| 02:55:41 | spung | is away: actually, I *am* here. I'm just not paying attention to you |
| 03:16:08 | moray | hm, appearing only to be away? |
| 10:06:55 | mallum | spung: you there ? |
| 10:59:29 | spung | mallum: i am now |
| 10:59:32 | spung | is back (gone 08:03:54) |
| 10:59:55 | mallum | spung: ello |
| 11:00:21 | mallum | spung: I added -desktop switch to mbcontrol |
| 11:00:50 | mallum | spung: did you add the DESKTOP property to today ? |
| 11:03:14 | spung | no |
| 11:03:18 | spung | mallum: nice |
| 11:03:36 | mallum | slacker |
| 11:03:36 | spung | mallum: i'm redoing gpe-today |
| 11:04:03 | spung | mallum: and i'm having exams :\ |
| 11:04:32 | mallum | GdkAtom window_type; |
| 11:04:33 | mallum | GdkAtom window_type_desktop; |
| 11:05:25 | mallum | window_type = gdk_atom_intern("_NET_WM_WINDOW_TYPE", FALSE); |
| 11:05:25 | mallum | window_type_desktop = gdk_atom_intern("_NET_WM_WINDOW_TYPE_DESKTOP", FALSE); |
| 11:05:25 | mallum | gtk_widget_realize(window); |
| 11:05:25 | mallum | gdk_property_change(window->window, window_type, XA_ATOM, 32, GDK_PROP_MODE\ |
| 11:05:26 | mallum | _REPLACE, (guchar *) &window_type_desktop, 1); |
| 11:07:16 | spung | mallum: thanks :) |
| 11:08:18 | mallum | spung: now you have no excuse |
| 11:08:27 | mallum | spung: make sure you copy and paste that |
| 11:08:36 | mallum | spung: and keep it somewhere safe |
| 11:08:58 | spung | mallum: done that :) |
| 11:08:59 | mallum | spung: I need to test if the -desktop switch works |
| 11:09:15 | mallum | spung: nautilus is broke on my desktop so I cant test :( |
| 11:10:11 | spung | mallum: hmm.. i'd test it |
| 11:10:27 | spung | mallum: i mean, i'll try |
| 11:10:49 | spung | mallum: is it in handhelds.org CVS? |
| 11:10:58 | mallum | spung one sec |
| 11:11:53 | mallum | spung: ok commited |
| 11:12:06 | mallum | spung: it currently compiles with lots of warnings |
| 11:12:13 | mallum | spung: having got round to sorting that out |
| 11:12:20 | mallum | spung: run in Xnest to test |
| 11:12:24 | spung | ok |
| 11:12:31 | mallum | spung: do a 'Xnest :1 -auth /dev/null' |
| 11:12:36 | mallum | then export DISPLAY=:! |
| 11:12:38 | mallum | then export DISPLAY=:1 |
| 11:12:41 | mallum | matchbox & |
| 11:13:41 | mallum | ignore the export DISPLAY=:! |
| 11:14:01 | spung | mallum: apt-get install'ing xnset |
| 11:14:04 | spung | xnest |
| 11:19:58 | spung | mallum: USA won england? |
| 11:20:24 | mallum | spung: 2-1 brazil-england |
| 11:21:05 | spung | ohh |
| 11:21:12 | spung | and USA won? |
| 11:24:48 | mallum | USA v Germany ? |
| 11:24:54 | spung | yea |
| 11:25:15 | spung | mallum |
| 11:25:17 | spung | aclocal: couldn't open `configure.in': No such file or directory |
| 11:25:17 | spung | configure.ac:3: error: possibly undefined macro: AM_INIT_AUTOMAKE |
| 11:25:17 | spung | configure.ac:4: error: possibly undefined macro: AM_CONFIG_HEADER |
| 11:25:17 | spung | configure.ac:35: error: possibly undefined macro: AM_CONDITIONAL |
| 11:25:17 | spung | autoheader2.50: error: AC_CONFIG_HEADERS not found in configure.ac |
| 11:25:18 | spung | automake: couldn't open `configure.in': No such file or directory |
| 11:26:05 | mallum | spung: run ./autconfigure.sh first |
| 11:26:18 | spung | yeah, this is what happened when i run it |
| 11:26:32 | mallum | spung: are you running debian ? |
| 11:26:35 | spung | yes |
| 11:26:47 | mallum | unstable ? |
| 11:26:50 | spung | sort of |
| 11:27:02 | mallum | sort of ? |
| 11:27:07 | spung | bits of woody |
| 11:27:13 | spung | bits of old unstable |
| 11:27:19 | spung | bits of updated unstable |
| 11:27:23 | spung | you know... |
| 11:27:44 | spung | it's not an uptodate unstable |
| 11:28:06 | mallum | hmmm |
| 11:28:51 | mallum | make sure you have these installed ( with the _same_ versions ): |
| 11:28:53 | mallum | ii autoconf 2.53-4 automatic configure script builder |
| 11:28:53 | mallum | ii autoconf2.13 2.13-44 automatic configure script builder (obsolete |
| 11:28:53 | mallum | ii automake1.5 1.5-3 A tool for generating GNU Standards-complian |
| 11:28:53 | mallum | ii autoproject 0.15.1-2 create a skeleton source package for a new p |
| 11:28:53 | mallum | ii autotools-dev 20020320.1 |
| 11:29:10 | spung | hmm ok |
| 11:47:32 | mallum | any joy ? |
| 11:49:22 | spung | make installing |
| 11:50:54 | mallum | cool |
| 11:50:59 | mallum | just getting some lunch brb |
| 11:55:02 | spung | mallum: yay |
| 11:55:05 | spung | mallum: it works :) |
| 12:02:40 | mallum | the compile ? |
| 12:09:12 | mallum | ? |
| 12:16:06 | mallum | spung: are you awake ? |
| 12:16:10 | mallum | cmarqu: hello there ;) |
| 12:16:13 | spung | mallum: yay |
| 12:16:17 | spung | mallum: the desktop switch |
| 12:16:19 | cmarqu | Hi all. |
| 12:16:21 | mallum | spung: it works ? |
| 12:16:26 | spung | mallum: yes |
| 12:16:32 | spung | mallum: but there are some other bugs |
| 12:16:32 | mallum | spung: heh, cool :) |
| 12:16:38 | mallum | spung: such as ? |
| 12:16:51 | mallum | expects a few ... |
| 12:16:54 | spung | mallum: yea |
| 12:17:01 | spung | mallum: the taks list |
| 12:17:12 | spung | mallum: it stayed there after i killed all apps |
| 12:17:28 | mallum | spung: yeah, theres a few probs with that :( |
| 12:18:15 | mallum | spung: what else ? |
| 12:20:13 | spung | mallum: hmm.. |
| 12:20:19 | spung | mallum: it's not working this time |
| 12:20:59 | mallum | spung: what happens ? |
| 12:21:29 | spung | mallum: sometimes the windows come back |
| 12:21:33 | spung | sometimes they don't |
| 12:22:39 | spung | mallum: that time popup is not showing up in the right place |
| 12:22:47 | spung | mallum: and it won't disappear |
| 12:23:22 | mallum | spung: what windows ? after a mbcontrol -desktop ? |
| 12:23:27 | spung | yes |
| 12:23:51 | mallum | ok, well get DESKTOP type built into today and I'll test and fix |
| 12:24:02 | spung | =) |
| 12:24:15 | mallum | hows the xkbd docs going too ? |
| 12:24:26 | spung | mallum: sorry :\ |
| 12:24:39 | spung | mallum: will look at it after the Chemistry exam |
| 12:26:41 | mallum | spung: cool |
| 12:26:52 | mallum | spung: just tell me if you dont want to do it |
| 12:27:17 | mallum | spung: but at least mail me some detailed instructions on what you did so others can benifit |
| 12:28:39 | spung | mallum: no, i want to it really |
| 12:31:27 | mallum | hey pb |
| 12:31:30 | mallum | :( |
| 12:31:41 | pb | hey mallum, what's up? |
| 12:31:53 | mallum | pb: brazil ;) |
| 12:32:06 | pb | mallum: yeah, gutted |
| 12:32:29 | mallum | pb: still, we did better than anyone expected ( in the tournament overall ) |
| 12:33:24 | cmarqu | Ah, summer solstice is just an hour away... |
| 12:33:31 | cmarqu | pats emacs :) |
| 12:33:32 | mallum | pb: anyway, more importantly I've implemented a '-exit' switch in mbcontrol |
| 12:33:58 | mallum | cmarqu: do you know how to do a 'backwards delete' ( like C-U in bash ) in emacs ? |
| 12:35:13 | pb | mallum: yeah, not all that bad a result in the grand scheme of things, but always disappointing when we do lose :-/ |
| 12:35:16 | pb | mallum: yay! |
| 12:36:03 | mallum | pb: I did the desktop switch too for spung, but its a little buggy at the mo |
| 12:36:21 | mallum | pb: I'll get a pre2 out later today |
| 12:36:26 | pb | mallum: cool |
| 12:36:49 | pb | wow, loads of new packages in unstable today |
| 12:38:48 | pb | mallum: huh, one of my programs is now making the dock crash as well :-| |
| 12:39:02 | mallum | pb: :( |
| 12:39:13 | mallum | pb: crash on Xerrors ? |
| 12:39:29 | pb | mallum: could be, I get |
| 12:39:37 | pb | X Error of failed request: BadDrawable (invalid Pixmap or Window parameter) |
| 12:39:37 | pb | Major opcode of failed request: 62 (X_CopyArea) |
| 12:39:37 | pb | Resource id in failed request: 0x800001 |
| 12:39:37 | pb | Serial number of failed request: 258053 |
| 12:39:37 | pb | Current serial number in output stream: 258078 |
| 12:39:42 | pb | and then sometimes the dock segfaults |
| 12:39:59 | mallum | pb: yah I think I need to trap the Xerrors |
| 12:40:05 | mallum | pb: I've noticed it too |
| 12:40:23 | cmarqu | mallum: No, sorry. One would need to write a little function I guess. |
| 12:40:26 | pb | I ran it up again and this time it worked, so seems that it's some kind of intermittent failure |
| 12:43:06 | pb | mallum: ... and now it's happening all the time, seems to behave just like xstroke :-( |
| 12:43:48 | mallum | pb: do a ps |
| 12:43:59 | mallum | pb: the dock has died and some apps are left over |
| 12:44:13 | cmarqu | Lightning here, I better shut down for a while. Later. |
| 12:44:15 | pb | mallum: right, I have loads of monolaunches still lurking |
| 12:44:19 | mallum | pb: when the dock comes back up, they do something which cuase it to crash |
| 12:44:33 | pb | mallum: well, I can restart the dock and the left-over monolaunches seem to sort themselves out okay |
| 12:44:44 | pb | mallum: as in, they appear inside the dock and it doesnt crash |
| 12:45:01 | mallum | pb: yes but sometimes it does ? |
| 12:45:19 | mallum | pb: I think it could be a race condition |
| 12:45:37 | pb | mallum: well, no, it always seems to crash when I start up either xstroke or gpe-what when the dock is already running |
| 12:46:05 | mallum | pb: hmm, is gpe-what in cvs ? |
| 12:46:11 | pb | mallum: yup |
| 12:46:29 | mallum | pb: give me a little bit and I'll try and figure out whats going on# |
| 12:46:33 | pb | okay, cool |
| 12:46:37 | mallum | pb: Im working on the dock at the moment# |
| 12:46:44 | pb | mallum: right |
| 12:47:01 | mallum | pb: just trying to plug a menu into it first .... |
| 12:47:16 | mallum | pb: if I get a chance I'll put the tray into a lib too |
| 12:47:35 | mallum | pb: especially if you are using it for stuff too |
| 12:47:41 | pb | mallum: excellent |
| 12:48:40 | pb | mallum: yeah, gpe-what was my first attempt at a prototype context-sensitive help system for gpe; dunno if this is actually a good mechanism to use, but we'll see |
| 12:59:49 | mallum | pb: Im just adding to the sprawl that is mmenu |
| 12:59:59 | pb | mallum: right |
| 13:00:27 | mallum | pb: I've got to split it into more than 1 source file ... |
| 13:00:35 | pb | heh :-) |
| 13:02:06 | mallum | pb: its just grown and grown and grown ... |
| 13:02:12 | mallum | pb: and before you knwo it ... |
| 13:03:31 | pb | mallum: yeah, I know how it happens. |
| 13:03:53 | mallum | damn its 1600 lines :( |
