cmarqu_ | It works. |
| 13:17:53 | cmarqu_ | :) |
| 13:19:06 | pb_ | yay! :-) |
| 13:23:01 | cmarqu_ | Now, what do we do with the resource files. Should we set them with gtk_rc_add_default_file()? In libgpewidget? |
| 13:23:21 | pb_ | Yah, that sounds like a good idea. |
| 13:23:59 | cmarqu_ | Then do gtk_rc_get_default_files (); in every app that uses gpe styles. |
| 13:24:44 | pb_ | Um, what does that do? |
| 13:24:58 | cmarqu_ | The problem is that it is almost impossible to get the current font and make e.g. a bold version of it. |
| 13:25:17 | cmarqu_ | "The set of these default files can be retrieved with gtk_rc_get_default_files() and modified with gtk_rc_add_default_file() and gtk_rc_set_default_files(). |
| 13:25:20 | cmarqu_ | " |
| 13:26:18 | cmarqu_ | gtk_init only reads: "Unless modified, the files looked for will be .gtkrc in the users home directory, and $localstatedir/gtk/gtkrc ($localstatedir defaults to /usr/local/etc)." |
| 13:26:55 | pb_ | hmm |
| 13:27:04 | cmarqu_ | wonders if gtk_rc_add_default_file() "modifies" things so that gtk_innit is affected. |
| 13:27:23 | cmarqu_ | Well, I can check this. |
| 13:40:15 | cmarqu_ | pb_: Where should that main gtkrc file go (in CVS, and also when installed)? |
| 13:40:35 | cmarqu_ | i.e., what is $localstatedir? |
| 13:43:59 | pb__ | argh nvidia strikes again |
| 13:48:35 | cmarqu_ | pb__: Did you get my question? |
| 13:48:41 | pb__ | no. |
| 13:48:48 | cmarqu_ | pb_: Where should that main gtkrc file go (in CVS, and also when installed)? |
| 13:48:54 | cmarqu_ | i.e., what is $localstatedir? |
| 13:49:03 | pb__ | /usr/share/gpe? |
| 13:49:30 | cmarqu_ | Yeah. Is that location in the makefile yet? |
| 13:49:47 | pb__ | Which makefile? |
| 13:50:06 | cmarqu_ | Well, should be in multiple makefiles... |
| 13:50:12 | cmarqu_ | Okay, I'll check. |
| 13:51:02 | cmarqu_ | Do I add gtkrc-gpe to libgpewidget? And where do I add the gtk_rc_add_default_file() function? |
| 13:51:28 | pb__ | Yes, I think so, and probably in init.c::gpe_application_init. |
| 13:52:12 | cmarqu_ | Ah, I overlooked the init.c file... |
| 13:54:38 | cmarqu_ | Should the per0user file be ~/.gpe/gtkrc or ~/.gtkrc-gpe? |
| 13:54:52 | pb__ | The former I think. |
| 14:07:25 | cmarqu_ | ln -s /etc/init.d/gpe-dm /etc/rc2.d/S99gpe-dm |
| 14:07:31 | cmarqu_ | ln: /etc/rc2.d/S99gpe-dm: File exists |
| 14:07:31 | cmarqu_ | make: *** [install-dm] Error 1 |
| 14:07:48 | cmarqu_ | (that is on the desktop) |
| 14:08:17 | pb__ | oops |
| 14:09:05 | pb__ | wanna check in a fix for that, or bugzilla it and I'll do it later? |
| 14:09:18 | pb__ | I'm going out for a bit, back later. |
| 14:09:29 | cmarqu_ | bugzilla :) |
| 14:23:58 | prpplague | howdy |
| 14:24:19 | moray | hello |
| 14:27:27 | prpplague | you guys gurus on the gpe project? |
| 14:28:01 | moray | well, some of us are developers, anyway ;) |
| 14:29:26 | prpplague | i'm familiar with mallum's matchbox, but not the other apps |
| 14:29:58 | prpplague | if your on the mailling list you'll see a post i did a few minutes ago about using fltk with GPE |
| 14:30:11 | prpplague | are you guys limiting contributions to gtk apps? |
| 14:31:20 | moray | prpplague: well, we want a consistent UI, and the GPE libraries are based on GTK... |
| 14:33:45 | moray | prpplague: do you have existing software you want to integrate with GPE, or is this for new development? |
| 14:37:00 | prpplague | moray: i've been developing a forms application, you drag and drop standardized questions onto a blank form, when you save the form, it compiles the form for arm |
| 14:37:39 | prpplague | moray: its currently in fltk, but i've been contemplating doing a re-write either in qpe or gtk |
| 14:38:26 | moray | ah |
| 14:39:25 | moray | prpplague: well, with the current version you can certainly use it *in* or *with* GPE (which you couldn't with the framebuffer-based QT stuff) |
| 14:40:53 | moray | prpplague: (GPE uses X, so you're not restricted to using only GPE programs) |
| 14:40:58 | prpplague | moray: agreed, but i'm looking for a project similiar to GPE to throw my support into |
| 14:41:41 | prpplague | moray: i assume there will be a learning curve moving to GTK, but it might be a better overall choice |
| 14:42:24 | moray | prpplague: well, I'd say so, but I'm possibly biased; I was using GTK for desktop development before doing handheld stuff anyway |
| 14:56:16 | prpplague | moray: do you use any gui dev tools like glade? |
| 14:57:33 | moray | prpplague: people are using Glade; it's encouraged (but not required) that you use libglade to load in the XML definitions rather than let Glade generate code for you |
| 14:58:15 | moray | quite a few things around at the moment have been Glade-based code |
| 14:59:28 | moray | (the reason for the slight discouragement against having Glade generate code is that it appears to make things bigger than they otherwise need to be, so for purely practical reasons) |
| 14:59:54 | prpplague | moray: yes, i'd assume so |
| 15:07:10 | pb_ | cmarqu_: I think you only want to do the gtk_widget_set_name()/gtk_rc_parse_string() once, when the widget is first created, not every time it's written to. |
