Previous day | Jump to hour: 01 02 03 04 05 06 07 08 09 10 11 12 13 14 15 16 17 18 19 20 21 22 23 | Next day

Seconds: Show Hide | Joins: Show Hide | View raw
Font: Serif Sans-Serif Monospace | Size: Small Medium Large

Click in the nick column to highlight everything a person has said.
The Logo icon identifies that the person is a core developer (has commit access).

Notice: Only Gecko based browsers prior to FF4 support the multipart/mixed "server push" method used by this log reader to auto-update. Since you do not appear to use such a browser, this page will simply show the current log, and not automatically update.

#rockbox log for 2011-06-10

00:00:21kugelthis is the main aspect of buflib, to reduce fragmentation
00:01:11sideralBecause (1) their allocation is mostly static / init-time only, (2) dealing with moving data is hard / complex for them, and (3) the benefits for the system are likely minor.
00:01:40sideralI think it's better to tolerate some fragmentation if at all possible
00:01:59kugelthe benefit doesn't depend on (for example) dircache, but what's hapenning around it
00:02:16kugelif something is freed before dircache, the hole wants to be filed
00:03:01 Join bug2000 [0] (~bug@unaffiliated/bug2000)
00:03:33sideralOne idea that crossed my mind: Would it make sense to simply hand out buffer space to clients in fixed-size chunks (regions, arenas) and let the clients do the entire high-level alloc management themselves, possibly with the help of a lib? That would always allocate/free big chunks (at the expense of some internal fragmentation), and no data would ever have to be moved around. Could the audio / codec code deal with multiple noncontiguous big ch
00:03:43pixelmasideral: I didn't see freqmod talking too often in the other channel (though that's a bit in this channel's nature) anyway but thought he would read it from time to time
00:03:55bug2000sideral: I found it! I think.
00:04:03pixelmaso I guess it needs a bit of patience
00:04:30sideralpixelma: OK, thanks for your efforts!
00:04:50kugelfreeing (always) creates holes, no matter how you do the allocation
00:05:12bug2000sideral: I think I know how to reproduce it.
00:05:49sideralbug2000: cool! shall I reopen the FS task?
00:06:02bug2000sideral: Ya.
00:06:18bug2000sideral: I believe you can reproduce it by deleting a file using Rockbox.
00:06:32bug2000Or at least thats how I think I've done it.
00:08:06sideralbug2000: OK, I've reopened FS #12129 <http://www.rockbox.org/tracker/task/12129>. I'm keen on seeing what you found! :)
00:08:07bluebrother-bothttp://www.rockbox.org/tracker/task/12129 3Duplicate database entries (bugs, unconfirmed)
00:08:30kugelsideral: I'm not sure what your suggestion achieves what a dumb malloc doesn't
00:10:08sideralkugel: The idea is that there wouldn't be any external fragmentation: One module cannot fragment the memory allocated to another
00:10:51kugelbut each free would leave a "big chunk" hole
00:11:44kugeland big chunks are almost always a memory waste
00:11:49bug2000kugel: Off the record, I love eating Kugel.
00:12:01sideralNo, each high-level free would leave a hole inside the module's arena. The the lowest level, the arena can only deallocated as a whole
00:13:18sideralYes, each module wastes whatever it cannot use of its chunk. But at the lowest level, there wouldn't be any fragmentation from small allocations
00:13:48siderals/The the/At the/
00:13:49 Quit bertrik (Ping timeout: 255 seconds)
00:13:51kugelfragmentation from small allocs is better than that from big chunks :)
00:14:25kugelno, we cannot afford wasting memory on big chunks, really
00:14:41sideralNot if you can tolerate noncontiguous big chunks in memory-hungry clients, which is my question actually
00:14:58kugelbuffering can't handle that, no
00:15:54sideralTo me it seemed like a smaller change than making dircache and tagcache fit for moving allocations
00:16:56bug2000sideral: Firefox found a time to die.
00:16:56sideralClients only needing small allocations could also share their arenas, reducing memory waste
00:17:12bug2000sideral: Basically, delete a file using Rockbox firmware. The duplication should happen if I got it right.
00:17:34sideralbug2000: Do you have the dircache enabled?
00:17:55kugelsideral: there's still not negible waste on big chucks. you know we have targets with less than 2MB memory (and 400k audio buffer)
00:17:58bug2000sideral: Where is that in the options?
00:18:08sideralbug2000: Is it the file that you deleted that will get duplicated?
00:18:20kugeland adding support to holes, which are even externally created, to buffering could easily double its code size
00:18:27bug2000sideral: It's EVERYTHING that gets duplicated.
00:19:01sideralbug2000: dircache is in general settings -> system -> disk
00:19:26kugelbtw, if anyone knows how to reduce this Makefile rules to one, I'm all ears: http://pastie.org/2045037
00:19:51sideralbug2000: can you make a backup of your database files for me, with a small example file set? Perhaps you can reproduce this with a simulator?
00:19:53 Quit ender` (Quit: Little expense had been spared to create the impression that no expense had been spared.)
00:20:35bug2000sideral: I see no "disk" there.
00:20:36sideralkugel: OK, it was only an idea. Moving allocations in a C program frighten me :)
00:20:56bug2000sideral: Sure, just link me to a download link for the linux sansa clipv1.
00:20:59kugelyou surely need to be careful
00:21:03bug2000I can't really firefox right now but I can wget.
00:21:39kugelthe ability for non-movable buffers is actually only meant for the audio buffer, not the other allocs
00:21:48sideralbug2000: Ah, the ClipV1 doesn't support dircache AFAIK
00:22:10bug2000sideral: Then no dir cache :P
00:23:11sideralbug2000: I can't cut you a sim build right now (need to leave in a few minutes). I'm not sure, perhaps we offer sim builds on our web site?
00:23:25bug2000sideral: Firefox is broken :P
00:24:13sideralbug2000: Use konqueror, links, or any other browser?
00:24:26bug2000Yey w3m :P
00:25:35bug2000bah, w32 file >.>
00:29:08 Join robin0800 [0] (~robin0800@149.254.61.232)
00:34:27 Quit liar__ (Ping timeout: 258 seconds)
00:37:57sideralkugel: Look up "pattern rules" in the GNU Make info file
00:37:58 Join Scromple [0] (~Simon@115-64-195-104.tpgi.com.au)
00:40:25bug2000Yey. Firefox is back.
00:42:00kugelsideral: I have such a rule
00:42:31kugelhttp://pastie.org/2045110 is the complete makefile
00:43:01kugelthe pattern rule seems to not work for test_*.c and main.c since they're always remade
00:43:31sideralkugel: That pattern rule is for compilation, not linking
00:44:05bug2000sideral: I fail to find the linux simulator.
00:44:08sideralI thought you wanted to remove the redundancy from the link rules?
00:44:44kugelsideral: I don't know how to setup a pattern rule for linking (the binaries have no extension) and it doesn't work for all C files
00:44:45sideralbug2000: Do you feel comfortable compiling your own? It's not difficult.
00:44:58bug2000sideral: Ya, I'm ok with compiling.
00:46:13sideralbug2000: http://www.rockbox.org/wiki/UiSimulator
00:47:46sideralkugel: The pattern would be like: % : %.o : $(LIB)
00:48:40sideraland then the command can refer to $@, $<, $^ in the usual fashion. And $* returns the string that was matched
00:49:58kugelI don't get how % works. what does it match?
00:50:34 Quit mudd1 (Quit: Ex-Chat)
00:50:42 Quit efyx (Remote host closed the connection)
00:50:53bug2000sideral: svn co sure takes time.