| 13:04:06 | pb | crumbs |
| 13:14:02 | spung | oh boy, audiogalaxy's dead |
| 13:14:10 | spung | where am i going to get my playalongs now? |
| 13:14:35 | pb | spung: sounds like a nightmare |
| 13:14:43 | spung | indeed |
| 13:15:01 | spung | audiogalaxy had every jamey aebersold play-a-long imagined |
| 13:15:19 | spung | oh well, gift is getting there |
| 13:16:01 | mallum | pb: hmmm, I changed it so I could pass a switch with a position to it arbituarly with a switch. I planed to get the dock to lauching it and point it towards a menu structure in /tmp |
| 13:16:10 | mallum | pb: I now think that too much of a hack |
| 13:16:34 | mallum | pb: and I should probably spend my time tidying the mmenu src and turning the core of it into a lib |
| 13:17:12 | pb | mallum: mmm, you mean so that the dock can restore the menu from a previous session or something? |
| 13:17:33 | pb | cmarqu: that was a short thunderstorm |
| 13:17:50 | mallum | pb: no so it can produce another menu with thing like 'add X to dock', remove Y from dock etc etc |
| 13:17:57 | pb | malum: ah, I see |
| 13:18:05 | cmarqu | pb: Indeed, but enough rain for our plants I suppose |
| 13:18:25 | pb | mallum: does sound like a bit of a hack :-| |
| 13:22:32 | mallum | pb: yeah Im being lazy |
| 13:25:30 | pb | mallum: is there any property apps can set on their windows to control which buttons matchbox will put in the titlebar? In particular, gpe-appmgr has a mode where it ignores the "close" button; it would be cool if we could stop that button from showing up at all. |
| 13:27:47 | pb | mallum: also, can programs supply their own icon to be used in the drop-down window list? |
| 13:29:04 | mallum | pb: matchbox could be set up to not show a close button. The problem is alot of apps igonore the protocol and then end up with no close button when they need one which is annoying |
| 13:29:16 | pb | mallum: ah, hmm, I see |
| 13:29:35 | mallum | pb: what you can do you is popup a 'are you sure you want to close' window when you get the delete event ? |
| 13:29:44 | mallum | pb: and ignore it if you wish |
| 13:29:52 | mallum | pb: matchbox wont force closure |
| 13:29:55 | pb | mallum: yup, that would work |
| 13:30:05 | mallum | pb: I thinkt hats the 'standard' way |
| 13:30:34 | mallum | pb: as for the icon, you should be able to set a WM_HINT that contains an xpm of the app |
| 13:30:46 | mallum | pb: mb will pick up on this and use it |
| 13:30:48 | pb | mallum: right, cool |
| 13:30:57 | mallum | pb: mb will also scale it if it is bigger that 16x16 |
| 13:31:11 | pb | even better :-) |
| 13:31:23 | mallum | pb: I would guess there is maybe a standard gtk call to set that |
| 13:31:44 | pb | mallum: very probably, I'll look at the api |
| 13:31:54 | cmarqu | pb: Hmm, your fix fix for endtime < starttime has some, uh, flaws :) |
| 13:32:02 | pb | cmarqu: oh dear |
| 13:32:41 | cmarqu | pb: Well, at least its behaviour is not totally nice. |
| 13:32:46 | pb | mallum: ah yes, it's probably gdk_window_set_icon (judging from the name) |
| 13:32:53 | pb | cmarqu: can you elucidate? |
| 13:33:12 | cmarqu | pb: I set an appt at 13:30, end time 13:32 --> end time becomes 14:32. |
| 13:33:25 | pb | cmarqu: hmm |
| 13:34:08 | cmarqu | The ideal behaviour would be: let the end time follow the start time automatically, all the time |
| 13:34:14 | mallum | pb: found this; |
| 13:34:17 | mallum | void |
| 13:34:18 | mallum | set_window_icon (GtkWidget * this_window) |
| 13:34:18 | mallum | { |
| 13:34:18 | mallum | #include "pixmaps/icon.xpm" |
| 13:34:18 | mallum | GdkPixmap *pixmap; |
| 13:34:19 | mallum | GdkBitmap *mask; |
| 13:34:21 | mallum | pixmap = gdk_pixmap_create_from_xpm_d (this_window->window, &mask, |
| 13:34:23 | mallum | &this_window->style->bg[GTK_STATE_NORMAL], icon_xpm); |
| 13:34:25 | mallum | gdk_window_set_icon (this_window->window, NULL, pixmap, mask); |
| 13:34:27 | mallum | return; |
| 13:34:29 | mallum | } |
| 13:34:32 | cmarqu | That is, even when the dialog is not closed. |
| 13:34:35 | pb | mallum: yup, looks like the thing |
| 13:35:31 | pb | cmarqu: the 13:32 -> 14:32 thing is because of a DST thing I overlooked |
| 13:36:02 | pb | cmarqu: I'll fix it. |
| 13:36:09 | cmarqu | pb: Ah, okay. The old ugly DST thing... storm had problems with that too. |
| 13:36:37 | pb | cmarqu: yeah, the way the various libc time functions interact with DST and timezones is all fairly messy |
| 13:36:57 | cmarqu | ACK |
| 13:38:35 | pb | cmarqu: it's not completely clear to me whether the way mktime() behaves is a bug or a feature, but gpe-calendar is going to stop depending on those functions soon anyway so I'm not too worried. |
| 13:39:14 | cmarqu | pb: Cool. |
| 13:40:36 | cmarqu | wonders how to maybe mark up the fact that one of the times displayed in the calendar list is on a different day... |
| 13:41:05 | pb | cmarqu: I think that whole UI needs some reworking |
| 13:42:09 | cmarqu | pb: Oooh. segfault. Now, when I was making the end time at a day earlier than the start day, it told me that I cannot have the end time earlier then start time, I click okay, then segfault. |
| 13:42:17 | pb | ugh |
| 13:42:40 | cmarqu | At least that data didn't get saved. |
| 13:43:19 | pb | cmarqu: can you see where it's crashing? |
| 13:43:46 | cmarqu | pb: Will check. |
| 13:44:06 | cmarqu | pb: Can the calendar show little icons in the list? |
| 13:44:16 | pb | cmarqu: which list? |
| 13:44:28 | cmarqu | In the time list. |
| 13:44:49 | pb | sorry, I still don't follow |
| 13:45:40 | cmarqu | pb: The list where my appointments show up. Like "11:00-12:00 Go eat." |
| 13:45:49 | pb | cmarqu: in the day view? |
| 13:45:55 | pb | cmarqu: or the future view? |
| 13:46:00 | cmarqu | pb: Yes. Day view. |
| 13:46:12 | cmarqu | Sorry, it's not a list I guess :) |
| 13:46:27 | pb | cmarqu: okay, so what do you want the icon to look like? |
| 13:46:29 | cmarqu | But in the future view, it would be nice too. |
| 13:46:34 | pb | cmarqu: and is it just decorative, or is it button that does something? |
| 13:47:11 | cmarqu | pb: A small bell when we have an alarm, and a small "calendar sheet" when the time is at a different day. |
| 13:47:29 | moray | argh |
| 13:47:29 | pb | cmarqu: okay, sounds good |
| 13:47:30 | cmarqu | If I could toggle the bell on and off, that would be cool. |
| 13:47:40 | moray | how come code that worked last night doesn't work this morning? |
| 13:47:45 | pb | cmarqu: first step is to find some pixmaps, want to have a go at that? |
| 13:47:50 | pb | moray: :-( |
| 13:47:56 | cmarqu | In storm, we had a "hollow" bell for "no alarm set". |
| 13:48:04 | moray | (FSVO 'morning') |
| 13:48:19 | cmarqu | pb: Bell is in storm, gotta look for the calendar page. |
| 13:48:37 | cmarqu | #0 0x40348b20 in free () from /lib/libc.so.6 |
| 13:48:38 | cmarqu | #1 0x40348ac3 in free () from /lib/libc.so.6 |
| 13:48:38 | cmarqu | #2 0x401947ab in g_free () from /usr/lib/libglib-1.2.so.0 |
| 13:48:38 | cmarqu | #3 0x805cca2 in event_db_destroy (ev=0x81819a8) at event-db.c:311 |
| 13:48:38 | cmarqu | #4 0x8059d1b in click_ok (widget=0x8134000, d=0x8149e78) at event-ui.c:282 |
| 13:48:38 | cmarqu | #5 0x400b51c5 in gtk_marshal_NONE__NONE () from /usr/lib/libgtk-1.2.so.0 |
| 13:48:40 | cmarqu | #6 0x400e4b7c in gtk_signal_remove_emission_hook () from /usr/lib/libgtk-1.2.so.0 |
| 13:48:49 | pb | ugh, that's what I was afraid of |
| 13:48:55 | moray | pb: I'm going to claim it's libgpewidget's fault...: open("/usr/local/share/gpe/pixmaps/default/dir-up.png", O_RDONLY) = -1 ENOENT (No such file or directory) |
| 13:48:55 | moray | --- SIGSEGV (Segmentation fault) --- |
| 13:48:55 | moray | +++ killed by SIGSEGV +++ |
| 13:49:12 | pb | moray: agreed |
| 13:49:20 | cmarqu | pb: Wait. I realize I didn't rebuild libgpewidget. |
| 13:49:28 | pb | moray: file a bug, set the priority to "extreme", assign it to dc |
| 13:49:54 | mallum | pb: hmm, where does lib gpe expect its icons ( running gpe-what on desktop ) |
| 13:50:08 | pb | mallum: $PREFIX/share/gpe/pixmaps/default |
| 13:50:28 | pb | mallum: prefix is /usr/local by default, so if you just do "make install" in gpe-icons it should come out right |
| 13:51:07 | mallum | pb: cool |
| 13:51:13 | mallum | pb: thats all been updated :) |
| 13:51:19 | pb | cmarqu: shouldn't be necessary, libgpewidget hasn't changed in any exciting ways I can think of |
| 13:51:24 | pb | cmarqu: though I suppose you can never be sure :-| |
| 13:51:52 | cmarqu | pb: New libgpewidget doesn't change it. Need more bt stuff or other gdb output? |
| 13:51:58 | pb | mallum: yay, gdk_window_set_icon seems to work |
| 13:52:05 | mallum | pb: :)) |
| 13:52:08 | pb | mallum: though the gpe-calendar icon doesn't look too good scaled down to 16x16 :-| |
| 13:52:18 | pb | cmarqu: no, lemme look at the code |
| 13:52:31 | cmarqu | pb: This shows the bell: http://storm-pim.sourceforge.net/screenshots/storm-7.png |
| 13:53:52 | pb | cool |
| 13:54:06 | cmarqu | Ah, the bell is from gnomecal :) |
| 13:54:32 | cmarqu | One can use the bell to drag the appt around. |
| 13:55:10 | moray | pb: which address is dc using for bugzilla? / is there a way to find out user addresses for bugzilla? |
| 13:55:26 | cmarqu | moray: dctanner@magenet.com |
| 13:56:49 | pb | moray: well, in the specific case of dc, he has some other bugs assigned to him so the query page would show you. In general, I'm not sure. |
| 13:58:21 | mallum | pb: maybe scale it in the gimp and set it to 16x16 ? |
| 13:58:33 | pb | mallum: yeah, that might help |
| 13:59:06 | pb | mallum: part of the problem is that most of our icons expect to be displayed with an alpha channel, they look a bit rough without |
| 13:59:25 | cmarqu | pb: Should the pixmaps be in gpe-icons or in gpe-calendar? I can check them in. |
| 13:59:33 | pb | cmarqu: stick 'em in gpe-icons |
| 13:59:54 | cmarqu | pb: Okay. As alarm-active/alarm-inactive ? |
| 14:00:39 | pb | cmarqu: fine. |
| 14:01:20 | mallum | pb: ok |
| 14:01:28 | mallum | pb: no png support yet in mb |
| 14:01:36 | cmarqu | pb: Ah, they are in CVS already, as gpe-icons/storm/alarm.xpm and alarm-shaded.xpm |
| 14:01:43 | mallum | pb: but there no doubt will be eventually |
| 14:02:01 | pb | mallum: would be great to have that, certainly |
| 14:02:34 | mallum | pb: would it be possible that when the dock is killed, it catches the signal and kills all its children ? |