| 15:19:55 | cmarqu_ | pb_: So, where is it created? THe set_name is necessary to make it show up with the new style at all. |
| 15:20:16 | prpplague | can you even buy an ipaq h3600 new anymore? |
| 15:20:29 | pb_ | cmarqu_: about line 559 |
| 15:20:34 | pb_ | label_result = gtk_label_new (""); |
| 15:20:56 | pb_ | prpplague: don't think so. |
| 15:21:22 | cmarqu_ | pb_: Ah. Let me play with it. |
| 15:21:49 | cmarqu_ | pb_: BTW, adding the gtkrc file in libgpewidget doesn't work :/ |
| 15:21:57 | pb_ | cmarqu_: why not? |
| 15:22:14 | cmarqu_ | pb_: I don't know... it works when I do this in the app itself. |
| 15:23:00 | pb_ | cmarqu_: well, gpe-login doesn't use gpe_application_init()... |
| 15:23:18 | moray | prpplague: oldest I see in stock on a random site is 3760 |
| 15:23:41 | cmarqu_ | pb_: Oh. Should it call that function? |
| 15:24:14 | pb_ | cmarqu_: well, that function has some undesirable effects for gpe-login, namely it tries to create ~/.gpe |
| 15:24:33 | cmarqu_ | :/ |
| 15:25:06 | pb_ | The directory creating probably doesn't belong in that function really. |
| 15:27:10 | cmarqu_ | Hmm. I don't feel like fiddling with that. |
| 15:27:23 | pb_ | Fair enough. |
| 15:28:03 | prpplague | moray: hmm, wonder what the support is like on the 3800 |
| 15:28:05 | moray | cmarqu_: it's not going to immediately do harm to use it with it making the directory, as long as it gets fixed before long... |
| 15:28:50 | cmarqu_ | pb_: About label_result... this could potentially get some other value than "Login incorrect", right? So I don't want to set the style alwasy, just when the login is incorrect. |
| 15:29:09 | moray | prpplague: http://www.handhelds.org/projects/h3800.html (don't know if that's up-to-date) |
| 15:29:18 | pb_ | cmarqu_: With the program the way it stands, that's the only thing it's used for. |
| 15:29:22 | cmarqu_ | moray: Well, it's fine as it is for now, marked up with FIXMEs... |
| 15:29:34 | pb_ | cmarqu_: I can't think of any other text that's likely to end up in that widget, really. |
| 15:29:44 | cmarqu_ | pb_: Okay. In case it gets other use, this has to be revisited then. |
| 15:29:48 | pb_ | Sure. |
| 15:30:14 | prpplague | moray: yes, i was reading that page, seems everything is supported except some more extensive battery management |
| 15:31:53 | prpplague | 's h3670 died last week |
| 15:33:10 | cmarqu_ | pb_: static const gchar *user_gtkrc_file = "~/.gpe/gtkrc"; |
| 15:33:26 | cmarqu_ | This doesn't really work like that, right? It doesn't expand the tilde. |
| 15:33:33 | pb_ | Right. |
| 15:33:48 | pb_ | Well, I assume so - it's conceivable that Gtk does that, but I doubt it. |
| 15:33:48 | cmarqu_ | strncat/snprintf it is then? |
| 15:33:53 | pb_ | Yup. |
| 15:34:06 | cmarqu_ | snprintf, since you don't string.h yet. |
| 15:36:51 | pb_ | moray: I added support to keylaunch for specifying an alternate config file on the command line. Are you in a position to make a new package at the moment? |
| 15:37:56 | pb_ | cmarqu_: hmm, mibus seems to use this idiom, which seems quite concise: |
| 15:38:00 | pb_ | fn = g_strdup_printf ("%s/.gpe/gpe-kbd", g_get_home_dir()); |
| 15:38:25 | moray | pb_: not just currently, but I should be in a few minutes - I'm about to do a 0.5.3 install to fix things/check for bugs |
| 15:38:34 | pb_ | moray: cool |
| 15:38:38 | cmarqu_ | pb_: Ah, nice. |
| 15:39:19 | moray | some of the other glib stuff is quite nice too - it's worth having a look at the API docs |
| 15:40:01 | cmarqu_ | Yeah, I remember from a time long ago... :) |
| 15:41:20 | cmarqu_ | Does it hurt that gpe_application_init has gtk_init and gpe-login too? |
| 15:42:07 | pb_ | You don't want to call that twice. |
| 15:42:19 | cmarqu_ | Hmm. |
| 15:42:24 | pb_ | If you're using gpe_application_init, take the gtk_init out of gpe-login. |
| 15:42:47 | cmarqu_ | Okay. But one is gtk_init (&argc, &argv); |
| 15:42:57 | cmarqu_ | the other one gtk_init(argc, argv); |
| 15:43:11 | pb_ | S'ok. |
| 15:43:19 | cmarqu_ | Good. |
| 15:43:25 | pb_ | Look at the way gpe_application_init is called. |
| 15:44:04 | cmarqu_ | Ah, yeah. |
| 15:46:19 | prpplague | mallum: dude |
| 15:49:36 | cmarqu_ | Hey, it works. Checking in. |
| 15:49:49 | mallum | hey |
| 15:50:01 | mallum | I got a bluetooth card now :)) |
| 15:50:08 | pb_ | mallum: way cool, which one? |
| 15:50:11 | moray | mallum: ah. found anything to do with it? ;) |
| 15:50:59 | mallum | pb_: not a sleeve , I just got a compact flash card for 50 quid |
| 15:51:17 | pb_ | mallum: fair enough, one of the Socket ones? |
| 15:51:18 | mallum | moray: my phone is bluetooth too, so hopefully connect to it |
| 15:51:23 | prpplague | mallum: 50 quid not too bad |
| 15:51:23 | mallum | pb_: sorry 75 quid |
| 15:51:56 | mallum | pb_: it looks like a cheap unbranded one |
| 15:52:38 | pb_ | mallum: oh right |
| 15:54:16 | mallum | pb_: I couldn't find the sleeve |
| 15:54:55 | mallum | pb_: :( |
| 15:55:04 | mallum | pb_: what is stating x now - gpe-dm ? |