00:51:57 Quit domonoky (Read error: Connection reset by peer)
00:53:21 Join Epicide [0] (~joe@cpe-174-109-086-069.nc.res.rr.com)
00:54:47kugelalso makes implicit rules confuse me
00:55:03kugelit's making stuff with rules I never declared
00:55:26sideralbug2000: we also have a git mirror: http://www.rockbox.org/wiki/GitVersionControl#The_public_Rockbox_git_repository
00:56:21 Join mshathlonxp [0] (~athlonmpp@5ad7b04e.bb.sky.com)
00:56:23sideralkugel: Well, you seem to have specified at least some nonexisting target that Make thinks it can synthesize
00:57:32bug2000sideral: WAHH. It's REALLY late.
00:58:08sideralbug2000: Take your time. I need to go offline anyway shortly
01:00
01:00:28kugelsideral: :?
01:00:57kugelhow does % work? how does it know which .o files are meant with %.o?
01:02:32kugelah the all: $(TARGETS) confuses it
01:04:35 Join johnney5 [0] (~ident@pool-138-89-75-24.mad.east.verizon.net)
01:07:11sideralkugel: I don't quite remember and would have to RTFM. I think it deduces from the targets it needs to make −− or did you have to specify after the second colon (and dependencies belong into separate dependency rules)? It's late... :)
01:10:33kugelgot it down, I think
01:10:45kugelnow*
01:16:30 Quit sideral (Quit: Leaving.)
01:16:57 Quit HaimN (Ping timeout: 276 seconds)
01:17:29 Join HaimN [0] (~quassel@95.86.119.222)
01:22:39 Quit Curulan (Ping timeout: 240 seconds)
01:23:58 Quit simonlnu (Remote host closed the connection)
01:36:34 Quit robin0800 (Read error: Connection timed out)
01:36:44 Quit Keripo (Quit: Leaving.)
01:37:24 Join robin0800 [0] (~robin0800@149.254.60.163)
01:49:54 Join [Saint] [0] (~st.lasciv@124-197-3-117.callplus.net.nz)
01:57:05 Join mshaxp [0] (~athlonmpp@5ad7b04e.bb.sky.com)
01:58:39***Saving seen data "./dancer.seen"
01:59:00 Quit mshathlonxp (Disconnected by services)
01:59:05 Nick mshaxp is now known as mshathlonxp (~athlonmpp@5ad7b04e.bb.sky.com)
02:00
02:15:32 Join Keripo [0] (~Keripo@c-76-28-198-27.hsd1.wa.comcast.net)
02:18:43 Quit Epicide (Ping timeout: 240 seconds)
02:27:34 Join Epicide [0] (~joe@cpe-174-109-086-069.nc.res.rr.com)
02:36:14 Join simonlnu [0] (Oguy5z8jMt@unaffiliated/simonrvn)
02:40:00 Quit HaimN (Remote host closed the connection)
02:42:18 Quit keyb_gr (Ping timeout: 240 seconds)
02:42:50 Quit robin0800 (Ping timeout: 250 seconds)
02:53:46 Quit GeekShadow (Quit: The cake is a lie !)
03:00
03:06:58 Quit Epicide (Quit: leaving)
03:09:35 Quit johnney5 (Quit: —I-n-v-i-s-i-o-n— 3.2 (July '10))
03:24:21 Join evilnick [0] (~evilnick@ool-18bee3a9.dyn.optonline.net)
03:24:34 Quit evilnick (Changing host)
03:24:34 Join evilnick [0] (~evilnick@rockbox/staff/evilnick)
03:24:40 Nick jhMikeS is now known as MosquitoInMyRoom (~jethead71@rockbox/developer/jhMikeS)
03:24:50 Nick MosquitoInMyRoom is now known as jhMikeS (~jethead71@rockbox/developer/jhMikeS)
03:25:25 Nick jhMikeS is now known as StandingOnMyPant (~jethead71@rockbox/developer/jhMikeS)
03:26:16 Nick StandingOnMyPant is now known as jhMikeS (~jethead71@rockbox/developer/jhMikeS)
03:36:24jhMikeSDo we consider it kosher to delete posts with attachments someone says are "bad"?
03:40:03 Join Zarggg [0] (~zarggg@24.229.139.169.res-cmts.sm.ptd.net)
03:41:22JdGordonbad how? (in the forum you mean?)
03:43:10jhMikeSin flyspray
03:44:03jhMikeSsome bum patches posted in 12021 for the VGM codec
03:50:28JdGordonyeah, delete them
03:53:46 Join timccc [0] (~aoeu@112.166.15.141)
03:55:36 Part timccc
03:58:43***Saving seen data "./dancer.seen"
04:00
04:14:18 Join Fangs [0] (~Fangs@cpe-71-72-63-25.neo.res.rr.com)
04:15:12 Quit [Saint] (Quit: Imagination is for turbo-nerds who can't handle how kick-butt reality is. I'm a kick-butt reality master! I would rather die, than be imaginative. I mean that.)
04:16:43 Join [Saint] [0] (~st.lasciv@124-197-3-117.callplus.net.nz)
04:20:51FangsHey can someone tell me if buying a new ipod just to put linux on it is reasonable?
04:21:26FangsAlso is there full functionality with rockbox's linux like any other computer with linux on it?
04:21:44kisakumm ... while linux is simular to rockbox, it is not related to rockbox
04:22:18Fangsk
04:22:19Fangs:)
04:22:32 Quit [7] (Disconnected by services)
04:22:33[Saint]what is this "linux" of which you speak?
04:22:35JdGordonhttp://www.google.com.au/url?sa=t&source=web&cd=6&ved=0CFQQFjAF&url=http%3A%2F%2Fdaniel.haxx.se%2Fblog%2F2008%2F11%2F28%2Frockbox-is-not-linux%2F&ei=43_xTf3OCcup0AGehIm6BA&usg=AFQjCNHbHNDYyT0967lEnr7d28YdruniIA&sig2=rQf15O07Nrno9XBmIp1WqQ
04:22:38Fangsbut would I be able to screw around with kernels
04:22:39 Join TheSeven [0] (~TheSeven@rockbox/developer/TheSeven)
04:22:42JdGordondammit
04:22:42[Saint]Rockbox != linux
04:22:46JdGordonhttp://daniel.haxx.se/blog/2008/11/28/rockbox-is-not-linux
04:23:21kisakalso, as far as I know, ipodlinux has not been ported to anything newer than the ipod v5 (ipod video)
04:23:29Fangsoh
04:23:52[Saint]Rockbox is Rockbox, Linux is Linux....you're you, I'm me, etc. ;)
04:24:21Fangsbut it is italics right?
04:24:34Fangswell it sonds legal enough to me
04:25:09*[Saint] fails to parse the statement.
04:25:29Fangsat the same time I think I'm too fresh to do this solo
04:26:07[Saint]Yes, there is a functional Rockbox port for the iPod Classic, no, it's not Linux. The port is classed as "Unusable" at this stage, however, this does not mean that it's *actually* unusable.
04:26:30[Saint]It's very usable in fact, it just doesn't yet meet the criteria for "Unstable".
04:27:27Fangswell this seems useful. it's not exactly what I was thinking it was, but
04:27:32Fangsjailbreaking just seems so restrictive
04:27:38Fangsirony?
04:28:04kisakjailbreaking?
04:28:19kisakhow is jailbreaking related to rockbox?
04:28:51FangsRockbox is related to ipods
04:28:55[Saint]Presently, installation requires the use of a third party loader (seperate from Rockbox) called emCORE. The installation is straightforward, however there is no dual-boot, so you will lose the ability to use the Apple firmware (this can be undone, though), and installation also requires formatting the device completely so you'll lose anything that isn't backed up.
04:28:59Fangsolder ipods, but ipods none the less
04:29:07Fangso
04:29:27[Saint]Also, there's some issues with USB for some people, and battery life is somewhat reduced compared to the OF.