| 14:02:57 | pb | mallum: what do you mean? |
| 14:03:36 | cmarqu | pb: gpe-icons/storm/storm-calendar-mini.xpm could be used for the other icon. |
| 14:04:16 | pb | cmarqu: ah, I see the bug |
| 14:04:29 | pb | cmarqu: it's actually a latent thing that's been there forever, missing "*" in event_db_alloc_details() |
| 14:04:40 | cmarqu | pb: Cool. You are always so quick... |
| 14:04:57 | mallum | pb: part of the problem is that if the dock is killed, then all its child windows should die too. otherwise they linger around and when the dock is relaunched they all try and dock at more or less the same time - which I *think* is causing the crash |
| 14:05:09 | pb | mallum: ah, I see |
| 14:05:23 | mallum | pb: I need to make sure that when the dock dies, it takes its kids with it |
| 14:06:16 | pb | mallum: right, yes. |
| 14:06:41 | pb | mallum: Being able to kill and restart the dock and have all the children automatically put themselves back in seems like quite a cool thing, but I guess from a practical perspective it isn't actually what you want. |
| 14:06:43 | moray | keylaunch works fine, but doesn't have support for a global configuration file - what's the best solution over that, since it would be nice if people didn't have to set it up manually? we can either patch it to look for a global file if there's no personal one, or have a wrapper that copies the global one to your homedir - an extra thing to consider is that once it's running it keeps checking whether the file it read has been updated |
| 14:07:49 | mallum | pb: yes, but the dock remembers what it launched, so it then goes and launches the whole lots again |
| 14:08:00 | pb | mallum: right |
| 14:08:04 | mallum | pb: so you get double the apps |
| 14:08:12 | pb | mallum: well, can you have it just send a delete event to each of its child windows when it dies? |
| 14:08:42 | mallum | pb: yeah but I have to catch the actual signal thats killing it too |
| 14:08:46 | pb | cmarqu: the crash and the DST thing should be fixed in CVS now |
| 14:08:56 | mallum | pb: for example if I run it from a shell and ctrl-c it |
| 14:09:06 | pb | mallum: yeah, the only signals you have to worry about are SIGINT and SIGTERM |
| 14:09:30 | pb | mallum: if someone sends it SIGQUIT or SIGKILL, they probably aren't expecting it to clean up after itself. |
| 14:10:06 | pb | mallum: at the moment it mostly gets killed by SIGSEGV on my machine, but that's probably not something you need to legislate for :-) |
| 14:10:06 | mallum | pb: right |
| 14:10:21 | mallum | pb: heh :) |
| 14:11:14 | pb | mallum: see gpe-login for example, it catches those signals and kills off its child processes (though it actually does that by sending them signals as well, rather than with X events) |
| 14:13:58 | pb | moray: dunno, I guess my inclination would be patch it to look for a global config file in /etc if the user doesn't have his own. |
| 14:13:59 | mallum | pb: BTW gpe-what appears to not be calling the tray_handle_events call |
| 14:14:07 | mallum | pb: that could be causing the problem |
| 14:14:08 | pb | mallum: ah, oops |
| 14:14:38 | mallum | pb: also you never want to call show(), tray will look after mapping the window for you |
| 14:14:50 | mallum | pb: just replace show with tray_init |
| 14:15:01 | moray | pb: yes, I think that's probably best - can we live without keylaunch automatically noticing if you then create one? |
| 14:15:06 | pb | mallum: yeah, trouble is that I think gtk_widget_show doesn't just map the window, it does some other stuff too. |
| 14:15:16 | pb | moray: sure, I wouldn't have a problem with that |
| 14:15:26 | pb | moray: maybe you could arrange for it to look again if you send it SIGHUP, or something |
| 14:15:40 | pb | mallum: but I'll give it a go, see what happens. |
| 14:16:19 | mallum | pb: hmm maybe call realise instaed ? |
| 14:16:49 | pb | mallum: yeah, realize is a partial solution (and in fact gpe-what already does that for other reasons) but I have a feeling it doesn't do everything that's needed. |
| 14:16:56 | pb | mallum: but I might be wrong about that, let's wait and see. |
| 14:18:15 | pb | mallum: it's just ClientMessage that tray_handle_event wants, right? |
| 14:19:03 | pb | btw, anybody tried out esd? |
| 14:20:20 | cmarqu | pb: now, once I had an invalid appt, the calendar popup doesn't come up again when I want to change the date. |
| 14:20:29 | cmarqu | pb: Didn't try esd, no. |
| 14:20:50 | pb | cmarqu: hmm |
| 14:21:28 | cmarqu | pb: The one for the start time works. |
| 14:22:30 | mallum | pb: yep. |
| 14:24:04 | mallum | pb: is it possible to send some void *data to a signal callback ? |
| 14:24:41 | pb | mallum: when you register a signal handler, you can set a void* pointer that's passed to the handler, is that what you mean? |
| 14:25:10 | pb | mallum: oh, sorry, ignore me, I have Gtk on the brain :-( |
| 14:26:00 | mallum | pb: yeah like gtk callbacks, but with signals :) |
| 14:26:06 | pb | mallum: I don't think you can do it with ordinary Unix signals |
| 14:26:13 | mallum | pb: :(( |
| 14:27:02 | pb | mallum: it might be possible with sa_sigaction style handlers |
| 14:27:37 | mallum | pb: yeah I think Id seen that |
| 14:28:15 | mallum | pb: but I may not need it now. If I do a kill(0, 15) and make all my kids and me in the same process group |
| 14:29:38 | pb | mallum: you probably want to put the kids in the process group but leave yourself out of it, then you can do "kill (-getpid(), 15)" and not have to worry about sending yourself the signal |
| 14:30:59 | mallum | pb: aha |
| 14:31:29 | mallum | pb: how do I initially create a process group with no actual members though ? |
| 14:31:42 | pb | mallum: do I have to do XSelectInput() to get ClientMessages, or do they always come through anyway? |
| 14:31:55 | mallum | pb: you should just get em |
| 14:31:55 | pb | mallum: process groups are intangible things, you don't have to create them |
| 14:32:06 | mallum | pb: no Im remebering now |
| 14:32:20 | pb | mallum: just do setpgid() with the group number you want, and you should be all set. |
| 14:32:45 | pb | mallum: see gpe-login again for an example |
| 14:32:49 | mallum | pb: on the first fork I call setgid(0,0) on the first fork then the gpid will be thats pid |
| 14:33:07 | mallum | pb: which I can get from the fork and make sure all other have that ? |
| 14:33:15 | pb | mallum: yup |
| 14:33:33 | pb | mallum: or you could call getpid() in the parent, and use that as the pgid value for all the children |
| 14:34:31 | mallum | pb: could be simpler |
| 14:34:43 | mallum | pb: and it dont mean Im in that group ? |
| 14:35:02 | cmarqu | BBL |
| 14:35:04 | pb | mallum: nope, group ids and process ids are completely separate |
| 14:35:32 | mallum | pb: cool but is it assured that'll be unique ? |
| 14:35:39 | pb | mallum: the fact that your process's pid and your children's pgid are the same number is just an amusing coincidence as far as the OS is concerned. |
| 14:36:18 | mallum | pb: okey dokey |
| 14:36:18 | pb | mallum: yes, pgids have to be unique -- by convention, most apps will only set the pgid to the pid of a process that they are in charge of, to avoid collisions. |
| 14:36:30 | mallum | pb: cool |
| 14:36:46 | mallum | pb: I need to read thoruhg my advanced unix programming book again me thinks |
| 14:36:49 | pb | heh |
| 14:39:08 | mallum | what does signal (SIGCHLD, SIG_IGN); do ? |
| 14:40:10 | pb | mallum: tells the kernel that you aren't interested in knowing when your children die. |
| 14:40:27 | pb | mallum: by default, dead children will hang around as zombies until you wait() for them |
| 14:40:37 | pb | mallum: but if you set SIGCHLD to ignored, the kernel will clean them up automatically. |
| 14:40:42 | mallum | pb: aha |
| 14:40:54 | mallum | pb: Id didn't realise the kernel would do that for you |
| 14:41:25 | pb | mallum: the kernel can be generous at times |
| 14:41:29 | mallum | pb: so no need to waitpid ? |
| 14:41:33 | pb | mallum: right |
| 14:41:46 | mallum | pb: oooh :) |
| 14:42:25 | pb | mallum: I added the message handling to gpe-what, doesn't seem to have made much difference. |
| 14:42:51 | mallum | pb: still crashing > |
| 14:42:52 | mallum | ? |
| 14:42:56 | pb | $ ./gpe-what |
| 14:42:56 | pb | gpe-what TRAY-DEBUG: found dock, sending opcode |
| 14:42:56 | pb | DOCK-DEBUG: woohoo, got a message |
| 14:42:56 | pb | DOCK-DEBUG: type is of tray |
| 14:42:56 | pb | DOCK-DEBUG: attempting dock |
| 14:42:57 | pb | DOCK-DEBUG: docking with 80, -90 : 200x200 |
| 14:42:59 | pb | TRAY-DEBUG: ouch got reparented |
| 14:43:01 | pb | gpe-what TRAY-DEBUG: message recieved |
| 14:43:03 | pb | Gdk-ERROR **: BadWindow (invalid Window parameter) |
| 14:43:07 | pb | serial 137 error_code 3 request_code 18 minor_code 0 |
| 14:43:09 | pb | [1] 3256 Segmentation fault dock |
| 14:43:48 | pb | I'm not sure what's causing the Gdk error, might be I have some bad window handle floating around in gpe-what itself. |
| 14:45:44 | mallum | pb: remove the dock atom stuff, thats not needed - though I doubt its that causing the problem |
| 14:46:07 | pb | mallum: do you know what the "request_code" means, can I match that up to a particular xlib call in my source code? |
| 14:46:29 | mallum | pb: you can look it up ( IIRC ) in Xproto.h |
| 14:46:34 | pb | mallum: cool |
| 14:46:55 | pb | mallum: X_ChangeProperty apparently |
| 14:47:15 | pb | which I suppose would point the finger at tray_set_xembed_info |
| 14:47:33 | mallum | pb: the way I normally trap them is , set an error handler with XSetErrorhandler and then run through gbd with a breakpoint on that |
| 14:47:40 | mallum | pb: yep checking |
| 14:47:48 | mallum | pb: Xerrors are a nightmare |
| 14:48:22 | pb | maybe my ClientMessage handling is still broken... |
| 14:48:29 | mallum | pb: the spec recommends you trap them |
| 14:48:43 | mallum | pb: reparenting outside of the window manager is a tricky bussiness |
| 14:48:57 | pb | I'm passing tray_handle_event the "display" and "window" values that came from the XAnyEvent structure, are those the wrong ones? |
| 14:49:14 | pb | mallum: right. |
| 14:49:30 | mallum | pb: the window should be your apps window , and it probably wont matter where the display comes from |
| 14:50:16 | pb | mallum: okay, I'll try that |
| 14:52:04 | mallum | pb: does this look ok to you ? |
| 14:52:06 | mallum | pid_t pid, mypid; |