| 15:55:20 | pb_ | mallum: yep |
| 15:56:13 | mallum | pb_: and its running stuff in Xsession-gpe.d |
| 15:56:47 | pb_ | mallum: right. |
| 15:57:40 | mallum | pb_: and does it run gpe-login before that ? |
| 15:58:34 | pb_ | yes |
| 15:58:39 | pb_ | mallum: see the web page for details |
| 15:59:16 | mallum | pb_: cool, checking the source too ... |
| 16:04:28 | mallum | pb_: it looks to me like xkbd is not working because there are no clients running and the server gets reset so it loses the modmap setting needed by xkbd for it to work |
| 16:04:47 | pb_ | mallum: gpe-dm connects itself to the server to avoid that very situation. |
| 16:05:05 | pb_ | mallum: so there should be at least one client at all times. |
| 16:05:36 | mallum | pb_: yes I see it opening a window ? |
| 16:06:09 | mallum | pb_: but you are creating a 0x0 window, which I dont think will work |
| 16:06:24 | mallum | pb_: you can start tinyX with a switch to stop it reseting |
| 16:06:29 | pb_ | mallum: it seems to work for me |
| 16:07:08 | mallum | pb_: let me add a dummy script in there to open an xterm off screen |
| 16:08:33 | pb_ | ok |
| 16:08:38 | cmarqu_ | pb_: About the gpe-login bug you assigned to me... would /etc/X11/gpe-login.setup be run before gpe-login comes up (or shows its login scree)? |
| 16:08:54 | pb_ | cmarqu_: see the web page :-) |
| 16:09:01 | cmarqu_ | Okay :) |
| 16:09:38 | moray | hm |
| 16:09:57 | moray | (ipkg is segfaulting if I do anything) |
| 16:10:18 | pb_ | moray: mmm, lots of people seem to have been stricken by that. |
| 16:12:35 | moray | pb_: do you know if there's an ipkg version that works, and is new enough to cope with installing (e.g.) gpe-apps? |
| 16:12:49 | pb_ | moray: no :-( |
| 16:12:54 | cmarqu_ | pb_: I am not convinced running an extra app is the right way. To not be in the way, the user info screen should go away with a single tap. |
| 16:13:03 | mallum | pb_: hmm running with Xipaq -noreset makes no difference :( |
| 16:13:13 | cmarqu_ | pb_: But now, imagine somebody has found the ipaq. |
| 16:13:37 | pb_ | moray: it seems that the latest version doesn't handle dependencies right: someone was trying to install earlier and, although they had gpe-apps installed, virtually none of its dependencies were there. |
| 16:13:45 | pb_ | mallum: hmm |
| 16:13:56 | cmarqu_ | pb_: The finder might tap the screen, but might not know how to get back to the user info to return it. |
| 16:14:04 | cmarqu_ | OTOH, that's how wince does it. |
| 16:14:38 | cmarqu_ | (return it == return the ipaq to the owner) |
| 16:14:45 | pb_ | cmarqu_: I would have thought there was enough space on the screen to display the user info alongside the login dialog. |
| 16:15:27 | cmarqu_ | pb_: Uh. Isn't that complicated? (For me, it is.) |
| 16:15:49 | pb_ | cmarqu_: why is it complicated? |
| 16:16:18 | cmarqu_ | Well, I wouldn't know where to start. |
| 16:17:17 | cmarqu_ | As an extra app, displaying stuff inside the gpe-login window? Isn't that complicated? |
| 16:17:23 | mallum | pb_: hmm it just sergfualts with strace |
| 16:17:28 | pb_ | cmarqu_: Not inside the window necessarily. |
| 16:17:39 | cmarqu_ | Ah. Hmm. |
| 16:17:52 | pb_ | cmarqu_: You can overlap another window on top of gpe-login, or change its size and display stuff next to it. |
| 16:18:20 | mallum | pb_: by the way ifd you want to create an invisable window you should make a 'inputonly' window |
| 16:18:27 | pb_ | mallum: ok. |
| 16:18:32 | mallum | pb_: look at source of xmonobut - that does it if IIRC |
| 16:18:37 | pb_ | mallum: what does inputonly do, exactly? |
| 16:19:15 | mallum | pb_: sets up a window that only wants input, has no 'visual' output |
| 16:19:25 | mallum | pb_: just general X wierdness |
| 16:19:25 | pb_ | mallum: cool, sounds like just the thing. |
| 16:19:54 | mallum | pb_: its so wierd the keyboard is not coming up |
| 16:20:05 | pb_ | mallum: btw, does matchbox do anything with WM_ICON_NAME? |
| 16:20:22 | mallum | pb_: dont think so ... |
| 16:20:48 | pb_ | oh well, just wondering if there was a way to get a pretty icon for rxvt. |
| 16:21:05 | pb_ | mallum: does gpe-login manage to display a keyboard if you run it standalone, without gpe-dm? |
| 16:21:16 | mallum | pb_: lets give it ago ... |
| 16:21:27 | mallum | pb_: does rxvt set WM_ICON_NAME ? |
| 16:21:40 | pb_ | mallum: seems to, yeah, via XSetIconName() |
| 16:22:29 | mallum | pb_: running gpe-login on its own I get a keyboard ! |
| 16:22:31 | pb_ | mallum: I'm not quite sure what WM_ICON_NAME actually means: my dim recollection is that window managers like fvwm used to go looking for an icon with that name in /usr/share/pixmaps. |
| 16:22:39 | pb_ | mallum: hmm, well, that's something I suppose. |
| 16:23:23 | pb_ | mallum: does gpe-dm produce any kind of debugging messages if you run it manuallu? |
| 16:23:32 | mallum | pb_: nope :( |
| 16:23:43 | pb_ | hmph. |
| 16:24:15 | pb_ | very strange. |
| 16:24:20 | moray | right, ipkg_0.99.19_arm.ipk seems to be the newest ipkg not to segfault (on an otherwise clean 0.5.3); it produces lots of debugging output, but if I ipkg install gpe-apps it doesn't pull in any of the dependencies at all |