04:29:48Fangswell it'd be a bare device so Im not worried
04:31:00kisakFangs: jailbreaking is something you do with iOS devices, rockbox has not been ported to those devices
04:31:19[Saint]http://www.freemyipod.org/wiki/EmCORE_Installation <−− All you'll need to know re: installing emCORE & Rockbox on the iPod Classic
04:31:59Fangsalright Im thinking it over, yeah I was going to buy the newest. My ipod video has been smashed to bits and disassembled for parts
04:32:23Fangsnot so sure about the rockbox anymore
04:32:33Fangswhat else is rockbox
04:32:34Fangs?
04:32:51[Saint]What?
04:33:09[Saint]WHat on earth do you mean "what else is it"?
04:33:10kisakhttp://www.rockbox.org/wiki/WhyRockbox
04:33:13[Saint]*What
04:33:58Fangssince im here and Im not sure I want to buy an old ipod
04:34:13[Saint]There's quite literally no alternative....so, if you want to buy a Classic, and you don't want to use the original Apple firmware....Rockbox is the only choice.
04:34:25[Saint]The other choice is being happy with the Apple firmware.
04:35:05Fangswhich is cool in it's own right. Niche and all, looking through all the programs
04:35:19kisakalternatively I'd recommend a 4 GB Sansa Fuze and stick a 16GB microSD card in it, to total 20GB overall
04:36:17kisakyou can get the player for $35 or less usually
04:36:27 Join kugel_ [0] (~kugel@rockbox/developer/kugel)
04:36:38Fangscool cool, im using a 2nd gen nano with a cracked screen and a partially nonfunctional scroll wheel, so really anything would be an upgrade
04:36:45Fangsconsidering the options....
04:38:14Fangslooks nice and cheaper
04:39:41Fangsooh crossfader you guys are great :D
04:39:47kisakthe only thing about those Fuze models is that the screen cover is quite thin, it's worth the $4 to get a shell to stick the player in
04:40:14Fangsyeah I've got a problem with that
04:40:26 Quit kugel (Ping timeout: 276 seconds)
04:40:41Fangsfall once and it's like $300 gone in seconds
04:41:10kisakhow do you lose $300 on a $35 player?
04:41:42Fangsthe ipod videos weren't cheap
04:41:46[Saint]betting on it to win the Kentucky Derby? ;)
04:42:07Fangsat some point, even though it probably made for 30
04:42:09kisakalso, I've dropped my iPod classic a couple of times over the last couple years
04:42:21Fangsyeah but I had a phone in my pocket as well
04:42:25kisakthe hardcase has protected it well enough
04:42:37Fangsthe screen went crunch against the other screen
04:42:51 Quit pixelma (Disconnected by services)
04:42:53 Join pixelma_ [0] (quassel@rockbox/staff/pixelma)
04:42:55 Nick pixelma_ is now known as pixelma (quassel@rockbox/staff/pixelma)
04:44:40 Quit amiconn (Disconnected by services)
04:44:40 Join amiconn_ [0] (quassel@rockbox/developer/amiconn)
04:44:49 Nick amiconn_ is now known as amiconn (quassel@rockbox/developer/amiconn)
04:45:55FangsAlright well im off
04:46:05 Quit Fangs (Quit: Fangs)
04:46:20 Join buttsfredkin [0] (~47db2762@giant.haxx.se)
04:48:22 Quit buttsfredkin (Client Quit)
05:00
05:04:36 Join Rob2222 [0] (~Miranda@p4FFF197D.dip.t-dialin.net)
05:08:08 Quit Rob2223 (Ping timeout: 246 seconds)
05:46:55 Quit simonlnu (Quit: .)
05:58:45***Saving seen data "./dancer.seen"
06:00
06:22:02 Join Horscht [0] (~Horscht@p5DD56F03.dip.t-dialin.net)
06:22:03 Quit Horscht (Changing host)
06:22:03 Join Horscht [0] (~Horscht@xbmc/user/horscht)
06:24:08 Quit Horschti (Ping timeout: 252 seconds)
06:25:13 Join robin0800 [0] (~robin0800@149.254.61.35)
06:26:45JdGordon[Saint]: whats the latest on the theme?
06:26:56JdGordonI skimmed the logs but didnt see much happening
06:30:27[Saint]I've not had a lot of time to work on it presently.
06:30:48[Saint]I'm poking at a WIP presently that removes the tabs (but still retains the functionality)
06:31:13JdGordonchange the tabs to something that looks more "buttony" and should be fine
06:31:51 Quit kugel_ (Ping timeout: 255 seconds)
06:32:14[Saint]I'm removing them completely. Right now, the plan of action is to launch the context popup with the current track text, and the volume popup with the volume icon.
06:32:54[Saint]it'd be kinda handy if the touch area "none" respected */&....it seems it doesn't?
06:33:25[Saint]Oh....that's right, I also need to poke at the source to get %cs working.
06:33:53[Saint](well, it works...just the "playlist" and "ptichscreen" values don't display)
06:33:59 Join kugel [0] (~kugel@rockbox/developer/kugel)
06:35:16[Saint]I'll poke you as soon as I've got something I feel is worth testing.
06:35:45JdGordonits 4 lines of code :)
06:35:58JdGordonoh, you mean the skin?
06:36:16[Saint]Yeah...I know, I've just not had the time for much/any Rb hacking over the past week or so.
06:36:27[Saint]Real life keeps getting in the way :P
06:36:37JdGordondamn real life!
06:36:45[Saint]And yeah, I did mean the theme.
06:37:31 Quit robin0800 (Quit: Leaving)
06:38:36 Quit krazykit (Ping timeout: 268 seconds)
06:51:50 Part belak
06:59:33 Join Judas_PhD [0] (~kevin@misterfluffy.dsl.xmission.com)
07:00
07:20:53 Join chattr [0] (~mike@94.166.189.72.cfl.res.rr.com)
07:21:06 Part chattr ("gone")
07:29:08 Quit Judas_PhD (Quit: This is a quitting message)
07:37:18 Quit wtachi (Ping timeout: 258 seconds)
07:40:09 Quit scorche (Disconnected by services)
07:40:16 Join scorche` [0] (~scorche@rockbox/administrator/scorche)
07:40:52 Join Stummi [0] (~Stummi@rockbox/developer/Stummi)
07:52:08 Quit Stummi (Quit: Leaving...)
07:58:47***Saving seen data "./dancer.seen"
08:00
08:03:29 Join sideral [0] (~sideral@213.165.85.248)
08:03:29 Quit sideral (Changing host)
08:03:29 Join sideral [0] (~sideral@rockbox/developer/sideral)
08:04:17 Join liar [0] (~liar@clnet-p09-185.ikbnet.co.at)
08:13:37 Join MethoS- [0] (~clemens@134.102.106.250)
08:15:54 Quit n17ikh (Remote host closed the connection)
08:24:08 Join B4gder [0] (~daniel@rockbox/developer/bagder)
08:26:01 Join GodEater [0] (~quassel@rockbox/staff/GodEater)
08:31:04 Quit factor (Read error: Connection reset by peer)
08:37:12 Join factor [0] (~factor@74.197.205.204)
08:42:40 Join einhirn [0] (~Miranda@p54850E33.dip0.t-ipconnect.de)
08:42:46 Join ender` [0] (krneki@foo.eternallybored.org)
08:43:10 Quit einhirn (Client Quit)
08:47:46 Quit Scromple (Quit: Gone)
08:51:43 Join n17ikh [0] (~n17ikh@c-174-56-150-44.hsd1.sc.comcast.net)
09:00
09:08:28 Join bertrik [0] (~bertrik@rockbox/developer/bertrik)
09:12:30 Join Zagor [0] (~bjst@rockbox/developer/Zagor)
09:17:24 Join petur [0] (~petur@rockbox/developer/petur)
09:18:17 Join Highlander [0] (~Highlande@2a01:e35:2ec2:dcd0:224:1dff:fec0:aef0)
09:22:13 Quit kramer3d (Quit: Leaving)
09:28:02 Join tchan1 [0] (~tchan@c-69-243-144-187.hsd1.il.comcast.net)
09:28:35 Quit bertrik (Ping timeout: 260 seconds)
09:29:02 Quit sasquatch (Quit: WeeChat 0.3.2)
09:29:26 Join sasquatch [0] (~username@p4FF2DCB8.dip.t-dialin.net)
09:30:12 Quit tchan (Ping timeout: 250 seconds)
09:36:25 Quit Highlander (Quit: Quitte)
09:36:43 Join mem_ [0] (~mem@mem-irc.netnod.se)
09:48:00bug2000This is one annoying bug. I was able to reproduce it last night and now I can't.