| 14:52:06 | mallum | mypid = getpid(); |
| 14:52:06 | mallum | pid = fork(); |
| 14:52:07 | mallum | switch (pid) { |
| 14:52:07 | mallum | case 0: |
| 14:52:08 | mallum | setpgid (0, mypid); |
| 14:52:10 | mallum | execlp("/bin/sh", "sh", "-c", cmd, NULL); |
| 14:52:27 | pb | mallum: looks good to me |
| 14:52:31 | mallum | and then I have a ; |
| 14:52:33 | mallum | static void |
| 14:52:33 | mallum | cleanup_children(void) |
| 14:52:34 | mallum | { |
| 14:52:34 | mallum | kill(-getpid(), 15); |
| 14:52:34 | mallum | exit(0); |
| 14:52:36 | mallum | } |
| 14:52:47 | pb | yup, can't fault that. |
| 14:52:59 | mallum | and I've pointed the various signals at that ( I point an Xclose event there too ) |
| 14:55:12 | pb | mallum: I get the same results using my window id from a global variable |
| 14:55:38 | mallum | pb: ok commit it and I'll have a play |
| 14:56:30 | pb | mallum: done |
| 14:59:05 | mallum | pb: ok one sec |
| 14:59:32 | pb | going to the postbox, back in 5 minutes |
| 15:11:02 | pb | moray: might be nice if keylaunch could link in the same "run once" code that gpe-appmgr has, ie the thing that spots if you try to launch a program that's already running and just brings it to the front of the window stack. |
| 15:12:15 | mallum | pb: I cant see that gpe-what is calling tray_handle_events ? |
| 15:12:55 | pb | mallum: it calls tray_handle_event (singular) from filter() in gpe-what.c |
| 15:13:00 | pb | mallum: is that the function you mean? |
| 15:13:55 | mallum | pb: yes |
| 15:14:18 | mallum | pb: ahh yes I see |
| 15:14:59 | pb | mallum: it does seem to be getting called, I see TRAY_DEBUG messages coming out on the terminal. |
| 15:15:55 | pb | mallum: doing raw X stuff without upsetting Gdk and Gtk too much can be a bit icky, the "filter" thing seemed to be the best way of obtaining the events, although it looks a bit nasty. |
| 15:17:48 | mallum | pb: hmmm, it docks for me but as a big white square :( |
| 15:17:54 | mallum | rectangle rather |
| 15:17:58 | pb | huh |
| 15:18:09 | pb | that's further than I got |
| 15:18:26 | pb | try putting the "gtk_widget_show" back in, see if that makes the window come out looking any better |
| 15:18:34 | mallum | pb: ok |
| 15:25:06 | mallum | pb: ok I got a crash |
| 15:25:16 | pb | mallum: yay |
| 15:25:22 | pb | mallum: well, kind of |
| 15:25:26 | mallum | :) |
| 15:25:41 | mallum | I think this is gonna take a little figuring out |
| 15:26:01 | mallum | need to figure out how to gtk_show() a window, _without_ mapping it or its not gonna work |
| 15:26:18 | mallum | as soon as the window maps, the windowmanger will grab it |
| 15:26:26 | pb | mallum: I see |
| 15:26:31 | mallum | as its a direct child of the root |
| 15:26:57 | mallum | with the tray.c it only maps once its been parented by the dock |
| 15:27:00 | pb | right, so I guess the dock is trying to grab it as well but failing, hence the segfault presumably |
| 15:27:10 | pb | mallum: okay, I'll have a dig around in gtk. |
| 15:27:13 | mallum | pb: yep |
| 15:27:30 | pb | mallum: worst case, we can make a new routine that's identical to gtk_widget_show but the the map taken out, and call that instead |
| 15:27:43 | mallum | pb: though im thinking the dock may be reading the size wrong |
| 15:27:55 | mallum | pb: and therefore the window is going all wierd |
| 15:28:21 | pb | mallum: hmm, right |
| 15:28:43 | pb | mallum: well, I just do gtk_widget_set_usize(), dunno exactly what that translates to in X terms. |
| 15:28:47 | mallum | pb: I'll play |
| 15:28:51 | pb | cool |
| 15:29:07 | mallum | pb: I think that sets wm size hints rather than 'basic' window size |
| 15:29:18 | mallum | pb: the dock is only reading the basic window size |
| 15:29:34 | pb | mallum: well, Gtk has a separate function to set the "default" size, I thought that was the one that set wm hints. |
| 15:29:52 | pb | mallum: but I might have got that mixed up, I dunno. |
| 15:30:20 | mallum | pb: dont worry, the dock shouldf read the wm size hints anyway |
| 15:31:27 | pb | okay, cool |
| 15:33:42 | mallum | I should probably make the dock pick up sigsenv 's too for now !! |
| 15:33:49 | pb | heh :-) |
| 15:34:05 | moray | pb: maybe we should move the single-instance code (back) out of appmgr, if at least keylaunch should be using it too? |
| 15:34:19 | pb | moray: yes, that might be a good idea. |
| 15:34:26 | moray | for keylaunch it would be simplest if it could just run a command line thing... |
| 15:34:53 | pb | moray: it probably doesn't belong in libgpewidget, but we can make a new libsingleinstance or something. |
| 15:36:36 | pb | moray: right, yeah. Well, if the code was in a library it would be easy to make a simple command line wrapper round it. |
| 15:39:01 | moray | though if we're hacking keylaunch about anyway we could just go all the way and fork a gpe- version I suppose |
| 15:39:28 | moray | (the global config. file thing is different, and we could try feeding that to at least the Debian maintainer) |
| 15:39:29 | pb | moray: sure, it's not as if it's a huge program so I don't expect ongoing maintenance to be too much of a headache. |
| 15:39:54 | pb | moray: yep, agreed. |
| 15:46:02 | mallum | pb: even with me getting the size hints, its still wierd :( |
| 15:46:10 | pb | mallum: hmm |
| 15:46:17 | mallum | pb: must be some magic with that show() |
| 15:47:03 | pb | mallum: well, gtk_window_show does loads of stuff, including some kind of geometry computation. |
| 15:47:08 | pb | mallum: I'm looking at it now. |
| 15:47:48 | mallum | pb: this is a gnome implementation of the same thing -> http://cvs.gnome.org/lxr/source/libegg/libegg/tray/ |
| 15:47:58 | mallum | pb: probably a lot more advanced than mine |
| 15:48:55 | mallum | pb: I could make tray_init block until it finds a tray |
| 15:49:14 | pb | mallum: gack, we can't just copy gtk_window_show, because it calls all kinds of functions that are declared static inside gtkwindow.c |
| 15:49:48 | pb | mallum: the only way to override it will be to subclass GtkWindow, which is certainly possible but will take a while. |
| 15:50:42 | pb | mallum: looking at the libegg thing now |
| 15:51:09 | pb | mallum: ah, they use a GtkPlug and just set its xid to zero, maybe we could try that |
| 15:51:24 | mallum | pb: if I make tray_init block until it finds and tray and then deals with the messaging events before returning it may just work |
| 15:51:37 | mallum | pb: ahh yes plug may make sense |
| 15:52:48 | pb | mallum: I've checked it in, wanna have a go? |
| 15:53:02 | mallum | pb: sure one sec |
| 15:54:05 | mallum | pb: boom :( |
| 15:54:22 | pb | :-( |
| 15:54:49 | pb | what happened? |
| 15:56:04 | mallum | pb: it crashed first time :( |
| 15:56:07 | pb | :-( |
| 15:56:16 | mallum | pb: second time, didn't crash but nothing appeared |
| 15:57:33 | pb | hmm, right, same here :-( |
| 15:59:10 | mallum | pb: the code is very similar in structure to minw |
| 15:59:12 | mallum | mine |
| 15:59:19 | pb | mallum: yep |
| 15:59:52 | mallum | theres probably something simple Im doing wrong |
| 16:02:11 | pb | well, I could try actually dumping the eggtrayicon code into gpe-what, would we expect that to work? |
| 16:03:11 | mallum | pb: in theory yes |
| 16:03:24 | mallum | pb: if it dont then Im definetly doing something wrong |
| 16:05:54 | mallum | pb: am tidying up the tray.c code now anyway |
| 16:06:03 | mallum | pb: and checking it against libeggtray |
| 16:06:10 | pb | mallum: okay, cool |
| 16:09:13 | pb | argh, eggtrayicon is GTK 2 :-( |
| 16:09:20 | pb | it uses the GLib type system |
| 16:10:13 | mallum | :( |
| 16:10:30 | mallum | looks like theres no option but to switch to gtk2 then ;) |
| 16:10:33 | pb | heh |
| 16:14:10 | mallum | pb: the gtkcode is trapping Xerrors |
| 16:14:17 | mallum | eggcode that is |
| 16:14:27 | mallum | so I guess I'll impplement that too |
| 16:14:31 | pb | good plan |
| 16:14:31 | mallum | that could be enough |
| 16:31:30 | moray | shall I add a screenshots page to the site, with links to all the ones in images/ ? |
| 16:31:43 | pb | moray: good idea |
| 16:31:45 | moray | so we can point QPE types at it etc. ;) |
| 16:31:49 | pb | moray: also, hassle dc to make the new shots |
| 16:32:25 | pb | moray: it looks quite a lot better with a new matchbox, decent theme and so on. |
| 16:32:39 | moray | yes, I ought to sort that out on my own ipaq too... |
| 16:33:30 | moray | though I was wondering if we should find a bright-background GTK theme for the default? - that gives higher visibility in most conditions |
| 16:34:00 | pb | moray: well, the xenophilia theme has a pretty bright background |
| 16:34:19 | moray | ok, some of the screenshots I've seen people taking looked quite dark to me |
| 16:34:27 | moray | I don't know which themes they were though |
| 16:34:47 | pb | moray: I think most of the shots on gpe.handhelds.org are done with ThinGreeber, which uses gray60 or something as the background. |
| 16:37:40 | pb | mallum: I fiddled a bit more with the tray code. If I set the window size before realizing it (duh) then the dock displays a white square of the right size, which I suppose is some kind of improvement. |
| 16:38:18 | mallum | pb: actually a square ? |
| 16:38:28 | mallum | pb: I think we can do it with the plug code |
| 16:38:34 | pb | yep, it's meant to be 16x16 and it looks like that's what we get. |
| 16:38:43 | pb | mallum: okay |
| 16:39:12 | mallum | pb: I think you need to pass the window id of the dock to the plug function somehow |
| 16:39:39 | mallum | pb: you shouldn't need to call handle events, as the gtk plug stuff should take care of the XEmbed messageing |
| 16:39:39 | pb | mallum: hmm |
| 16:40:36 | pb | okay |
| 16:42:20 | mallum | pb: thats assuming the plug Xembed stuff in gtk1.2 is the same.... |
| 16:42:56 | pb | mallum: right, yeah. |
| 16:44:37 | mallum | pb: I've tided up the code |
| 16:45:07 | mallum | pb: do you want me to mail it over to you ? - < pb@[insert cool domain here].org ? |
| 16:45:22 | pb | sure, pb@nexus.co.uk is best |
| 16:46:32 | mallum | pb: ok on its way now |
| 16:46:36 | pb | cool |
| 16:47:09 | mallum | pb: after calling tray_init , you can now call tray_get_window() , which will give you the dock window ( if found ) |
| 16:48:35 | pb | okay, I'll try it. |
| 16:48:46 | mallum | pb: also Im trapping errors now |
| 16:49:01 | mallum | pb: and Im gonna trap errors in the dock too |