| 16:24:35 | mallum | pb_: Im sure I support WM_ICON_NAME |
| 16:25:18 | pb_ | oh right |
| 16:25:44 | mallum | pb_: I'll look into why its not getting one for rxvt |
| 16:26:00 | mallum | pb_: but Im sure on metacity rxvt just gets the defualt icon |
| 16:27:32 | pb_ | mallum: xprop shows me: |
| 16:27:36 | pb_ | WM_ICON_NAME(STRING) = "rxvt" |
| 16:27:45 | pb_ | so it does seem to be setting it. |
| 16:27:56 | pb_ | possibly not to any useful value though :-| |
| 16:28:10 | mallum | pb_: uh maybe it just the wording to put beside the icon rather than the actual pixmap |
| 16:28:22 | mallum | pb_: that would make sense |
| 16:28:51 | pb_ | mallum: ah, hmm. |
| 16:29:47 | pb_ | mallum: I'm sure older window managers used to let you set a pixmap for rxvt. Maybe that behaviour has just fallen out of favour. |
| 16:29:56 | mallum | pb_: look in base_client.c for how matchbox gets the actual icon - It just uses X calls so Im not 100% sure what the hints are |
| 16:30:13 | mallum | pb_: sure I think it just does that with the WM_CLASS setting |
| 16:30:36 | mallum | pb_: theres no reason why matchbox could not do that in a later version |
| 16:30:50 | pb_ | mallum: ah right, cool. |
| 16:30:52 | mallum | pb_: but I wopuld like to get the Xsettings stuff working first and tie it into that |
| 16:31:31 | pb_ | righto. |
| 16:33:30 | mallum | pb_: I have to try compiling the xsetting implentation with signed char flag |
| 16:33:35 | cmarqu_ | pb_: gpe-login's Makefile has two "all" targets. |
| 16:33:51 | mallum | pb_: that may well fix it on powerpc |
| 16:36:25 | mallum | pb_: I switched back to nullxdm for now |
| 16:36:42 | mallum | pb_: strace is segfualting so its pretty impossible for me to find out whats going on :( |
| 16:36:46 | pb_ | cmarqu_: oops |
| 16:37:50 | mallum | pb_: could I just request that in the next ipkg of gpe-dm / gpe-login , it prints/logs something if the xkbd startup fails ? |
| 16:37:58 | pb_ | mallum: sure |
| 16:38:24 | mallum | pb_: Its very wierd :( |
| 16:38:34 | pb_ | mallum: I'll probably make a new release shortly, just trying to straighten out some issues with xkbd and autolock mode. |
| 16:38:36 | mallum | pb_: I gotta fix that modmap crap in xkbd anyway |
| 16:38:58 | mallum | pb_: ok, well if you want me to do anything with xkbd just say the word |
| 16:40:25 | mallum | pb_: I think I'll flash my other greyscale ipaq tommow and get a clean install on that |
| 16:40:37 | mallum | pb_: my 3600 ipaq eventually needs a reflash too :( |
| 16:41:13 | pb_ | mallum: right |
| 16:41:29 | cmarqu_ | pb_: And $(CC) is not set. |
| 16:41:42 | moray | hm, it really shouldn't take that long for 'whoami' to return... |
| 16:41:52 | pb_ | cmarqu_: it isn't supposed to be set. |
| 16:41:55 | mallum | pb_: one more quick que - the bluez tools are not in debian ? |
| 16:42:09 | pb_ | mallum: thought they were in unstable now, lemme check. |
| 16:42:10 | cmarqu_ | pb_: Okay then :) |
| 16:42:19 | pb_ | bluez-utils - Utilities for controlling Bluetooth devices |
| 16:42:27 | pb_ | try that one. |
| 16:44:55 | moray | hub.c: USB new device connect on bus2/1, assigned device number 53 |
| 16:44:56 | moray | usb_control/bulk_msg: timeout |
| 16:44:56 | moray | usb.c: USB device not accepting new address=53 (error=-110) |
| 16:44:56 | moray | hub.c: USB new device connect on bus2/1, assigned device number 54 |
| 16:44:56 | moray | usb_control/bulk_msg: timeout |
| 16:44:56 | moray | usb.c: USB device not accepting new address=54 (error=-110) |
| 16:45:00 | moray | argh |
| 16:46:45 | moray | I'd just use PPP, but NFS-over-PPP is *so* slow... |
| 16:47:30 | mallum | pb_: hmm, I dont think they've made it into debian-powerpc |
| 16:48:17 | pb_ | mallum: check buildd.debian.org? |
| 17:02:14 | mallum | pb_: will do |
| 17:02:24 | mallum | pb_: will probably play tommorow now ... |
| 17:04:08 | pb_ | mallum: can you add an option to xkbd to set its window name? |
| 17:04:39 | mallum | pb_: sure |
| 17:04:53 | mallum | pb_: what do you want todo that for ? |
| 17:04:59 | pb_ | mallum: to foil xsi :-/ |
| 17:06:02 | pb_ | mallum: when gpe-login is active in autolock mode, it keeps and xkbd inferior running at all times. Unfortunately, if you then use xsi to try to start an xkbd, it latches on to the invisible gpe-login one, and nothing useful happens. |
| 17:14:54 | pb_ | cmarqu_: I expect you have to g_free (user_gtkrc_file) in init.c |
| 17:15:35 | cmarqu_ | Argh. |
| 17:21:17 | pb_ | btw, your changelog talks about |
| 17:21:22 | pb_ | * init.c (main): Add default resource files, doesn't work yet. |
| 17:21:32 | pb_ | but there is no main() in init.c. |
| 17:25:39 | mallum | pb_: arg. Let me look at xsi ... it really is a hack it latching on to window titles like that |
| 17:29:21 | pb_ | mallum: cool, please do. |
| 17:30:08 | cmarqu_ | pb_: Grr. This is what pcl-cvs generates. How do you do such things? Only check in things that belong to one app at a time? |
| 17:30:24 | mallum | pb_: who is the author of xsi ? |