09:49:10 Part mem_
09:58:50***Saving seen data "./dancer.seen"
10:00
10:01:06sideralI'm feeling with you, bug2000. I'm following several of those on our tracker :)
10:01:18 Join antil33t [0] (~antil33t@124-197-33-15.callplus.net.nz)
10:01:58bug2000sideral: It's just WEIRD.
10:02:04sideralOne suggestion: log your experiments (changes, actions, and observations) to a text file.
10:02:35[Saint]"Reason for closing: PICS OR IT DIDN'T HAPPEN!" ;)
10:03:07sideralbug2000: Perhaps it only triggers if you delete a file that was previously tracked by the database?
10:03:18[Saint]I'm sure that'd get rid of a lot of the open bugs on the tracker :P
10:04:07bug2000sideral: Pretty sure I did generate a new database for it to work with.
10:05:33sideralbug2000: If it's a Heisenbug, you must look away to allow it to happen :)
10:13:23[Saint]forum admins: tinajane wants a ban.
10:19:25 Join efyx [0] (~efyx@lap34-1-82-225-185-146.fbx.proxad.net)
10:21:33 Quit [Saint] (Quit: Imagination is for turbo-nerds who can't handle how kick-butt reality is. I'm a kick-butt reality master! I would rather die, than be imaginative. I mean that.)
10:23:00 Join pamaury [0] (~quassel@rockbox/developer/pamaury)
10:26:15 Join [Saint] [0] (~st.lasciv@124-197-3-117.callplus.net.nz)
10:42:21*GodEater prods @ torne with a stick
10:43:57 Quit Keripo (Quit: Leaving.)
11:00
11:17:36 Join simonlnu [0] (3n4xUanc0i@unaffiliated/simonrvn)
11:35:38 Quit beve (Quit: brb)
11:51:02 Join ChickeNES [0] (~ChickeNES@128.135.100.102)
11:55:55 Quit ChickeNES (Client Quit)
11:56:35TorneGodEater: hi
11:58:54***Saving seen data "./dancer.seen"
12:00
12:09:37 Join ChickeNES [0] (~ChickeNES@128.135.100.102)
12:12:19 Quit ChickeNES (Client Quit)
12:15:08GodEaterI've got the details of that VM
12:21:09 Join keyb_gr [0] (~chatzilla@p4FF0314F.dip.t-dialin.net)
12:28:25 Quit liar (Quit: hallowed are the ori!)
12:29:47 Join liar [0] (~liar@clnet-p09-185.ikbnet.co.at)
12:31:17 Join GeekShadow [0] (~Antoine@reactos/tester/GeekShadow)
13:00
13:15:35 Quit jfc (Ping timeout: 240 seconds)
13:22:30 Quit pamaury (Ping timeout: 252 seconds)
13:26:51bug2000Why are the list and home button on the fuze reversed?
13:28:32Tornereversed compared to what?
13:28:49Tornewe generally don't care about the original firmware's behaviour, and prefer to make button mappings mroe or less consistent across rockbox instead
13:45:49 Join pamaury [0] (~quassel@sphinx.lix.polytechnique.fr)
13:45:49 Quit pamaury (Changing host)
13:45:49 Join pamaury [0] (~quassel@rockbox/developer/pamaury)
13:55:32 Join jfc [0] (~john@static-72-73-80-12.ptldme.east.myfairpoint.net)
13:58:56***Saving seen data "./dancer.seen"
14:00
14:03:18 Quit antil33t ()
14:28:30 Quit markun (Read error: Connection reset by peer)
14:43:36 Join krazykit [0] (~krazykit@206.183.185.8)
14:53:35 Join benedikt93 [0] (~benedikt9@unaffiliated/benedikt93)
15:00
15:03:07 Quit liar (Ping timeout: 258 seconds)
15:09:49 Join evilnick_B [0] (0c140464@rockbox/staff/evilnick)
15:11:27TorneALL COMMITTERS: if you haven't put your email into the spreadsheet yet, please do so ASAP. If you don't know what I'm talking about, PM me for a link :)
15:28:10 Quit krazykit (Read error: Operation timed out)
15:29:51 Join domonoky [0] (~Domonoky@rockbox/developer/domonoky)
15:32:14 Quit bluebrother (Ping timeout: 258 seconds)
15:42:39 Join bluebrother [0] (~dom@g231121230.adsl.alicedsl.de)
15:42:40 Quit bluebrother (Changing host)
15:42:40 Join bluebrother [0] (~dom@rockbox/developer/bluebrother)
15:44:33GodEaterTorne: have you managed to get the import working at full speed now?
15:44:43 Join n1s [0] (~quassel@rockbox/developer/n1s)
15:44:56TorneGodEater: I have it working how we will want it to work, we just need the author mapping
15:45:01TorneSpeed is another matter
15:45:05GodEater\o/
15:45:07Tornebut i don't really care
15:45:11Tornemy server can just sit and churn
15:45:17GodEaterwell I was warned that box was slow
15:45:23TorneI'm not doing it on porpoise
15:45:26Tornei'm doing it on my server
15:45:31GodEateroh ok
15:45:34Tornei'll just send a tarball over when it's done
15:45:39Tornemy machien is way faster indeed :)
15:45:54GodEatercan we not give this task to preglow?
15:46:00Tornehm?
15:46:02GodEaterwho has the most ninja server evar?
15:46:08TorneIt won't help a lot
15:46:14TorneIt's not that slow.
15:46:23TorneI expect it will take well under 24 hours :)
15:46:28Torneso fuck it
15:46:38Tornei jsut need the author mapping, is the thing
15:46:45GodEaterwhere's the bottle neck then?
15:46:49Torneer
15:46:52Tornehashing stuff, i expect
15:47:04Torneso it's probably down to single-core performance
15:47:16Torneand my server has reasonably fast/modern cores
15:47:29Torneyou probably can't manage to go more than *twice* as quick, so who cares.
15:47:36GodEaterisn't it still running minecraft too ? ;)
15:47:43Tornei have multiple cores
15:47:46Torneminecraft can use the others ;)
15:47:53GodEaterlol
15:48:23Tornei'm looking at setting gerrit up on the machine while i wait to see if anyone else is going to fill in the trix
15:48:34Tornewe can decide what to do about people who haven't responded at some point
15:49:20 Quit mshathlonxp (Ping timeout: 240 seconds)
15:49:23Tornegah, why does the headless JRE still depend on cups and freetype
15:49:39kugelthe server of my lecturer is pretty fast, 8 core
15:49:42 Join mshathlonxp [0] (~athlonmpp@5ad7b04e.bb.sky.com)
15:49:57Tornegit doesn't parallelise much/atall
15:50:00Torneso hey.
15:52:37kugelZagor: did you do anything w.r.t to android in the build system since devcon?