| 16:53:05 | pb | mallum: hmm, still not getting anything show up in the dock |
| 16:53:09 | pb | these are the debug messages I see |
| 16:53:14 | pb | DOCK-DEBUG: woohoo, got a message |
| 16:53:14 | pb | DOCK-DEBUG: type is of tray |
| 16:53:14 | pb | DOCK-DEBUG: attempting dock |
| 16:53:14 | pb | DOCK-DEBUG: docking with 4, 9 : 1x1 |
| 16:53:14 | pb | DOCK-DEBUG: got a tray begin message |
| 16:53:14 | pb | DOCK-DEBUG: event loop: conf request |
| 16:53:17 | pb | DOCK-DEBUG: event loop: conf request |
| 16:53:20 | pb | DOCK-DEBUG: got map request |
| 16:54:25 | cmarqu | moray: http://storm-pim.sourceforge.net/screenshots.html is pretty bright too. I just don't remember what dc used for those shots... |
| 16:56:34 | mallum | pb: hmmm, lis that with the plug stuff ? |
| 16:56:37 | pb | yup |
| 16:56:45 | pb | trying a few things now |
| 16:56:49 | mallum | pb: well there definelty messaging each other |
| 16:57:30 | mallum | pb: and it looks like the plug is trying to map itself |
| 16:57:38 | pb | hmm |
| 16:57:50 | mallum | pb: let me allow it to map itself |
| 16:57:59 | mallum | pb: do you want to check the code into cvs ? |
| 16:58:12 | pb | just a moment, I'll do taht |
| 16:58:55 | pb | okay it's there |
| 17:00:45 | mallum | pb: Ok I got it mapped :) |
| 17:00:57 | pb | wow, what did you have to do? |
| 17:01:43 | mallum | pb: with my xembed stuff, it'll only get mapped when the app requests it via a client message with appropriate opcode |
| 17:02:12 | pb | I see |
| 17:02:17 | mallum | pb: Im guessing the gtk1.2 xembed protocol doesn't do that but reuests a map via a redirect ( normal way ) |
| 17:02:23 | mallum | pb: I just honoured that |
| 17:02:34 | pb | aha |
| 17:02:39 | mallum | pb: how ever, though the icon is now showing the window is _very_ long |
| 17:03:06 | pb | mallum: move the gtk_widget_set_usize() before the gtk_widget_realize(), see if that helps |
| 17:05:53 | mallum | pb: no thats causing all sorts of wierdness :( |
| 17:06:00 | pb | mallum: ah, what sort? |
| 17:06:53 | mallum | pb: its not even getting to the dock |
| 17:07:01 | pb | :-( |
| 17:07:39 | pb | hmm, okay, well, put it back the way it was, and add "gdk_window_resize (window->window, 16, 16);" just after the gtk_plug_construct call. |
| 17:17:00 | mallum | hmmm |
| 17:17:14 | mallum | dock has started segfualting all over the place :( |
| 17:17:21 | pb | :-( |
| 17:17:22 | mallum | not sure if its related |
| 17:17:34 | pb | My dock seems to have taken to just busy-looping. |
| 17:18:58 | pb | time to go home, back in 45 minutes or so |
| 17:52:52 | dc_ | afternoon :} |
| 17:53:25 | moray | hello... |
| 17:53:49 | dc_ | moray: hey :D |
| 17:55:20 | mallum | hey dc |
| 17:55:26 | dc_ | mallum: :D |
| 17:56:13 | mallum | dc_: gtk hell today :( |
| 17:58:43 | pb_ | #join #handhelds.org |
| 17:58:45 | mallum | hey pb |
| 17:58:52 | pb_ | :-| |
| 17:59:02 | mallum | I give up :(( |
| 17:59:03 | pb_ | hey mallum, any news? |
| 17:59:10 | pb_ | ah, not the news I was hoping for |
| 17:59:12 | mallum | pb_: nope :( |
| 17:59:26 | mallum | pb_: gtk is still asking for a 400x600 window :( |
| 17:59:30 | pb_ | gah |
| 17:59:53 | pb_ | well, right now it would probably be easier for me to rewrite gpe-what without Gtk |
| 18:00:10 | pb_ | but it seems like docking Gtk apps is something we ought to be able to do... :-/ |
| 18:00:40 | mallum | yes indeed |
| 18:01:52 | pb_ | moray, what version of gpe-appmgr were you running for the screenshots? |
| 18:02:14 | pb_ | mallum, so, did you try doing an explicit gdk_resize_window call? |
| 18:02:31 | mallum | pb_: yep |
| 18:02:37 | pb_ | and it didn't make any difference? |
| 18:03:00 | mallum | pb_: nope :( |
| 18:03:09 | pb_ | sigh |
| 18:03:12 | mallum | pb_: you could do it with gdk to get lower level control |
| 18:03:45 | pb_ | yeah, we could try, but I have a feeling it's really gdk at fault. |
| 18:04:38 | pb_ | how did you find out the window size? |
| 18:04:41 | mallum | pb_: its like it rufusing to do a window *that* small |
| 18:07:16 | pb_ | mallum: can I run the dock on my desktop without the rest of matchbox? |
| 18:10:19 | pb_ | mallum: I see the answer is yes. |
| 18:10:23 | pb_ | :-) |
| 18:12:41 | cmarqu | Thunderstorm again... well be back I guess, but if not, see you in 3 weeks. |
| 18:12:58 | pb_ | cmarqu: bye |
| 18:13:09 | dc_ | cmarqu: bye! |
| 18:13:11 | dc_ | returns |
| 18:13:13 | dc_ | hey pb_ :} |
| 18:13:31 | pb_ | yo dc_ |
| 18:15:38 | dc_ | pb_: news? |
| 18:16:15 | pb_ | dc_: well, mallum and I have had a dock-busting session this afternoon |
| 18:16:21 | dc_ | ow |
| 18:16:38 | dc_ | pb_: gpe-what is cool, I assume the help pixmap is supposed to dock? |
| 18:16:50 | pb_ | dc_: gpe-edit, gpe-todo and gpe-calendar now all set icons for the window manager to use in its window list, though none of them look very pretty scaled down to 16x16 :-/ |
| 18:16:58 | pb_ | dc_: yep, that's the idea |
| 18:17:31 | dc_ | pb_: I can try and arrange some nicer 16x16 icons for you :} |
| 18:17:36 | pb_ | dc_: that would be cool |
| 18:17:37 | dc_ | us :D |
| 18:17:46 | pb_ | :-) |
| 18:20:15 | mallum | pb_: sorry back now |
| 18:20:33 | dc_ | 8:26pm ~/iPAQ/Apps/gpe/base/gpe-edit: make |
| 18:20:34 | dc_ | cc `gdk-pixbuf-config --cflags` `gtk-config --cflags` -D_GNU_SOURCE -Wall -O2 -g -I../libgpewidget -DPACKAGE=\"gpe-edit\" -DPREFIX=\"/usr/local\" -DPACKAGE_LOCALE_DIR=\"/usr/local/share/locale\" -c -o main.o main.c |
| 18:20:34 | dc_ | cc -o gpe-edit main.o `gtk-config --libs` `gdk-pixbuf-config --libs` -g -L../libgpewidget -lgpewidget |
| 18:20:34 | dc_ | main.o: In function `main': |
| 18:20:34 | dc_ | /home/dctanner/iPAQ/Apps/gpe/base/gpe-edit/main.c:327: undefined reference to `what_init' |
| 18:20:35 | dc_ | collect2: ld returned 1 exit status |
| 18:20:36 | pb_ | mallum: right now, gpe-what seems to make the dock get stuck in client_get_x_pos (according to gdb, anyway) |
| 18:20:36 | dc_ | make: *** [gpe-edit] Error 1 |
| 18:20:39 | dc_ | pb_: what have you done ;p |
| 18:20:43 | pb_ | dc_: new libgpewidget required |
| 18:20:59 | mallum | pb_: let me check ... |
| 18:21:08 | mallum | pb_: are you running cvs code there ? |
| 18:21:13 | pb_ | mallum: yup |
| 18:21:16 | dc_ | pb_: it is installed |
| 18:21:40 | pb_ | dc_: hmm |
| 18:21:45 | mallum | pb_: cool |
| 18:22:07 | pb_ | mallum: I'll rebuild with -g |
| 18:22:36 | dc_ | that error makes no senfce |
| 18:22:41 | dc_ | sence |
| 18:22:49 | mallum | pb_: uh ? it just worked for me !!!! |
| 18:23:26 | pb_ | mallum: maybe I broke gpe-what, can you mail me the current code you're using for that? |
| 18:23:30 | pb_ | dc_: agreed |
| 18:23:44 | mallum | pb_: it was with no wm running |
| 18:24:02 | mallum | pb_: with wm running its big again |
| 18:24:04 | pb_ | mallum: hm, maybe that makes the difference |
| 18:24:11 | pb_ | aha |
| 18:24:16 | mallum | pb_: yeah, maybe the wm is getting at it somehow ... |
| 18:24:25 | dc_ | arg |
| 18:24:57 | pb_ | dc_: I checked all my code is checked in |
| 18:25:04 | dc_ | arg :{ |
| 18:25:05 | dc_ | hrm |
| 18:25:09 | dc_ | taht's evil |
| 18:25:13 | pb_ | dc_: do "objdump --dynamic-syms libgpewidget.so.0 | grep what" |
| 18:25:43 | mallum | pb_: looks like matchbox is getting hold of it, resizing it and then the dock is grabbing it ! |
| 18:26:04 | dc_ | objdump: libgpewidget.so.0: No such file or directory |
| 18:26:06 | dc_ | err... |
| 18:26:06 | pb_ | mallum: damn matchbox! :-) |
| 18:26:23 | pb_ | dc_: go to the directory where libgpewidget is |
| 18:26:23 | mallum | pb_: matchbox is doing its job :) |
| 18:26:35 | pb_ | ha. |
| 18:26:49 | dc_ | 0000ad6c g DF .text000002fc Base what_init |
| 18:27:02 | pb_ | dc_: well, that's all well and good |
| 18:27:09 | pb_ | dc_: you're sure that's the library you're linking against? |
| 18:27:14 | dc_ | yeah |
| 18:28:06 | pb_ | and you have your libgpewidget.so symlink in place? |
| 18:28:11 | dc_ | ah ha! |
| 18:28:18 | dc_ | there was one in /usr/lib |
| 18:28:22 | dc_ | for some reason :/ |
| 18:28:25 | dc_ | so I rm'd it, and volla |
| 18:28:35 | pb_ | dc_: cool |
| 18:29:09 | dc_ | brb |
| 18:33:38 | mallum | pb_: the xembed protocol gtk1.2 is using is totally different I think |
| 18:33:46 | pb_ | mallum: oh dear |
| 18:37:14 | pb_ | gtk 1.2 doesn't seem to actually make any mention of XEmbed as such, maybe it predates that standard. |
| 18:37:20 | mallum | pb_: no Im not sure it is |
| 18:37:32 | pb_ | pb@kc:~/gtk+1.2-1.2.10$ find . -type f | xargs grep -i xembed |
| 18:37:32 | pb_ | pb@kc:~/gtk+1.2-1.2.10$ |
| 18:37:33 | pb_ | :-/ |
| 18:38:13 | mallum | pb_: the problem seems to be matchbox is getting hold of it just before the dock and this is causing all the wierdness |
| 18:38:39 | mallum | pb_: matchbox will only grab a window if it maps itself as a direct child of the roo |
| 18:38:42 | mallum | root window |
| 18:38:56 | pb_ | mallum: right |
| 18:39:07 | mallum | pb_: one of the gtk calls is causing a map |
| 18:39:12 | mccarthy | pb_: how come I cant compile gpe-todo? |
| 18:39:17 | pb_ | mallum: can I manually reparent the window into the dock, or will that cause other bad stuff to happen? |
| 18:39:19 | dc_ | can't we just have some xlib dock which will do it? |
| 18:39:22 | mallum | pb_: if we can figure out which one it is it should work fine |
| 18:39:24 | mccarthy | pb_: I fixed my bugs BTW... |
| 18:39:28 | pb_ | mccarthy: cool |
| 18:39:29 | dc_ | mccarthy: :} |
| 18:39:35 | pb_ | mccarthy: dunno, libgpewidget out of date? |
| 18:39:43 | mccarthy | pb_: maybe... |
| 18:40:07 | mallum | pb_: hmmm I dont think so |
| 18:40:18 | pb_ | mallum: do you have a call to gtk_widget_show (or show_all) after the gtk_plug_construct? |
| 18:40:25 | pb_ | mallum: if so, that's the most likely thing to be doing the mapping. |
| 18:40:25 | mallum | pb_: all your app has to do is sent the dock |
| 18:40:29 | mallum | pb_: yes |
| 18:40:32 | mccarthy | pb_: nope... |
| 18:40:38 | mallum | pb_: maybe a flush there would fix it |
| 18:40:39 | mccarthy | pb_: main.c:86: `pmap' undeclared (first use in this function) |
| 18:40:49 | pb_ | mccarthy: are your gpe-todo sources up to date? |
| 18:40:51 | mccarthy | pb_: among others |
| 18:40:57 | mccarthy | pb_: I think so... |
| 18:40:57 | mallum | pb_: do you know how to do a Xflush in gtk/gdk ? |
| 18:41:36 | pb_ | mallum: just call XFlush directly I guess. |