| 17:39:26 | pb_ | mallum: no idea, moray might know. |
| 17:39:55 | pb_ | cmarqu_: write the changelog semi-manually, C-x 4 a. |
| 17:40:32 | mallum | pb_: it looks like its part of appmgr |
| 17:40:43 | mallum | pb_: so mimbus ? |
| 17:41:14 | pb_ | hmm |
| 17:41:15 | mallum | pb_: I'll mail him and advise he looks at recent monolaunch without relying on window titles |
| 17:41:18 | pb_ | * xsingleinstance |
| 17:41:19 | pb_ | * |
| 17:41:19 | pb_ | * Program to ensure that a single instance of an X window program |
| 17:41:19 | pb_ | * exists on the screen. |
| 17:41:19 | pb_ | * |
| 17:41:21 | pb_ | * Copyright (c) 2000, Merle F. McClelland for CompanionLink |
| 17:41:23 | pb_ | * Portions of this code, such as TryChildren and ClientWindow, are |
| 17:41:24 | cmarqu_ | pb_: Oh, I misread. I thought you meant the check-in comment. My fault. |
| 17:41:25 | pb_ | * based upon routines in the Xmu library, which is Copyright 1989, 1998, |
| 17:41:27 | pb_ | * The Open Group. All Rights Reserved. |
| 17:41:45 | pb_ | cmarqu_: I don't care too much what you put in the checkin comment. |
| 17:53:23 | cmarqu_ | Now the question is: where would I get the owner info from? |
| 17:54:39 | moray | pb_: we should probably have (working) xsi type code in a library (or back as a standalone program...), given that several things need it |
| 17:54:51 | moray | cmarqu_: erm... from the GPE database presumably ;) |
| 17:57:24 | cmarqu_ | moray: Ah. :) |
| 17:57:43 | pb_ | moray: indeed, seems like the kind of thing that should go in a library. |
| 17:58:48 | pb_ | moray: could you check in the control file etc for keylaunch, so I can build a new package myself? |
| 17:59:27 | moray | pb_: I'm just building a package now (not sure why the control file didn't get checked in before) |
| 17:59:35 | pb_ | moray: oh, cool. |
| 17:59:36 | moray | pb_: have you checked in all your changes? |
| 17:59:39 | pb_ | moray: yup |
| 17:59:41 | moray | fine |
| 18:01:32 | pb_ | hmm, of course Nils already included xsi in his libgpe library, so I suppose we could just use that one. Though I'd still prefer to see it somewhere else I think. |
| 18:05:09 | moray | if monolaunch does things more sanely, maybe we need to get mallum to rewrite it, anyway ;) |
| 18:06:45 | pb_ | yeah :-) |
| 18:07:22 | cmarqu_ | Is there any app that uses usqld yet? |
| 18:07:33 | pb_ | Don't think so. |
| 18:07:57 | moray | I think everyone's waiting for the schema etc.... :-/ |
| 18:08:16 | cmarqu_ | So, mine won't be the first, and I guess it doesn't make sense to have yet another app that uses sqlite directly... |
| 18:08:23 | pb_ | Yeah. At the moment each app has its own little database, hence no contention, so there isn't much demand for usqld. |
| 18:09:32 | pb_ | From what I understand, the schema itself is basically finished now, we're just waiting for the library that implements it. In theory anybody could write an app right now that talks directly to the schema, and I've been getting increasingly tempted to do that. |
| 18:09:50 | cmarqu_ | Can stuff that is to be filled in externally be a label in GTK? I added my name as a default, and it ends up in the XML file in glade. Of course. |
| 18:10:02 | pb_ | cmarqu_: Well, usqld itself is basically transparent. You can pretty much just search-and-replace sqlite_exec() with usqld_exec() and you will be all set. |
| 18:10:39 | pb_ | On the other hand, it might be kind of cute to abstract the database manipulation out into (yet another!) shared library, so that one could drop in new database backends without source changes, and ideally without even recompiling. |
| 18:10:57 | moray | ok, new keylaunch uploaded; anyone with a working Familiar setup feel free to test... |
| 18:11:03 | cmarqu_ | Too hard for me! :) |
| 18:11:04 | moray | goes to eat before his food burns |
| 18:11:05 | pb_ | cmarqu_: Yeah, glade is kind of dumb like that. |
| 18:11:30 | pb_ | cmarqu_: But your app can fill in the proper value of the label at run time, overwriting the default value. |
| 18:11:36 | pb_ | moray: cool, I'll try it momentarily. |
| 18:11:38 | cmarqu_ | Ah. Good. |
| 18:12:04 | pb_ | I suspect all these useless default strings are one of the reasons why glade-generated interface.c files are so massive. |
| 18:12:25 | pb_ | ... another key one being the fact that it insists on giving every widget a textual name. |
| 18:12:46 | cmarqu_ | I have a table now, with labels all over the place, and then a scrollable GtkText widget for the address. |
| 18:12:59 | pb_ | cool! |
| 18:13:14 | cmarqu_ | Not sure if it looks too nice, but well... |
| 18:13:19 | pb_ | scap? |
| 18:14:03 | cmarqu_ | Well, it's on my desktop only. |
| 18:14:07 | cmarqu_ | :) |
| 18:14:14 | pb_ | Ah. |
| 18:15:05 | cmarqu_ | I could send you the XML file though. |
| 18:16:15 | pb_ | Sure |
| 18:16:19 | cmarqu_ | There's several things I don't like with it. The scollable address part looks like you could edit it :/ |
| 18:16:44 | pb_ | Use a smaller font and make it non-scrollable? :-) |
| 18:17:11 | cmarqu_ | Well, people might have loooong addresses. |