15:52:51kugelstill getting the sdk version error
15:53:28 Join GodEater_ [0] (~quassel@rockbox/staff/GodEater)
15:54:07 Quit GodEater (Ping timeout: 250 seconds)
15:57:10 Join GodEater [0] (~quassel@64.9.146.129)
15:57:10 Quit GodEater (Changing host)
15:57:10 Join GodEater [0] (~quassel@rockbox/staff/GodEater)
15:57:58 Quit GodEater_ (Ping timeout: 252 seconds)
15:58:59***Saving seen data "./dancer.seen"
16:00
16:00:33 Quit GodEater (Read error: Connection reset by peer)
16:00:35Zagorkugel: you mean other than add it?
16:00:45Zagorhttp://build.rockbox.org/shownewlog.cgi?rev=29991;type=android480x800
16:01:04kugelhm
16:01:11kugelwhy do I still get the sdk error?
16:01:34Zagorare you sure your build client has the path set up correctly?
16:01:45kugelyes
16:02:53 Join GodEater [0] (~quassel@rockbox/staff/GodEater)
16:03:09kugelZagor: http://pastie.org/2047785
16:04:17Zagoryes, but does your runclient.sh not set the path?
16:04:42kugelwhy should it? it's exported
16:06:00Zagorjust checking
16:06:51 Join GodEater_ [0] (~quassel@rockbox/staff/GodEater)
16:07:19 Quit GodEater (Ping timeout: 240 seconds)
16:07:31 Join wtachi [0] (~wtachi@cpe-065-190-012-236.nc.res.rr.com)
16:10:05kugelZagor: did it work at the devcon already? rbclient.pl wasn't updated since then
16:10:25Zagoryes it did
16:11:29kugelI told you about that error at the devcon already
16:12:25Zagorwell, it's difficult to debug a problem I cannot reproduce
16:13:52kugelZagor: adding the tools folder to PATH helped
16:14:11Zagorgreat
16:14:44kugelI expected ANDROID_{S,N}DK_PATH to be sufficient, though, as those are needed for building
16:20:41 Quit domonoky (Read error: Connection reset by peer)
16:24:46 Join Judas_PhD [0] (~kevin@misterfluffy.dsl.xmission.com)
16:26:17 Join edboyer93 [0] (~eric@pool-72-94-145-226.phlapa.fios.verizon.net)
16:30:38 Quit Judas_PhD (Remote host closed the connection)
16:31:04 Join Judas_PhD [0] (~kevin@misterfluffy.dsl.xmission.com)
16:33:41 Quit B4gder (Quit: Konversation terminated!)
16:47:18 Quit Judas_PhD (Quit: This is a quitting message)
16:48:05 Join krazykit [0] (~krazykit@206.183.182.189)
16:58:22 Join Buschel [0] (~c2795aa3@giant.haxx.se)
17:00
17:00:51 Quit Buschel (Client Quit)
17:02:48 Part Zagor
17:03:35 Nick kugel is now known as kugelp (~kugel@rockbox/developer/kugel)
17:05:59kugelpgevaerts: I mostly finished implementing the buflib enhancements
17:06:28 Quit bluebrother (Disconnected by services)
17:06:29 Join bluebroth3r [0] (~dom@rockbox/developer/bluebrother)
17:07:51 Quit sideral (Quit: Leaving.)
17:09:29 Quit bluebrother-bot (Ping timeout: 255 seconds)
17:09:47 Quit petur (Quit: *plop*)
17:20:17 Join T44 [0] (~Topy44@f049181145.adsl.alicedsl.de)
17:20:18 Join bluebrother-bot [0] (~bluebroth@g231121006.adsl.alicedsl.de)
17:20:39 Quit n1s (Ping timeout: 240 seconds)
17:23:26 Quit Topy (Ping timeout: 255 seconds)
17:34:48*[Saint] wonders about RaaA and the build system...
17:34:53[Saint]How is signing handled?
17:35:11TorneIt isn't
17:35:21TorneThat's one ofd the reasons it's not being built successfully there yet
17:35:29[Saint]Aha.
17:35:31TorneWe discussed it at devcon and have some ideas.
17:36:17[Saint]basically all that can be done is build up to point of signing, then pass the build back to a trusted central machine....or have just one trusted machine handle all builds...no?
17:36:42 Join ChickeNES [0] (~ChickeNES@128.135.100.102)
17:36:49[Saint]otherwise you'd need to uninstall each time, and, that'd suck.
17:37:23Tornewe're probably going to end up actually authenticating clients somehow
17:37:26Torneand choosing which ones to trust
17:37:30 Join user890104 [0] (~Venci@6bez10.info)
17:37:37GodEater_and then sign builds from them
17:37:42Torneit's not okay to just let clients build unsigned binaries and have the server sign it because anyone can connect as a client and upload anything
17:37:46GodEater_from what I recall of the talk
17:38:03*[Saint] nods.
17:38:04TorneYeah, so we'll either have the trusted clients sign it with their own keys and then the master can resign them
17:38:15Torneor we will just authenticate the slave/master connection
17:38:18[Saint]I've been wondering how the signing would work with the build system for some time now.
17:38:30Torneprobably the trusted clients will be "ones run by committers"
17:38:39GodEater_I'm trying to work out which is the easier to implement
17:39:12GodEater_I think signing the builds from each trusted client is probably easier
17:39:21Torneprobably
17:39:21GodEater_rather than authenticating the connection
17:39:29[Saint]multiple machines+resigning would be faster...but one trusted machine "easier"..I assume?
17:40:05 Quit pamaury (Remote host closed the connection)
17:40:09GodEater_Torne: the alternative is we send out the Rockbox master key to all trusted clients?
17:40:27Tornei'd vote for not doing that on general principle unless it proves to be really hard to resign
17:40:30GodEater_though I guess that risks the entire system in the event of a client compromise
17:40:41Torneyeah. revocation.
17:40:43[Saint]indeed.
17:41:02[Saint]Not handing out the key should be avoided, if possible.
17:41:17GodEater_nice double negative there [Saint]
17:42:23kugelp(re)signing trusted builds on the master very doable doesn't it?
17:42:48Tornekugelp: should be,m unless the android tools are terrible
17:43:22kugelpTorne: I think you can just repack the apk and sign again
17:43:38Torneis there some easy way to verify the signature, though?
17:44:16kugelpwhatever jarsigner can do
17:44:57 Quit advcomp2019__ (Read error: Connection reset by peer)
17:45:24 Join advcomp2019__ [0] (~advcomp20@unaffiliated/advcomp2019)
17:56:23 Quit GodEater_ (Ping timeout: 252 seconds)
17:59:03***Saving seen data "./dancer.seen"
18:00
18:21:18 Quit edboyer93 (Quit: Leaving)
18:22:56Torneok, so i'm not giving up on more people filling in email addresses int he list *yet*, but eventually we will have to as some people are no longer around. What should we do with people who haven't been active for a long time but are still in the commit history? Guess an email for them from committers subscriptions? Use a @rockbox.org address? Use something obviously fake?
18:29:10 Join thomasjfox [0] (~thomasjfo@rockbox/developer/thomasjfox)
18:29:56CIA-58New commit by 03thomasjfox (r29992): RaaA / sim: Don't abort directory read if we encounter files larger than 2 GB in a directory
18:33:06CIA-58r29992 build result: All green
18:36:16*Torne builds a list of everyone who hasn't committed recently for comparison
18:39:51Torneheh, there's 53 usernames that haven't committed since before 2008
18:40:45 Quit benedikt93 (Ping timeout: 252 seconds)
18:41:44 Join hobby16 [0] (~ubuntu@162.123.68.86.rev.sfr.net)
18:42:15hobby16hello
18:43:24 Join robin0800 [0] (~robin0800@cpc3-brig8-0-0-cust703.3-3.cable.virginmedia.com)
18:49:16[Saint]Hi hobby16, welcome to the channel. This channel is specific to Rockbox, if you have a question regarding ROckbox just ask away. If you're looking for a social channel, this isn't the right place ;)
18:51:34linuxstbTorne: Can email addresses be "corrected" at a later date if old committers come forward?