| 18:42:12 | mccarthy | pb_: must not have been.... I hope I didn't hose gpe-calendar fixing things off an older CVS image.... |
| 18:42:15 | pb_ | mccarthy: best to check, it seems to build okay for me at the moment but there was a period a little while ago when it was broken. |
| 18:42:36 | pb_ | mccarthy: don't worry, gpe-calendar doesn't change that rapidly :-) |
| 18:42:53 | mccarthy | pb_: I suck at CVS should I do a "checkout" or an "update" to make sure I am up to date? |
| 18:42:54 | mallum | pb_: arg still getting it :( |
| 18:42:58 | pb_ | mccarthy: update |
| 18:43:08 | pb_ | mallum: :-( |
| 18:43:24 | dc_ | so what's on the agenda today? |
| 18:43:35 | pb_ | dc_: well, fixing this pesky dock thing |
| 18:43:46 | dc_ | okay.. |
| 18:43:49 | pb_ | dc_: other than that, dunno, any thoughts? |
| 18:43:52 | dc_ | :/ |
| 18:43:54 | mccarthy | pb_: mind if I add the "feature" that gpe-tod will remember things that are clicked? (it really bugs me that I check things off the list and they are unchecked next time I run it!) |
| 18:43:58 | dc_ | I have to go and see... |
| 18:44:02 | dc_ | spider man! |
| 18:44:03 | dc_ | :/ |
| 18:44:07 | dc_ | at half past. |
| 18:44:11 | mallum | dc_: it rocks !!!!!!!!!! |
| 18:44:14 | mccarthy | dc_: spiderman rocks |
| 18:44:19 | dc_ | mallum: oooh good :DDD |
| 18:44:33 | pb_ | heh, I'll have to go see it if it comes that highly recommended |
| 18:44:37 | mccarthy | dc_: I liked it better than SW2 |
| 18:44:43 | dc_ | mccarthy: wow |
| 18:44:49 | dc_ | okay, I'm seeing it then :D |
| 18:44:54 | pb_ | mccarthy: sure, go ahead |
| 18:44:58 | dc_ | pb_: I'll tell you what I think tongith |
| 18:45:19 | pb_ | dc_: excellent, I await your review |
| 18:45:20 | mccarthy | pb_: is it just me or does the sql database *not* get updated when things are "checked"? |
| 18:45:39 | pb_ | mccarthy: well, it _should_ do, but I'm not 100% certain. |
| 18:46:55 | mccarthy | pb_: nope... I fixed it before, but never committed. I will simply fix it again... |
| 18:47:12 | pb_ | mallum: |
| 18:47:17 | pb_ | 117 if (p->mapped && !p->aligned_east) |
| 18:47:18 | pb_ | (gdb) print p |
| 18:47:18 | pb_ | $1 = (Client *) 0x804ef60 |
| 18:47:18 | pb_ | (gdb) print d->head_client |
| 18:47:18 | pb_ | $2 = (Client *) 0x804ef60 |
| 18:47:18 | mccarthy | pb_: any hope of getting "hidden" checked events? |
| 18:47:19 | pb_ | (gdb) print d->head_client->next |
| 18:47:21 | pb_ | $3 = (struct _client *) 0x804ef60 |
| 18:47:23 | pb_ | :-( |
| 18:47:34 | pb_ | mccarthy: sure, at some point |
| 18:47:40 | pb_ | mccarthy: faster if you implement it :-) |
| 18:47:42 | mallum | pb_: delete ~/.mbdock |
| 18:47:52 | mallum | pb_: see if that makes a difference |
| 18:48:16 | mccarthy | pb_: I think dc has got to get preferences set up for us first... or else how should it *know* to hide things? |
| 18:48:19 | mallum | pb_: gtk_plug_new needs the xid of the dock |
| 18:48:48 | pb_ | mallum: yeah, I was wondering about that. |
| 18:48:49 | dc_ | hu?? |
| 18:48:50 | mallum | pb_: yet we need the gtk_window plug returns to create the plug ! |
| 18:48:55 | pb_ | mallum: exactly! |
| 18:49:00 | dc_ | mccarthy: what's this? |
| 18:49:04 | mallum | pb_: I've never noticed that before |
| 18:49:19 | mallum | pb_: I may be able to find the tray without a window ... |
| 18:49:24 | mccarthy | dc_: are you still working on preferences? (for gpe-todo perhaps?)? |
| 18:49:36 | dc_ | mccarthy: I was working on gpe-config |
| 18:49:42 | dc_ | familiar config |
| 18:49:49 | dc_ | mccarthy: not app config, yet.... |
| 18:49:52 | dc_ | heh |
| 18:50:11 | mallum | pb_: yes I can chop up tray c for it to work |
| 18:50:15 | pb_ | mallum: I deleted that file. If I run gpe-what, stop it with ^C, and run it a second time, the dock reliably starts looping. |
| 18:50:34 | mccarthy | dc_: you could always make gpe-config have app-specific notebooks (that then are simply opened by the apps themselves when necessary)... as in make a library?! |
| 18:50:35 | pb_ | mallum: okay, gotta be worth a try. |
| 18:50:50 | mallum | pb_: becuase of all the wierd reparenting strange things are happening |
| 18:50:58 | pb_ | mallum: ah ok |
| 18:51:06 | mallum | pb_: I guess the client is never getting deleted from the list or something |
| 18:52:05 | dc_ | mccarthy: I think that would be overkill :/ |
| 18:52:16 | dc_ | mccarthy: the gpe-config apps are all seperate themselves |
| 18:52:35 | dc_ | mccarthy: I think having some sort of config functions will be better :} |
| 18:52:43 | dc_ | mccarthy: but we'll have to work on those :{ |
| 18:52:45 | dc_ | :) |
| 18:56:10 | mallum | pb_: is there a gdk thing to get the current display handle without passing a window to it ? |
| 18:56:48 | pb_ | mallum: yah, GDK_DISPLAY() |
| 18:56:48 | dc_ | pb_: hmm, the 16x16 icons don't look too bad :/ |
| 18:56:51 | dc_ | on my desktop |
| 18:57:14 | pb_ | dc_: no, they're certainly not disastrously bad |
| 18:57:18 | pb_ | dc_: would probably be better with alpha |
| 18:58:39 | dc_ | indeed |
| 18:58:46 | dc_ | brb |
| 18:58:56 | mccarthy | what are the 16x16 icons for? |
| 18:59:05 | dc_ | window titlebard |
| 18:59:07 | dc_ | bars |
| 18:59:55 | pb_ | mallum: I made a special nobbled gdk_map_window to stop the plug widget from getting mapped, but now it doesn't show up at all. |
| 19:00:14 | mccarthy | pb_: definitely needed a puch_item in the list-ui.c in order to save which todo's are actually *done* (without opening the edit window that is) |
| 19:00:21 | dc_ | bah |
| 19:00:26 | pb_ | mccarthy: right |
| 19:00:37 | dc_ | Qt dudes with transparent menus |
| 19:00:44 | dc_ | mallum: put that feature back in :D |
| 19:00:46 | pb_ | dc_: transparent menus are sick |
| 19:00:57 | dc_ | pb_: mallum had them in mment |
| 19:00:59 | dc_ | mmneu |
| 19:01:44 | moray | surely they just mean that you can't read the menu if you happen to have the wrong thing underneath? |
| 19:02:00 | moray | which is more of a problem on the ipaq where you aren't dragging windows about etc... |
| 19:02:06 | mccarthy | pb_: where would "purge completed" or "hide completed" buttons go?! (there is no more room) |
| 19:02:14 | pb_ | mccarthy: toolbar? |
| 19:03:18 | mccarthy | pb dc: are there icons that might actually be intuitive for purging and/or (un)hiding todos? |
| 19:03:31 | mccarthy | or do a need a drop-down menu |
| 19:05:15 | mallum | dinner time .... |
| 19:07:07 | pb_ | mallum: well, seems you were right, have a look at the current code in CVS :-/ |
| 19:08:11 | pb_ | mccarthy: dunno, I'd rather avoid a menu if possible, but I can't think of any icons that would fit the bill. |
| 19:08:13 | mccarthy | pb_: I guess it might actually be useful for me to have an account on your machines after-all.... none of the gpe stuff will compile on the ipaqs anymore... |
| 19:08:24 | pb_ | mccarthy: sure, send me your ssh key. |
| 19:08:31 | mccarthy | pb_: a little nuclear bomb might look like a purge :) |
| 19:08:37 | pb_ | :-) |
| 19:09:32 | pb_ | mccarthy: well, I suppose people should be able to figure it out either by experimenting or from the tooltips, even if the icons are a bit cryptic. |
| 19:11:26 | pb_ | mccarthy: one option would be to move completed items to the end of the list, so they would be the first things to go off the screen, and just purge them some (configurable) number of days after they're marked completed. |
| 19:13:27 | pb_ | mccarthy: ssh europa.armlinux.org |
| 19:14:14 | mccarthy | pb_: it asked me for a password... maybe I am using the wrong ssh... |
| 19:14:43 | pb_ | mccarthy: more likely one of our mailers scrunged the key, can you send it to me again as a MIME attachment rather than inline? |
| 19:14:56 | pb_ | mccarthy: if you can log into hh.org, the same ssh should work for europa |
| 19:15:06 | mccarthy | pb_: maybe... I just changed mailers.... |
| 19:15:11 | mccarthy | :) |
| 19:15:28 | moray | pb_: could you set up an account for me too? |
| 19:15:54 | pb_ | mccarthy: by the way, we're trying to maintain GNU-format changelogs for gpe-calendar and gpe-todo (plus a few other programs), can you add an entry there for the changes you made? If you use emacs, C-x 4 a is the magic key sequence. |
| 19:16:00 | pb_ | moray: sure, send me your ssh key. |
| 19:16:40 | mccarthy | pb_: dont use emacs... so I have no idea what you are talking about.... |
| 19:17:11 | pb_ | mccarthy: well, emacs is not obligatory :-) Just look at the ChangeLog file in each source directory, and add a new entry at the top for the changes you make. |
| 19:17:20 | moray | (hm, do the rest of us get to spam all the channels asking for money, if lilo does?) |
| 19:17:51 | pb_ | moray: only if you want people to set up morayfree.net, I suspect. |
| 19:18:25 | moray | :) |
| 19:18:57 | pb_ | mccarthy: okay, try again, the new key is in place |
| 19:19:04 | moray | (well, I'm already on several other IRC networks, so I only get it on a couple of channels anyway I suppose.) |
| 19:19:19 | pb_ | mccarthy: it does look differently formatted, I think the first one got line-wrapped :-( |
| 19:19:30 | mccarthy | pb_: yep. works now... |
| 19:19:38 | pb_ | mccarthy: okay, cool |
| 19:19:48 | pb_ | mccarthy: let me know if you need new packages installed, or anything like that |
| 19:19:58 | dc_ | bbl spiderman time :D |
| 19:20:02 | pb_ | dc_: enjoy! |
| 19:20:05 | dc[spiderman] | :D |
| 19:20:07 | mccarthy | pb_: ok. do you have the necessary stuff here to compile gpe-media? |
| 19:20:17 | pb_ | mccarthy: dunno, you'll have to try it and see. |
| 19:20:46 | mccarthy | pb_: what is the recommended way of getting CVS sources here? same as I do on my bock? (checkout?) |
| 19:21:35 | pb_ | mccarthy: yep. |
| 19:21:44 | pb_ | moray: okay, you should be set up on europa now as well |
| 19:22:48 | mallum | pb_: I've played with getting a window xid to the plug but still no joy :( |
| 19:22:59 | mallum | pb_: anyway now at least I know its possible |
| 19:23:04 | pb_ | mallum: :-| |
| 19:23:05 | moray | pb_: hm, I'm getting asked for a password - any clue at your end why it's not working? |
| 19:23:40 | mallum | pb_: I think Im gonna leave it to battle another day with a fresh head :) |
| 19:23:45 | pb_ | moray: one moment, I'll check. Might be worth trying "ssh -1", in case you have a protocol mismatch. |
| 19:23:51 | pb_ | mallum: yeah, me too. |
| 19:24:16 | mallum | pb_: I think I need the source of gtk at hand to do it |