| 18:17:21 | pb_ | btw, just uploaded new versions of gpe-what and gpe-kbd that know how to register themselves with the dock's "Add" menu. |
| 18:17:31 | cmarqu_ | Ah, nice. |
| 18:17:35 | pb_ | cmarqu_: true, I guess. |
| 18:18:15 | pb_ | What's the status with gpe-login, does the gtkrc stuff work well enough for me to upload a new version? |
| 18:18:57 | cmarqu_ | It needs the libgpewidget of course, but I think it's usable. I only tried on the desktop though. |
| 18:19:08 | pb_ | okay. |
| 18:19:37 | pb_ | Well, I'll do a new libgpewidget package shortly, and think about whether to risk respinning gpe-login again today. |
| 18:20:40 | cmarqu_ | Hmm. Do I need stuff for the ipkg and dpkg targets in libgpewidget, to install the gtkrc file? |
| 18:21:06 | pb_ | Probably want to go in the "install" target. |
| 18:21:52 | cmarqu_ | I have install. If that's all... |
| 18:22:06 | pb_ | Yup. |
| 18:22:31 | pb_ | I don't recommend you try adding stuff to the dpkg target... :-} |
| 18:23:14 | cmarqu_ | Right :) |
| 18:23:31 | cmarqu_ | Argh. I can't dcc since I'm on a different machine with irssi,. |
| 18:23:36 | cmarqu_ | mail? |
| 18:23:41 | pb_ | sure, pb@nexus.co.uk |
| 18:25:27 | cmarqu_ | It really isn't much... |
| 18:26:47 | pb_ | looks okay to me. |
| 18:27:32 | pb_ | Have you tried to fit it on the screen alongside gpe-login? |
| 18:27:58 | pb_ | Obviously that's going to be a bit of a challenge to get right on all machines: e.g. Psion 7 has a rather different screen shape to iPAQ. |
| 18:28:24 | cmarqu_ | No, I haven't done anything like that yet. |
| 18:29:14 | cmarqu_ | In storm, we had a nice solution IMO: make the name a button, then expand all the rest when the name is clicked. |
| 18:29:16 | pb_ | Fair enough. Maybe you could fix bug #31 at the same time :-) |
| 18:29:35 | pb_ | Yeah, that would certainly be an option. |
| 18:30:33 | cmarqu_ | Maybe one could cram gpe-login's OK button right of the password field? |
| 18:31:03 | pb_ | Yep. |
| 18:31:24 | pb_ | I think that might look better actually, the "OK" button seems a bit forlorn out on its own like that. |
| 18:39:35 | pb_ | if anybody has any bright ideas on bug #40, please speak out :-) |
| 18:42:52 | prpplague | howdy all |
| 18:43:02 | prpplague | anyone got gpe stuff running on the zaurus? |
| 18:49:12 | pb_ | prpplague: not that I know of, but I imagine it would "just work" if you have an appropriate X server. |
| 18:49:44 | prpplague | pb_: ya thats the key, i'm afraid i'm going to have to build it myself |
| 18:50:17 | prpplague | pb_: i'm really surprised that there is not already a x package for the zaurus |
| 19:08:00 | flypiper | AHHHHH finally installed :-) |
| 19:09:03 | moray | flypiper: have you found an ipkg program that works properly, or just fetched the right packages manually? |
| 19:10:07 | flypiper | I installed 5.3 and changed my src to unstable then installed task-x and gpe.. using the -force-depends option |
| 19:10:38 | flypiper | but haven't got logged in from login screen yet |
| 19:12:45 | flypiper | there... does it make you add another user to login? |
| 19:14:01 | flypiper | now I don't have any kinda menu in the brown bar at the bottom? am I missing an app? |
| 19:14:06 | moray | flypiper: you're encouraged not to run GPE as root unless you really want to do that |
| 19:14:16 | flypiper | ahhh Ok |
| 19:15:51 | moray | flypiper: assuming you've got a recent matchbox version you should be able to get a menu from the bottom panel itself, and then add whatever buttons you want? |
| 19:16:51 | flypiper | moray: It installed matchbox 3.1-1 I believe |
| 19:17:25 | moray | flypiper: well, try click-and-holding on the panel at the bottom of the screen? |
| 19:17:29 | flypiper | ahhh have to tap and hold.. got it |
| 19:19:47 | flypiper | NICE :-) |
| 19:20:19 | moray | flypiper: you 'ought' to get a menu by default of course, but that sort of thing is in flux just now... |
| 19:30:07 | pb_ | moray: arranging a menu by default is actually slightly tricky: it would probably be easier just to document the fact that people need to add one to the dock themselves. Plus of course some people may not want a menu at all, if they rely on gpe-appmgr to do launching. |
| 19:31:47 | moray | pb_: true, I was forgetting about gpe-appmgr there |
| 19:31:55 | moray | pb_: if gpe-appmgr started by default it would be fine I guess |
| 19:40:03 | flypiper | appmgr did start after login |
| 19:40:21 | flypiper | I have calendar,Text Editor and to-do icons |
| 19:40:50 | flypiper | but it looks like dillo is missing ...must have to add them |
| 19:40:59 | flypiper | s/them/that |
| 19:41:10 | moray | ok - gpe-appmgr doesn't approve of packages with icons, so not everything gets displayed... |
| 19:41:29 | flypiper | k |
| 19:44:02 | moray | hm, with the current packages xkbd doesn't seem to be coming up on the login screen...? |
| 19:59:01 | pb_ | moray: hmm, mallum reported the same thing. But it works for me, so I'm a bit stumped. |
| 19:59:20 | cmarqu_ | Glade uses automake. Now I need to add libgpewidget to my gpe-ownerinfo... how do I do that? |