18:52:23TorneNot once we publish the repository and people start using it
18:52:34 Join GodEater [0] (~quassel@5ad8c387.bb.sky.com)
18:52:34 Quit GodEater (Changing host)
18:52:34 Join GodEater [0] (~quassel@rockbox/staff/GodEater)
18:52:35TorneYou don't have to use the same email forever, so if they come back and start committing again they can use whatever they like
18:52:35linuxstbTorne: Also, I assume you've noticed some users are there twice? e.g. I have been both linuxstb and dave, Bagder has also been dast
18:52:47thomasjfoxlinuxstb: That would mess up the commit history -> new SHA1 ids
18:52:50Tornebut we can't change the address in old commits without destroying the history
18:53:06TorneThis is why the conversion I'm going to do is going to break existing git repos cloned from our git mirror ;)
18:53:14Tornebecause that mirror doesn't remap usernames at all and just uses useless strings
18:53:37Tornelinuxstb: so yeah. I can, technically, replace the addresses while i'm still testing it
18:53:44Tornefilter-branch is faster than doing the whole import/conversion again
18:53:49Tornebut once we publish it you can't do that any more
18:54:03Torneand yes, i know some people are in there multiple times
18:54:06Tornethat's fine
18:54:14Tornethose will just get folded together
18:54:23Tornezagor is even more people since he is also root and (no author) :)
18:55:52Tornebut anyway. yeah. we will have to decide what is going to be recorded as the "official" history
18:56:20Torneand the ability to change our minds drops off extremely rapidly once we are using git publicly :0
18:58:18 Join GodEater_ [0] (~quassel@027e17f7.bb.sky.com)
18:58:19 Quit GodEater_ (Changing host)
18:58:19 Join GodEater_ [0] (~quassel@rockbox/staff/GodEater)
18:59:31Tornei poked more stats btw and about 50% of the usernames on the list have committed in the last 18 months
18:59:50Torneso the other 50% may well not be paying much attention :)
19:00
19:01:00 Quit GodEater (Ping timeout: 244 seconds)
19:02:15 Quit thomasjfox (Remote host closed the connection)
19:02:25 Quit GodEater_ (Client Quit)
19:02:42 Join GodEater [0] (~quassel@rockbox/staff/GodEater)
19:04:06 Join pamaury [0] (~quassel@vit94-1-82-67-248-70.fbx.proxad.net)
19:04:06 Quit pamaury (Changing host)
19:04:06 Join pamaury [0] (~quassel@rockbox/developer/pamaury)
19:07:23 Quit hobby16 (Ping timeout: 250 seconds)
19:12:21amiconnTorne: We might be able to get some committer->email mappings from the ml archives.
19:12:32amiconnThe question is whether we want that
19:13:36 Join GodEater_ [0] (~bibble@027e17f7.bb.sky.com)
19:13:36 Quit GodEater_ (Changing host)
19:13:36 Join GodEater_ [0] (~bibble@rockbox/staff/GodEater)
19:15:46 Join promyloph [0] (~foo@unaffiliated/promyloph)
19:23:41 Join Buschel [0] (~chatzilla@p54A3B136.dip.t-dialin.net)
19:37:51 Join bertrik [0] (~bertrik@ip117-49-211-87.adsl2.static.versatel.nl)
19:37:51 Quit bertrik (Changing host)
19:37:51 Join bertrik [0] (~bertrik@rockbox/developer/bertrik)
19:40:14 Join xeniter [0] (~xeniter@194.48.133.8)
19:53:41 Join webguest28 [0] (~5039e7b1@giant.haxx.se)
19:53:52 Quit webguest28 (Client Quit)
19:54:11 Quit GodEater (Remote host closed the connection)
19:59:05***Saving seen data "./dancer.seen"
20:00
20:08:59 Join TheLemonMan [0] (~lem0n@ppp-36-20.26-151.libero.it)
20:09:41 Quit swilde (Remote host closed the connection)
20:16:37 Join sideral [0] (~sideral@213.165.85.248)
20:16:37 Quit sideral (Changing host)
20:16:37 Join sideral [0] (~sideral@rockbox/developer/sideral)
20:28:48 Join wodz [0] (~wodz@87-206-240-131.dynamic.chello.pl)
20:30:34Torneamiconn: indeed
20:30:34Torneamiconn: i'm not sure i want to just fill it in for people
20:30:49wodzTorne: You refered to some spreadsheet earlier today - what's that?
20:31:06Tornewodz: i sent a mail to -committers
20:31:24Tornewe're collecting email addresses for the git migration, for what you want to be associated with your commits
20:31:38Tornesince svn uses just a username wheraesa git uses realname <email>
20:32:00wodzhmm, I am not subscribed to -commiters that explain why I missed the context.
20:32:05TorneYou should be
20:32:11Tornebut yeah i just pm'ed you the link
20:32:24Tornefill in the email address you want me to use for the git migration, next to your username
20:32:34Torneand correct your real name if it's spelt wrong or not formatted right ;)
20:33:36amiconnWhat are you going to do if committers have more than one username? Will you merge them?
20:33:56Torneyes
20:33:59Tornethey dont' need to be unique
20:34:08Tornethere's no actual "account" or anything here
20:34:24Torneit's literally just throwing the svn usernames away and replacing them with an arbitrary string formatted as something that includes an email address.
20:34:28Torneso it doesn't have to be 1:1 :)
20:34:50amiconnSo how does one authenticate?
20:34:59Torneunrelated to this.
20:35:07Torneif we go with gerrit, you authenticate with an ssh pubkey
20:35:10Torneor a generated password
20:35:15Torneand whatever username you choose on gerrit
20:35:18Tornewhich is up to you
20:35:24Torneit's all self-service at that end
20:35:33TorneThis is just for how to identify authorship for existing commits in history
20:35:37Torneit has nothing to do with access control
20:36:18Torneif we don't go with gerrit then it may be slightly differnt but generally ssh public key auth is how it's done with git
20:41:43TorneI'll have a gerrit demo running sometime over the weekend, and will put up some instructions on how to get set up and try it out, so you can see what you think.
20:41:43Tornewe have a machine to host it that isn't in my living room now, so it's got enough bandwidth to have a full copy of the rockbox tree ;)
20:41:43wodzhmm is '+' sign valid in email address?
20:41:43Tornewodz: yes
20:41:43wodzweird
20:41:43DBUGEnqueued KICK Torne
20:41:43Tornelots of things don't like it, but it is valid by the RFCs
20:41:43Tornewodz: if you quote it almost everything is valid
20:41:43***Alert Mode level 1
20:41:43Torne"Hi There I'm Some Guy@231q535%%3"@foo.com is a valid email address
20:41:43jhMikeSTorne: you sent something to committers about the spreadsheet? I'm not seeing it :\
20:41:43***Alert Mode level 2
20:41:43TornejhMikeS: yesterday? day before?
20:41:43***Alert Mode level 3
20:41:43Tornespawned a 20+ post long argument about real names?
20:41:43***Alert Mode level 4
20:41:43Torne:)
20:41:43jhMikeSoh
20:41:43jhMikeSwaaaay back there ... lol
20:41:43***Alert Mode level 5
20:41:43Torneamiconn: i'm tempted to go with username@rockbox.org for the people who haven't replied
20:41:43***Alert Mode level 6
20:41:43Torneafter some deadline
20:41:43DBUGEnqueued KICK jhMikeS
20:41:43jhMikeSI'm not sure if bagder got my request to setup a rockbox.org address
20:41:43***Alert Mode level 7
20:41:43TornejhMikeS: well you can put it in now anyway
20:41:43***Alert Mode level 8
20:41:43Torneit's not going to receive any mail any time soon
20:41:43***Alert Mode level 9
20:41:43Tornewe are unliekly to actually switch for at least a week or two :)
20:41:43***Alert Mode level 10
20:41:43Torneamiconn: we can redirect them to some holding location or something until people come back to "claim" their mail, I guess?