| 19:24:26 | moray | pb_: er, I think that was probably an ssh2 key I sent you, are you running ssh1 there? |
| 19:24:28 | mallum | pb_: but gtk apps need to be able to dock |
| 19:24:51 | pb_ | moray: ah, try again now |
| 19:25:04 | pb_ | moray: I put your authorized_keys in the wrong directory. |
| 19:25:16 | moray | pb_: ok, it's working fine now, thanks :) |
| 19:25:20 | pb_ | moray: cool |
| 19:26:23 | mccarthy | pb_: do you normally keep libgpewidget up to date on these machines or should I keep it compiled in my CVS tree? |
| 19:26:34 | pb_ | mallum: I'm becoming increasingly suspicious that you just can't do it with Gtk1.2, short of some terrible hack like the gdk_map_window thing. |
| 19:26:49 | pb_ | mccarthy: keep in compiled in your own tree, I don't think it's installed at all. |
| 19:27:40 | mallum | pb_: I hope not :( |
| 19:27:46 | mallum | break now |
| 19:27:53 | mccarthy | pb_: how to get around the make ipkg non-root thingie? |
| 19:28:03 | pb_ | mccarthy, use fakeroot |
| 19:28:09 | mccarthy | pb_: ? |
| 19:28:19 | mccarthy | ignorant |
| 19:28:23 | pb_ | just do "fakeroot make ipkg" |
| 19:28:32 | mccarthy | pb_: oh.... |
| 19:28:41 | pb_ | it pretends that you have root privilege, which is good enough for ipkg. |
| 19:29:19 | pb_ | see the manpage for the details |
| 19:29:24 | mccarthy | pb_: I am getting gdk-pixbuf-comfig: command not found |
| 19:29:24 | mccarthy | ?! |
| 19:29:42 | pb_ | mccarthy: gdk-pixbuf probably not installed, I'll take care of it. |
| 19:29:57 | mccarthy | pb_: oh. I thought this was the machine you normally used... |
| 19:30:35 | pb_ | mccarthy: no, I have my own machine here in Cambridge. Europa is in Canada, it's one of the debian autobuilders. |
| 19:30:55 | pb_ | mccarthy: try again now |
| 19:31:45 | mccarthy | pb_: working ok now... I also got configure: error: can not find install-sh or install.sh in . ./.. ./../.. |
| 19:31:45 | mccarthy | trying to ./configure mibus's gpe-media |
| 19:32:12 | mccarthy | any ideas? |
| 19:32:36 | mccarthy | pb_: apt-get install gdk_imlib :) |
| 19:32:37 | pb_ | mccarthy: not really, I seem to have that file in my gpe-media directory |
| 19:32:56 | pb_ | mccarthy: what app needs that? we're trying to abolish gdk_imlib in favour of gdk_pixbuf |
| 19:33:05 | mccarthy | pb_: gpe-todo |
| 19:33:13 | mccarthy | main.c |
| 19:33:21 | mccarthy | I haven't looked at why.... |
| 19:33:35 | pb_ | mccarthy: try deleting the #include, it may just be a leftover artifact. |
| 19:33:52 | mccarthy | pb_: yep... |
| 19:33:57 | pb_ | ok cool |
| 19:34:07 | mccarthy | pb_: I *do* need sqlit installed thoough |
| 19:34:20 | mccarthy | cvs commits the new main.c :) |
| 19:34:53 | pb_ | sqlite is done. |
| 19:36:04 | mccarthy | pb_: should I provide my own ipkg-build? |
| 19:36:15 | pb_ | mccarthy: probably easiest for now |
| 19:36:16 | mccarthy | or do you want to do it system-wide |
| 19:36:37 | pb_ | mccarthy: well, if you put the file in your home directory somewhere, I can move it to /usr/local/bin easily. |
| 19:36:53 | mccarthy | pb_: its in ~mccarthy |
| 19:37:38 | pb_ | ok, done |
| 20:16:17 | mccarthy | pb_: why do the apps look for their "menu" icons in /usr/local/share/pixmaps ? |
| 20:17:33 | pb_ | mccarthy: well, they look in $PREFIX/pixmaps, because that's where they expect to install the icons. |
| 20:18:02 | mallum[tv] | pb_: this is driving me mmmmaddddd ddd ! |
| 20:18:12 | mccarthy | pb_: who decides what $PREFIX is? they are *already* installed in /usr/share/pixmaps |
| 20:18:36 | pb_ | mccarthy: it's set in the makefile, you can override it on the command line. |
| 20:21:15 | pb_ | argh, left my ipaq charger at the office :-( |
| 20:21:54 | mallum[tv] | pb_: I can now get it too get reparented, not touch the wm, but it refuses to map itself :( |
| 20:22:20 | pb_ | mallum[tv]: :-/ |
| 20:22:51 | pb_ | mallum[tv]: I think I might just make a subclass of GtkWindow that has the map disabled. From my dlsym experiments I reckon that will work. |
| 20:23:55 | mallum[tv] | pb_: yeah |
| 20:24:19 | mallum[tv] | pb_: there probably is a way of doing it without. but Its too painful now :*( |
| 20:24:27 | pb_ | mallum[tv]: agreed |
| 20:24:43 | pb_ | mallum[tv]: it's probably not worth making any heroic efforts if Gtk2 does proper XEmbed. |
| 20:27:22 | moray | pb_: (I've updated the appmgr screenshot) |
| 20:27:42 | pb_ | moray: cool |
| 20:28:27 | moray | and keylaunch should be patched to look for a global config file if there's no per-user one, I'll just test that etc. |
| 20:28:35 | moray | shall I put keylaunch into CVS? |
| 20:28:39 | pb_ | moray: yes please |
| 20:28:46 | mallum[tv] | pb_: yeah, just dont like being beaten by it :( |
| 20:29:06 | moray | pb_: actually, I guess I should check in a clean version of keylaunch first, to make it easier to see what went on later... |
| 20:29:33 | mallum[tv] | pb_: oh could you see quickly if xstroke is now docking ? |
| 20:30:17 | pb_ | mallum[tv]: not right this minute, I'm not able to talk to my ipaq just now. |
| 20:30:33 | pb_ | mallum[tv]: I don't have a serial cradle here and my dhcp server is acting up so it can't get on the network |
| 20:31:26 | mallum[tv] | pb_: np |
| 20:31:56 | mallum[tv] | pb_: I need to build tray.c as a little lib and drop it in anyway |
| 20:32:14 | pb_ | mallum[tv]: great. |
| 20:32:32 | mallum[tv] | pb_: probably do that next week now though |
| 20:32:54 | pb_ | argh, just realized that the subclassing idea doesn't work either :-( |
| 20:41:18 | pb_ | mallum[tv]: what it seems to come down to is, gtk_window_show() sets up various bits of internal state and a GtkWindow doesn't work properly unless you call that function. Unfortunately there is no way you can prevent it from also causing the window to be mapped. Worse, most of the actual work in there is done by other functions in gtkwindow.c that are static so you can't call them from a subclass. :-( |
| 20:41:53 | pb_ | oh well, maybe I'll stick with dlsym until we can switch to Gtk 2. |
| 20:42:36 | pb_ | moray: yeah, the appmgr screenshot looks a lot better now. |
| 20:44:24 | mallum[tv] | pb_: dlsym ? |
| 20:45:21 | mallum[tv] | pb_: I'll have another look with xscope or xmon and see if I can figure anymore out too |
| 20:45:30 | mallum[tv] | pb_: debugging X can be a real pain |
| 20:46:20 | pb_ | mallum[tv]: the current code in cvs uses dlsym to nobble gdk_map_window, it seems to work but it's as ugly as anything. |
| 20:48:06 | mallum[tv] | pb_: maybe we should hang on till we can switch to gtk2 |
| 20:48:36 | mallum[tv] | pb_: luckily most dockapps are simple enough to write in xlib ... |
| 20:48:50 | pb_ | mallum[tv]: yeah |
| 20:49:07 | pb_ | I might rewrite gpe-what in xlib, just so we have something that works for now. |
| 20:49:17 | mallum[tv] | pb_: I should try and build that libegg code and see if it works |
| 20:49:24 | pb_ | good plan |
| 20:49:33 | mallum[tv] | pb_: cool. lots of code in matchbox/util dir |
| 20:49:43 | mallum[tv] | pb_: and most of it quite readable |
| 20:49:49 | spung | is away: actually, I *am* here. I'm just not paying attention to you |
| 20:50:42 | mallum[tv] | pb_: if you get a chnace have a look at the xemned and tray specs at freedesktop.org, there may well have been something I've missed |
| 20:50:53 | pb_ | mallum[tv]: ok, will do |
| 20:51:11 | mallum[tv] | pb_: the old way of docking ( with the type_dock ) was simple but badly wrong |
| 20:51:24 | mallum[tv] | pb_: and only the wm could really run a dock |
| 20:51:42 | mallum[tv] | pb_: the tray specs and xembed allow you to get around that |
| 20:51:51 | pb_ | i see |
| 20:52:15 | mallum[tv] | pb_: type_dock means a window thats an actual dock ( not a docked app ) |
| 20:52:16 | pb_ | recompiles his kernel |
| 20:52:48 | mallum[tv] | pb_: 2.4.19 out ? |
| 20:53:10 | pb_ | mallum[tv]: no, just needed to change some config options :-/ |
| 20:53:36 | pb_ | kind of surprised 2.4.19 hasn't been released yet though, it's been a long time in the making |
| 20:53:38 | mallum[tv] | pb_: compiling kernels is fun :) |
| 20:53:49 | pb_ | mallum[tv]: I think it's overrated |
| 20:53:50 | mccarthy | pb_: install-sh is a symlink to /usr/share/automake/install-sh |
| 20:53:54 | mallum[tv] | pb_: I liked the way I first compiled an arm kernel and all the options changed :)) |
| 20:53:56 | pb_ | mccarthy: ah |
| 20:53:59 | mccarthy | pb_: so it seems that automake is not installed? |
| 20:54:34 | mccarthy | pb_: is there a way to let root login to gpe-login? |
| 20:54:45 | pb_ | mccarthy: latest gpe-login should allow root logins |
| 20:54:59 | pb_ | mccarthy: (might only be in CVS though, if so I'll make a new package in a moment) |
| 20:55:00 | mccarthy | pb_: latest ipk ot latest CVS? |
| 20:55:08 | mccarthy | ok. |
| 20:55:11 | pb_ | mccarthy: automake is installed now |
| 20:55:16 | mccarthy | pb_: thanks |
| 20:55:24 | mallum[tv] | spung: I just commited a new matchbox with the select menu fixed |
| 20:55:28 | pb_ | mccarthy: you can tell easily enough, if root doesn't show up on the drop-down menu then your gpe-login is too old :-) |
| 20:55:45 | mccarthy | pb_: no. I know mine is too old... |
| 20:56:12 | pb_ | mccarthy: looks like it is indeed in CVS only, building a new package now. |
| 20:56:12 | mccarthy | pb_: unless you uploaded a new ipk this afternoon (last few hours) then it isnt in unstable. |
| 20:56:14 | mallum[tv] | spung: monday I clean up the warning s. .... |
| 20:56:28 | mccarthy | pb_: *definitely* not in unstable |
| 20:57:01 | mallum[tv] | back to tv ..... |
| 20:57:18 | pb_ | mallum, wot you watching? |
| 20:58:39 | mccarthy | pb_: could you install libmad? |
| 20:59:12 | mccarthy | pb_: I think I also need the floating point ogg library, but I have no idea where that can be grabbed... :( |
| 20:59:19 | mccarthy | pb_: compiling gpe-media is a pain! |
| 20:59:28 | pb_ | mccarthy: agreed :-( |
| 20:59:38 | pb_ | mccarthy: I'll install those things in a minute. |
| 21:01:13 | pb_ | mccarthy: libmad is installed, and I just uploaded gpe-login 0.13 |
| 21:01:23 | mccarthy | pb_: great |
| 21:01:59 | mccarthy | pb_: id3tag.h: No such file or directory? wouldn't this be in libmad? (did you forget -devel?) |
| 21:02:07 | pb_ | mccarthy: there is a tarball of fixpvorbis in ~philb, suggest you just unpack it in your home directory and work from there. |
| 21:02:41 | mccarthy | pb_: cp: cannot open `/home/philb/fixpvorbis-0.1.tgz' for reading: Permission denied |