| 20:00:15 | pb_ | um, gpe-contacts probably does that too, have a look at its files. |
| 20:00:28 | cmarqu_ | Okay, thanks. |
| 20:01:21 | pb_ | moray: it would be interesting to know whether "ps" shows any xkbd child processes when gpe-login is running. |
| 20:01:56 | pb_ | moray: you might also try the very latest gpe-login from cvs, see if that behaves any differently. |
| 20:10:42 | moray | pb_: ah: xkbd |
| 20:10:42 | moray | xkbd: error while loading shared libraries: libXft.so.1: cannot open shared object file: No such file or directory |
| 20:11:14 | moray | I'll see if it works with libxft... |
| 20:13:30 | moray | also appears to need me to install the MS TT fonts... |
| 20:18:57 | flypiper | putting a screenshot of my display in scap.. after installing MS$ fonts |
| 20:25:17 | moray | pb_: xkbd appears fine now (after installing libxft and the MS TT fonts) |
| 20:25:23 | pb_ | moray: aha, cool. |
| 20:26:20 | pb_ | flypiper: cool |
| 20:27:22 | moray | libgdk_imlib isn't getting installed either (but is needed by appmgr) |
| 20:40:24 | cmarqu_ | Argh. There doesn't seem to be a make target which removes e.g. configure in a Glade project. |
| 20:40:45 | cmarqu_ | Or should I check in configure etc. into CVS? |
| 20:41:41 | moray | cmarqu_: is there not a 'distclean' or something similar? |
| 20:42:13 | pb_ | moray: hmm, the recent appmgrs shouldn't require imlib anymore, all that stuff should have been converted to use gdk_pixbuf. |
| 20:43:04 | cmarqu_ | moray: There is, but it doesn't clean as much as I want it to. maintainer-clean-am doesn't cut it either... |
| 20:43:44 | cmarqu_ | OTOH, gpe-contacts also has all that stuff... config.guess et al. |
| 20:44:03 | pb_ | cmarqu_: Yeah, generated files in CVS are a continuing bane of our collective lives. |
| 20:44:13 | cmarqu_ | Right... |
| 20:44:26 | pb_ | cmarqu_: config.guess isn't so bad, because it's static rather than being generated from other source files. |
| 20:44:39 | pb_ | cmarqu_: but, for example, every time I do "cvs update", I get: |
| 20:44:46 | pb_ | M GPE-Screen/Makefile.in |
| 20:44:46 | pb_ | M GPE-Screen/aclocal.m4 |
| 20:44:46 | pb_ | M GPE-Screen/config.h.in |
| 20:44:46 | pb_ | M GPE-Screen/configure |
| 20:44:46 | pb_ | M GPE-Screen/src/Makefile.in |
| 20:44:47 | pb_ | M GPE-Time/Makefile.in |
| 20:44:49 | pb_ | M GPE-Time/aclocal.m4 |
| 20:44:51 | pb_ | M GPE-Time/config.h.in |
| 20:44:53 | pb_ | M GPE-Time/configure |
| 20:44:55 | pb_ | M GPE-Time/src/Makefile.in |
| 20:44:57 | pb_ | etc etc |
| 20:45:08 | pb_ | So, definitely best to avoid checking these things in if at all possible. |
| 20:45:11 | cmarqu_ | I guess I could load the xml file, write the sources another time and check what is different to my current stuff. |
| 20:45:28 | cmarqu_ | pb_: .cvsignore for these files? |
| 20:46:19 | cmarqu_ | .cvsignore might not help though... |
| 20:46:33 | pb_ | cmarqu_: .cvsignore only really helps for files that haven't been checked in. |
| 20:47:16 | cmarqu_ | I don't find a place to add LINGUAS. It only is in generated files. :/ |
| 20:47:40 | pb_ | On the other hand, it can be equally frustrating if the generated files are _not_ in the archive, because then you need to run around finding the right versions of all the autotools in order to compile anything. I guess there's no silver bullet here. |
| 20:48:09 | cmarqu_ | Yeah... |
| 20:49:42 | pb_ | ah, the Opie bluetooth manager looks rather nice (in scap) |
| 20:51:37 | pb_ | moray: the new keylaunch seems to work fine, by the way. |
| 20:53:30 | pb_ | moray: one bizarre thing though: I press the calendar button, gpe-calendar starts (good). I press the launcher button, gpe-appmgr comes to the front (also good). I press the calendar button again, nothing happens (bad). If I press the calendar button a second time, gpe-calendar now comes to the front as I'd expect. |
| 20:58:18 | pb_ | flypiper: incidentally, what did you have to do to get GPE to install in the end? |
| 21:26:15 | cmarqu_ | gpe-ownerinfo in CVS. It doesn't really read the owner info from somewhere though. |
| 21:26:24 | pb_ | yay, excellent. |
| 21:26:44 | pb_ | wanna write a system setup utility so we have a way to configure the info? :-) |
| 21:26:53 | cmarqu_ | *cough* |
| 21:27:24 | cmarqu_ | I cannot even figure out where to set ALL_LINGUAS... |
| 21:28:16 | pb_ | Baffled by automake? You're in fairly good company there. |
| 21:28:44 | cmarqu_ | Hehe. |
| 21:29:21 | pb_ | Which program is it that creates these comical zero-length NEWS and README files, is that automake again? |
| 21:29:24 | cmarqu_ | Adding it to Makefile.am doesn't seem to help, it still complains when running autogen.sh |
| 21:29:38 | cmarqu_ | Seems so. |
| 21:29:57 | cmarqu_ | And these gettext messages in ChangeLog... |
| 21:30:17 | cmarqu_ | 2002-07-21 gettextize <bug-gnu-utils@gnu.org> |
| 21:30:17 | cmarqu_ | * Makefile.in.in: Upgrade to gettext-0.10.40. |
| 21:30:28 | cmarqu_ | How funny. :/ |
| 21:31:17 | pb_ | hmm, what's the deal with i18n? You ship a de.po in po/, which is cool, but there doesn't seem to be a .pot file there. There's a gpe-ownerinfo.pot in the parent directory, but it doesn't look quite like I would expect. |