20:41:43***Alert Mode level 11
20:41:43Torneamiconn: seems like a better option than guessing and publishing an address they don't want published.
20:41:43bertrikmaybe wait at least until release
20:41:43***Alert Mode level 12
20:41:43Tornebertrik: yes, probably
20:41:43gevaertsYes, it does seem wise to wait ubtil the release
20:41:56TorneI can still bring the git infrastructure up while we are still using svn, though
20:42:08Torneit's not a problem to import new commits from svn as long as nobody is committing real work to git yet
20:42:18Tornewe can have it up to look at and experiment with
20:43:05Torneit's not any different to the current git mirror in principle, it's just this version will have proper contributor names/addresses instead of dummy ones
20:43:12Torneand the branch/tag history will be slighly different
20:43:22wodzTorne: we need to sponsor you t-shirt with GO GIT or something :-)
20:43:22Torne(i am preserving all branches and tags because it's easy and doesn't really make things any bigger) :)
20:43:31Tornewodz: you realise i still don't actually like git
20:43:35Torneand would rather use something else ;)
20:43:45Torneit's ust that if we are going to use git i at least want to see it done properly ;)
20:44:02jhMikeSdone
20:44:20Tornethanks.
20:44:28TorneMost people who have committed in the last year or so have done it
20:44:36Tornethe majority of the blanks are from people who aren't really around any more
20:44:48Torneor at least not actually submitting to svn
20:46:49 Quit [Saint] (Quit: Imagination is for turbo-nerds who can't handle how kick-butt reality is. I'm a kick-butt reality master! I would rather die, than be imaginative. I mean that.)
20:47:12wodzabout updating wiki - TowerOfRockbox needs update with new record
20:48:43Torneheh
20:49:05jhMikeSI have a small helicopter I could fly into it
20:49:42gevaertsSpeaking of which, what's the status of the devcon minutes and the google blog?
20:50:00TorneOh crap yes
20:50:04TorneSomeone still needs to write that
20:50:04Torne:)
20:50:16Tornewould someone like to write that up this weekend? :) :) :)
20:50:18TorneI suck at writing
20:50:25Torneand maybe pick some photos
20:50:31Torne(i don't have any photos)
20:51:00gevaertsI half understood that GodEater was doing somethibg, but I'm not sure
20:51:44***Alert Mode OFF
20:51:50Tornehe implied he might
20:51:51Torneanyway if we can get that done say, within the next week, that'd be great
20:51:51Torneotherwise cat will probably bug me
20:52:46 Join benedikt93 [0] (~benedikt9@unaffiliated/benedikt93)
20:52:46CtcpIgnored 2 channel CTCP requests in 0 seconds at the last flood
20:52:46*Torne heads home, will be back later to check if any of the people who haven't filled in the list are actually on irc ;)
20:53:40 Join liar [0] (~liar@clnet-p09-185.ikbnet.co.at)
20:55:05jhMikeSLinusN surely must be involved here :)
20:56:03 Quit xeniter (Ping timeout: 276 seconds)
21:00
21:02:02kugelpwe definitely owe Torne some beer for doing this git transition work
21:04:04 Join {phoenix} [0] (~dirk@p5DF2DDF4.dip.t-dialin.net)
21:05:56 Nick tchan1 is now known as tchan (~tchan@c-69-243-144-187.hsd1.il.comcast.net)
21:06:01 Quit tchan (Changing host)
21:06:01 Join tchan [0] (~tchan@lunar-linux/developer/tchan)
21:39:51kugelpI have one question w.r.t. to buflib_alloc_maximum()
21:40:18kugelpbuflib has the handle table at the end of its buffer. for new allocations the handle table needs to be enlarged
21:40:38kugelpwhich might fail if buflib_alloc_maximum() handed out memory there
21:40:58gevaertsCan that ever actually happen?
21:41:24kugelpso, I'm thinking how to solve this. a) reserve sufficiently enough handles in _maximum(), b) ask the allocation to make space at the end, or c) both
21:42:20kugelpit's unlikely to happen since de-allocated handles are reused before enlarging the table
21:42:23gevaertsI mean, if you don't shrink after calling buflib_alloc_maximum(), something is wrong
21:42:43 Join [Saint] [0] (~st.lasciv@124-197-3-117.callplus.net.nz)
21:43:07kugelpthe shrink can possibly free space at the front and not at the end of the allocation (i.e. audio buffer)
21:43:14gevaertsoh
21:43:16gevaertsright
21:44:19kugelpalso the user code ideally shouldn't be concerned with the handle table at the end (or even know it exists)
21:46:03gevaertsMaybe add a rule that says that non-movable allocations can only shrink on one end? Would that hurt things?
21:46:17kugelpI have implemented b) currently, a) is easy to implement additionally so we'd have c)
21:46:19 Quit [Saint] (Client Quit)
21:46:23 Join realodd [0] (~4f759869@giant.haxx.se)
21:46:59gevaertsWell, it depends on what "sufficient" means
21:47:43 Quit rasher (Quit: Changing server)
21:47:49kugelpperhaps 64?
21:47:51gevaertsSo I'd say b, and possibly c, and insist that if you want a buffer at a fixed address, you can only shrink while keeping the start address the same
21:48:26gevaertsIf you do that, the a) bit is only an optimisation and it's not critical if you get the number a bit too small
21:48:43kugelpthat breaks my plans to resize the audio buffer at its front, having the pcm buffer (accessed by dma) at the end
21:48:47gevaertsBut I don't know if that rule is realistic given the code
21:48:48 Join rasher [0] (~rasher@rockbox/developer/rasher)
21:49:19kugelpso, effectively audio would have to stop to make room at the end
21:49:23gevaertsok
21:49:39gevaertsThat's not really very nice, so let's try to avoid it :)
21:51:24kugelp64 is only 256 bytes, and it's *really* unlikely there are going to be 64 simultanous allocations
21:51:51gevaertsSo c) would mean you only move when you happen to get the number wrong I guess, so you only may have to restart playback in rare cases?
21:52:01 Quit promyloph (Quit: WeeChat 0.3.5)
21:52:04kugelpif at all, yes
21:52:12gevaertsright, then I think that's best
21:52:39 Quit mshathlonxp (Ping timeout: 240 seconds)
21:52:45kugelpalright then
21:52:56 Quit mikroflops (Ping timeout: 246 seconds)
21:53:08kugelpbtw, I've implemented the API proposed on the ML now (with very minor changes to it)
21:54:20gevaertsWith the extra call to get a pointer? ;)
21:55:04kugelpyes :)
21:55:23kugelp(I've pushed everything to my github)
21:57:35 Join _jhMikeS_ [0] (~jethead71@c-68-61-166-99.hsd1.mi.comcast.net)
21:57:35 Quit _jhMikeS_ (Changing host)
21:57:35 Join _jhMikeS_ [0] (~jethead71@rockbox/developer/jhMikeS)
21:57:35 Quit jhMikeS (Disconnected by services)
21:57:36 Nick _jhMikeS_ is now known as jhMikeS (~jethead71@rockbox/developer/jhMikeS)
21:58:41realoddhey anybody here?
21:59:01 Quit realodd (Quit: CGI:IRC)
21:59:08***Saving seen data "./dancer.seen"
21:59:10 Join realodd [0] (~4f759869@giant.haxx.se)
21:59:59jhMikeSnobody but us chickens
22:00
22:01:21GodEater_Torne: I'll try and write up something cat-worthy over the weekend
22:01:23realoddi bet there's chicks out there too
22:01:27realoddanyhow
22:01:53realoddcan anyone help me make my ipod recognizable as a midi controller?