| 21:02:54 | pb_ | oops sorry, try again |
| 21:03:54 | pb_ | mccarthy: looks like id3tag.h is in libid3tag0-dev, installing that now |
| 21:04:31 | mccarthy | pb_: thanks... is mibus really expecting to have all this extra ipkgs required? |
| 21:04:48 | pb_ | mccarthy: dunno, you'd have to talk to mibus :-) |
| 21:05:01 | pb_ | mccarthy: the id3 thing is installed now, see if it fixes your problem |
| 21:07:51 | mccarthy | pb_: yep, but the vorbis stuff is a pain! |
| 21:09:29 | mccarthy | pb_: stupid question -> if I only have a libfixpvorbis.a how do I make something useful of it? (i.e., a .so?) |
| 21:10:43 | pb_ | mccarthy: gcc -shared -Wl,--whole-archive -o libvorbis.so libfixpvorbis.a |
| 21:11:23 | mccarthy | pb_: it didnt like *that* at all! |
| 21:11:34 | pb_ | mccarthy: what happened? |
| 21:11:50 | mccarthy | pb_: about a million things like "/usr/lib/gcc-lib/arm-linux/2.95.4/libgcc.a(tinfo2.o)(.bss+0x80): multiple definition of `signed char type_info node'" |
| 21:12:35 | mccarthy | pb_: if I ditch the --whole-archive bit, it doesn't complain at all.... |
| 21:13:08 | pb_ | mccarthy: right, but you probably end up with a very small library in that case |
| 21:13:37 | pb_ | mccarthy: try putting "-Wl,--no-whole-archive" right at the end of the command line, see if that sorts it out. |
| 21:14:09 | mccarthy | pb_: that worked |
| 21:14:15 | pb_ | ok, good |
| 21:15:00 | mccarthy | pb_: but the *linking* doesn't work... I wonder if he was using a different libfixvorbis version? |
| 21:15:08 | mccarthy | pb_: oh, well. I give up for today... |
| 21:15:17 | pb_ | mccarthy: as far as I know only one fixpvorbis was ever released |
| 21:15:30 | pb_ | mccarthy: and it did compile for me using that version at one point. |
| 21:17:01 | mccarthy | pb_: I think my .so is screwy... the linker claims that things are undefined, but they were in the fixvorbis *.c files... |
| 21:17:12 | pb_ | ah |
| 21:17:37 | pb_ | back later, dinner time |
| 21:17:43 | mccarthy | later all |
| 21:42:32 | spung | [21:59] <mallum[tv]> spung: I just commited a new matchbox with the select menu fixed |
| 21:42:35 | spung | mallum[tv]: hum? |
| 22:06:36 | dc[spiderman] | OMFG |
| 22:06:52 | dc_ | pb__: I advise you, not worth the money :/ |
| 22:07:03 | dc_ | spiderman kind of sucked. |
| 22:07:03 | dc_ | alot. |
| 22:07:07 | dc_ | american |
| 22:07:08 | dc_ | :/ |
| 22:07:15 | dc_ | but the cg was cool |
| 22:07:47 | dc_ | mallum[tv]: you watch eastenders, so you don't count ;P |
| 22:07:48 | dc_ | hehehe |
| 22:07:49 | dc_ | asnyway |
| 22:07:51 | dc_ | off to bed. |
| 22:08:06 | dc_ | it sucked. |
| 22:08:25 | dc_ | be up bright and early for some GPE hacking |
| 22:21:03 | mallum[tv] | dc ? |
| 22:21:16 | mallum[tv] | man kids of today |
| 22:21:18 | mallum[tv] | ;) |
| 22:22:50 | moray | yeah, going to bed early is ridiculous behaviour |
| 22:25:37 | mallum[tv] | moray: I meant the spiderman comments .... |
| 22:25:47 | mallum[tv] | moray: did he see the movie or the old tv show ? |
| 22:25:57 | moray | yes, he didn't seem too impressed, did he? |
| 22:26:19 | moray | mallum[tv]: well, unless his parents charge him to watch television I'd suspect the first... |
| 22:27:19 | mallum[tv] | heh |
| 22:27:50 | mallum[tv] | latest post on the familiar list - wtf is that all about ? |
| 22:28:01 | moray | hm, keylaunch is now working fine (and using a global config. file), but I can't be bothered to sort it out to install the config file / automatically get started... |
| 22:28:44 | pb__ | mallum[tv]: hmm, which post is that? |
| 22:29:17 | mallum[tv] | " having xkbd to stroke keys for me ..." |
| 22:29:25 | moray | oh, I had no idea really |
| 22:29:54 | moray | if he wants to remap keys, I don't know what that has to do with xkbd |
| 22:36:33 | moray | I was just wondering, if ipkg the program can happily install debs, what are the benefits of using the ipk format? |
| 22:36:45 | pb__ | moray: none |
| 22:36:57 | moray | that's what I was wondering, yes... |
| 22:37:00 | pb__ | moray: well, I suppose that's a bit harsh, it does have a smaller installed database format. |
| 22:37:36 | pb__ | moray: see the debian-arm mailing list archives for this month, there has been discussion of that very topic. |
| 22:37:50 | moray | I suppose there is a minor practical benefit in it being obvious that the ipk is the one with random stuff removed, but still... |
| 22:38:38 | mallum[tv] | ipkg can install debs right ? |
| 22:38:52 | pb__ | mallum[tv]: apparently, yes; I've never tried it though |
| 22:38:57 | moray | pb__: I read the thread up to a few days ago in the archive. I suppose I should resubscribe to the ARM lists again now, really, given that I'm actually doing stuff (stopped reading when I changed email addresses and wasn't doing anything relevant) |
| 22:39:18 | moray | pb__: I just used the existing debian/ setup to build and install keylaunch |
| 22:39:25 | pb__ | moray: cool |
| 22:40:12 | moray | it would seem quite nice in some ways if we did that elsewhere (as long as it didn't offend people too much...), but I'm not sure what the best thing is to do about producing reduced size debs - i.e. no man pages etc. |
| 22:40:25 | mallum[tv] | pb__: Im just wondering if ipkg could be fixed to use provides and then use it with udebs rather than ipkg, built from 'proper' debs |
| 22:40:30 | moray | we could always add an extra target in debian/rules, I suppose, as a short term thing |
| 22:40:51 | mallum[tv] | The big problem with ipkg seems no source packages and no quality control |
| 22:41:03 | mallum[tv] | and a lack of people to actually maintain the packages |
| 22:41:25 | pb__ | mallum[tv]: well, they're all three sides of the same (topographically implausible, I admit) coin. |
| 22:41:27 | moray | mallum[tv]: well, some of those are familiar problems, from familiar's point of view - but it's clear there's lots of duplicated work going on |
| 22:41:56 | pb__ | mallum[tv]: because there are no source packages, all the binaries are built by ad-hoc methods, and it's virtually impossible to reproduce a package that someone else made. |
| 22:42:38 | pb__ | mallum[tv]: which, in turn, is quite a large disincentive for people to actually pitch in and fix bugs. |
| 22:42:45 | mallum[tv] | pb__: exactly |
| 22:43:03 | moray | though of course with Debian, people still feel guilty about doing NMUs... |
| 22:43:20 | mallum[tv] | NMU's ? |
| 22:44:03 | pb__ | moray: well, yeah, there are social and/or political issues there, it's hard to get the balance right between people feeling too inhibited to make NMUs when they should, and people making random NMUs for packages they don't understand and introducing new bugs. |
| 22:44:20 | moray | mallum[tv]: non-maintainer upload of a (presumably bugfixed) package |
| 22:45:12 | moray | pb__: obviously, but it's a long way from the 'CVS spirit' of just fixing bugs when you see them - I suppose you could argue it's inevitable with the size of the project |
| 22:46:03 | moray | of course in principle you can just post the maintainer the patch, but many of them won't act on it terribly quickly |
| 22:46:18 | pb__ | moray: indeed |
| 22:47:18 | pb__ | I just created a new directory "gpe/games" in the CVS repository. You can probably guess what it's supposed to hold. :-) |
| 22:47:36 | moray | ah... |
| 22:47:38 | mallum[tv] | pb__: games ? |
| 22:47:51 | pb__ | mallum[tv]: you win. |
| 22:48:22 | mallum[tv] | :) |
| 22:48:53 | mallum[tv] | I have quake 4 ipkg'd |
| 22:49:02 | pb__ | :-) |
| 22:49:11 | moray | hm, is there something around that will convert RISC OS sprite files to a more useable format? |
| 22:49:24 | pb__ | moray: only on RISC OS I think |
| 22:49:51 | moray | pb__: right, I'll have to wait until I get that machine back from 'storage' then... |
| 22:50:42 | moray | or in principle, since I have a working arcem, I could feed them into it somehow and screengrab ;) |
| 22:50:47 | pb__ | heh :-) |
| 22:51:58 | mallum[tv] | pb__: did you used to hack on bbc's when your were a kid ? |
| 22:52:16 | pb__ | mallum[tv]: yep |
| 22:52:47 | mallum[tv] | pb__: then archemidees ? |
| 22:52:59 | pb__ | mallum[tv]: right again :-) |
| 22:53:17 | mallum[tv] | pb__: then risc pc ? |
| 22:53:36 | pb__ | mallum[tv]: ah no, I had discovered Linux by that time |
| 22:54:46 | pb__ | mallum[tv]: it was pretty hard to avoid Acorn machines in this city :-/ |
| 22:55:09 | mallum[tv] | pb__: yeah I can imagine |
| 22:55:24 | mallum[tv] | pb__: we had commodore pets at my school, then a load of beebs |
| 22:56:58 | pb__ | mallum[tv]: ah, I never used a PET. |
| 22:57:22 | pb__ | mallum[tv]: before my time I suppose :-) |
| 22:57:56 | mallum[tv] | pb__: this was my primary school |
| 22:58:15 | mallum[tv] | pb__: we got em pretty early on and were on 'tommorows world' |
| 22:58:24 | pb__ | mallum[tv]: wow :-) |
| 23:00:12 | mallum[tv] | pb__: the I got a vic20 at home |
| 23:00:37 | pb__ | mallum[tv]: ah, I used to have a C64 at home |
| 23:01:23 | mallum[tv] | pb__: c64 had the sprites :) |
| 23:02:19 | pb__ | right! :-) |
| 23:05:04 | mallum[tv] | hmmm, keithp has new xft stuff available that compiles seperatly to X and give better AA and _much_ lower memory usage |
| 23:05:10 | pb__ | hey, cool |
| 23:05:33 | mallum[tv] | I build them earlier no problem, but am yet to install on the ipaq |
| 23:05:48 | mallum[tv] | The xft stuff on the ipaq is very old |
| 23:06:44 | pb__ | mallum[tv]: just like the other x stuff then... |
| 23:08:55 | mallum[tv] | pb__: yep |
| 23:09:10 | mallum[tv] | pb__: jg has been ill and busy with hp stuff though ... |
| 23:09:23 | pb__ | mallum[tv]: right |
| 23:09:35 | pb__ | mallum[tv]: I was wondering why we hadn't heard much from Jim recently. |
| 23:12:39 | pb__ | heh, just noticed the skiffcluster webcam at http://www.handhelds.org/cam.html |
| 23:12:46 | pb__ | doesn't look like the most dynamic of scenes :-) |
| 23:16:01 | mallum[tv] | pb__: lets just hope they rearange the furniture soon ;) |
| 23:16:27 | pb__ | mallum[tv]: yup, I'm on the edge of my seat for the next instalment. |
| 23:20:00 | mallum[tv] | pb__: pure adranelin delivered through your browser every 10 minutes |
| 23:21:31 | pb__ | It's all too much, I'm going to bed. |
| 23:21:36 | pb__ | goodnight mallum[tv], moray |
| 23:22:06 | moray | goodnight pb |
| 23:23:17 | mallum[tv] | night |