| 21:32:48 | cmarqu_ | Yeah, kinda strange. The po/gpe-ownerinfo.pot can be created with make update-po, so that's fine. |
| 21:32:59 | pb_ | Righto. |
| 21:33:02 | cmarqu_ | A user doesn't really need the pot file. |
| 21:33:28 | pb_ | Yeah, true enough. |
| 21:33:48 | pb_ | I get this when running gpe-ownerinfo, does that mean I need a newer libgpewidget? |
| 21:33:49 | pb_ | -:1: error: invalid string constant "gpe_username", expected valid string constant |
| 21:34:07 | cmarqu_ | Needs a new libgpewidget. |
| 21:34:17 | pb_ | Suspected as much. |
| 21:34:19 | cmarqu_ | Well, one from this afternoon. |
| 21:34:31 | cmarqu_ | cvs update: warning: directory CVS specified in argument |
| 21:34:31 | cmarqu_ | cvs update: but CVS uses CVS for its own purposes; skipping CVS directory |
| 21:34:44 | cmarqu_ | I guess I did something wrong during cvs add :/ |
| 21:34:46 | pb_ | I've built an ipk for libgpewidget 0.13, just want to test it before I upload it to familiar. |
| 21:35:02 | pb_ | Yeah, that normally means you did "cvs add *" or something of the kind. |
| 21:35:35 | cmarqu_ | Right :/ |
| 21:36:16 | cmarqu_ | (Note to self: you can't hide anything from pb...) |
| 21:36:30 | pb_ | ha! |
| 21:36:32 | moray | hm, I still can't work out what sh in Familiar doesn't like about my case statement... :-/ |
| 21:36:46 | pb_ | moray: which of the multitude of shells are you using? |
| 21:38:14 | moray | pb_: ash apparently (hadn't done anything to change it) |
| 21:39:20 | moray | pb_: I thought I'd make a proper init script for gpe-dm (since I was bored with killing it manually), but I've clearly done something wrong |
| 21:42:14 | pb_ | moray: yeah, ash is about the best on offer in familiar 0.5.x; the busybox ash in 0.6 and unstable seems to be a bit better. |
| 21:42:22 | moray | (oh, maybe it's not the case statement anyway, since it's actually been printing the right text I notice) |
| 21:42:31 | pb_ | well, a proper init script for gpe-dm would certainly be a bonus. |
| 21:42:45 | pb_ | what's the actual problem you're having? |
| 21:43:25 | moray | it looks like the killproc I pasted in from the /etc/init.d/x isn't working, in fact |
| 21:43:41 | pb_ | ah |
| 21:44:49 | pb_ | (libgpewidget 0.13 is uploaded now.) |
| 21:44:59 | moray | anyway, it was leading to amusing results like this...: http://www.handhelds.org/scap/port.27255.png |
| 21:48:42 | pb_ | wow |
| 21:50:08 | cmarqu_ | I gotta add an ipkg target to the Makefile.am tomorrow... |
| 21:52:59 | pb_ | I just have a piece of boilerplate that I paste into all the makefiles to handle stuff like dpkg and ipkg. It might work in Makefile.am too, I dunno. |
| 21:54:34 | cmarqu_ | gpe-sketchbooks ipkg target doesn't look tailored for Makefile.am's, and I guess it works... |
| 21:54:54 | cmarqu_ | What apps use your template? |
| 21:55:13 | pb_ | gpe-su tends to be where I copy it from. |
| 21:55:40 | pb_ | I keep thinking maybe I should create a makefile fragment that can just be included from all these other places. |
| 21:59:31 | cmarqu_ | gets himself a beer after the hard coding work... :) |
| 22:01:06 | moray | ok, I finally have a working /etc/init.d/gpe-dm - further tidying up can wait till later |
| 22:03:03 | pb_ | wahey |
| 22:05:10 | moray | it would *possibly* be sensible for it (like the other dm scripts in Debian) to restart gpe-dm if it dies, but at the moment that doesn't seem to be a problem for us? |
| 22:07:02 | pb_ | Well, I haven't heard anybody reporting untimely demise of their gpe-dm. I guess there's always a risk it could fall foul of the OOM killer or something though. |
| 22:09:16 | moray | (actually, to be more correct, I think it's the display managers that restart X, rather than something restarting the display manager) |
| 22:10:05 | moray | pb_: what the Right Thing is on handhelds also depends what we end up doing about storage in volatile RAM etc.... |
| 22:10:24 | pb_ | ah, gpe-dm already restarts the X server if it dies. |
| 22:10:38 | pb_ | at least, it ought to -- I haven't tested that very recently. |
| 22:11:14 | moray | oh ok - is it meant to show the login screen again at that point, presumably? |
| 22:11:20 | pb_ | yeah |
| 22:11:42 | flypiper | must be what im getting trying to change theme's ? |
| 22:12:10 | flypiper | if for some reason the theme doesn't work it gives me the login screen again? |
| 22:12:17 | pb_ | you're probably crashing matchbox: that will end the session. |
| 22:12:27 | flypiper | ahhhhh |
| 22:12:58 | pb_ | if you look at the menu files, "Log out" is actually implemented with mbcontrol -exit. |
| 22:13:17 | flypiper | I just noticed that :) |
| 22:18:05 | pb_ | night everybody |
| 22:19:33 | flypiper | must have crashed more than matchbox.. won't let me login now:( |
| 22:42:39 | flypiper | it kinda looks like mbcontrol isn't working |
| 22:44:07 | moray | hm, and just when mallum isn't here to defend himself ;) |
| 23:41:27 | cmarqu_ | Night all. |
| 23:42:32 | moray | night cmarqu_ |