22:02:07realoddor help me find some music making software?
22:02:10realodd:)
22:02:43gevaertsThat's not going to be straightforward
22:03:33gevaertsI *think* midi over USB uses the audio class. There are patches for USB audio around, but they're not finished or integrated, and they don't do midi yet
22:05:04pamauryand it mainly depends on the usb controller
22:05:27kugelpgevaerts: so, I would like to have decisions regarding the API relatively soon so I can start porting it over to rockbox
22:06:16gevaertskugelp: what decisions still need to be taken? The naming?
22:06:17kugelpin the meantime I will add more test cases, but it's not simple to make up sufficiently complex scenarios
22:06:42kugelpgevaerts: and (more importantly) whether the API is good enough
22:06:56kugelpI thought it would be more controverse but it doesn't appear to be
22:07:20GodEater_kugelp: it's not contraversial because you're still not calling it malloc() ;)
22:07:33kugelpactually...:)
22:07:42*GodEater_ gets his pitchfork
22:07:46gevaertskugelp: you proposed it on the mailing list, so give it a few more days and it's fine :)
22:07:53kugelpactually, I thought on RaaA it might be malloc backed
22:08:03GodEater_it's acceptable there
22:08:19GodEater_those targets have bucketloads of memory compared to the usual targets
22:08:31jhMikeSGodEater_: lol...my sentiments exactly
22:08:57GodEater_it's not like we're looking after the memory there either really - there's a "real" OS to do it for us.
22:08:58gevaertsOn RaaA you're also likely not going to bother moving allocations around
22:09:04gevaertsThey do have an MMU after all
22:09:05jhMikeSsneaking malloc in the back door
22:09:12GodEater_jhMikeS: Sssssssh!
22:09:18kugelpmalloc doesn't mean memory waste :)
22:09:18*GodEater_ looks nervously over his shoulder
22:10:01jhMikeS;)
22:10:20kugelpthe amount of available memory isn't an argument for or against malloc. whether you can sensibly free() is one
22:11:46kugelpjhMikeS: too late, it's already sneaked in ;)
22:12:39jhMikeSthere are times I wouldn't mind it at all
22:15:20 Join [Saint] [0] (~st.lasciv@124-197-3-117.callplus.net.nz)
22:20:18 Quit TheLemonMan (Quit: Ex-Chat)
22:20:46 Join stoffel [0] (~quassel@p57B4B953.dip.t-dialin.net)
22:25:42 Join mshathlonxp [0] (~athlonmpp@5ad7b04e.bb.sky.com)
22:26:01realoddyeah i know it's tricky
22:26:07realoddbut where should i go
22:26:09realodd?
22:26:38realoddwhat other devices does rockbox emulate besides snes and that computer that runs on cassettes?
22:26:43 Join Noub [0] (~8e44613a@giant.haxx.se)
22:28:50Noubas opposed to random songs)?
22:29:05Nouband that ate the first half of my question, nice
22:29:32Noubcan rockbox shuffle albums as opposed to songs?
22:30:06gevaertsIt can shuffle directories.
22:30:09 Quit wodz (Quit: Leaving)
22:30:33Noubthat'd be the same as shuffling artists then?
22:30:50 Join hobby16 [0] (~ubuntu@162.123.68.86.rev.sfr.net)
22:31:01gevaertsWell, that depends entirely on your directory structure
22:31:05NoubI guess it would depend how the songs were transferred over, huh
22:31:08gevaertsHave a look at http://download.rockbox.org/daily/manual/rockbox-ipodvideo/rockbox-buildch11.html#x14-30700011.4.13
22:31:23 Quit [Saint] (Quit: Imagination is for turbo-nerds who can't handle how kick-butt reality is. I'm a kick-butt reality master! I would rather die, than be imaginative. I mean that.)
22:31:24Noubrhythmbox seems to do it by artist, I wonder if I can change that
22:34:15 Join evilnick_B_ [0] (0c140464@rockbox/staff/evilnick)
22:34:29Tornemost programs that use a file structure at all do artist\album
22:34:31Tornewhich would work fine
22:35:11 Quit evilnick_B (Ping timeout: 252 seconds)
22:35:18 Nick evilnick_B_ is now known as evilnick_B (0c140464@rockbox/staff/evilnick)
22:36:42Noubalbums with various artists would cause headaches though
22:36:54 Join mikroflops [0] (~yogurt@h-34-156.A238.priv.bahnhof.se)
22:37:05Tornemost programs that arrange files sensibly put those under Various Artists\album, or under the albumartist
22:37:08Torne:)
22:39:22 Quit benedikt93 (Quit: "There are three kinds of lies: lies, damned lies, and statistics." - Benjamin Disraeli)
22:41:04Tornedionoea: see my mail on -committers about email addresses for git migration
22:42:17dionoeaTorne: ?
22:42:37Tornetrying to collect committer names/addresses for svn2git purposes
22:42:44Torneyou haven't filled it in :)
22:43:04dionoeaoh. I'm not sure that I got that e-mail. Let me check
22:44:17TorneUnhelpful: also you
22:44:48Tornealso other people but they are not ehre using the same nick as their svn username
22:44:51Torne:)
22:46:10 Join kramer3d [0] (~kramer@ip98-169-194-252.dc.dc.cox.net)
22:46:10 Quit kramer3d (Changing host)
22:46:10 Join kramer3d [0] (~kramer@unaffiliated/kramer3d)
22:46:25dionoeaTorne: hum, I can't find that email
22:46:34gevaertsTorne: IrcNicks on the wiki might help
22:47:00Tornemaybe but i can't be bothered to do that work right now
22:47:08Tornesomeone else is free to, you have the link ;)
22:50:43Tornethere are still 16 people who haven't done it who have committed in the last 18 months. they are clearly not paying attention ;p
22:51:03Tornepeople who haven't committed this decade I can maybe understand ;)
22:53:14 Quit stoffel (Remote host closed the connection)
23:00
23:01:07gevaertsFebs, LambdaCalculus37, Lamed, MarcGuay, Mmmm, Moos, Nico_P, Paprica, Quelsaruk, TiMiD, Toni1, [IDC]Dragon, aliask, archivator, austriancoder, dan_a, debauched_sloth, dreamlayers, earHertz, eigma, elinenbe, gotthardt, hardeep, jyp, lostlogic, lowlight, matsl, midgey, midkay, niobos, phaedrus961, ptw419, roolku, safetydan, sdoyon, shotofadds, stevenm, stripwax, t0mas, theli_ua, tomal, tucoz, vitja: ping
23:01:29gevaertsTorne: done :)
23:01:30Tornehehe
23:01:58TornePeople just pinged: see history just above, or rockbox-committers list
23:02:18Torneif you don't have the committers mail with the link to the spreadsheet ask someone else (not posting it publicly)
23:02:46Tornegevaerts: it's much easier when people's names/emails/usernames/nicks are the same :)
23:04:34 Quit factor (Read error: Connection reset by peer)
23:05:21 Quit evilnick_B (Quit: Page closed)
23:14:08 Quit bertrik (Read error: Connection timed out)
23:14:45 Join bertrik [0] (~bertrik@ip117-49-211-87.adsl2.static.versatel.nl)
23:14:45 Quit bertrik (Changing host)
23:14:45 Join bertrik [0] (~bertrik@rockbox/developer/bertrik)
23:22:21 Join factor [0] (~factor@74.197.205.204)
23:28:46 Quit realodd (Quit: CGI:IRC)
23:41:51 Join mshaxp [0] (~athlonmpp@5ad7b04e.bb.sky.com)
23:42:34 Join realodd [0] (~4f759869@giant.haxx.se)
23:44:39 Quit mshathlonxp (Ping timeout: 240 seconds)
23:51:44 Quit hobby16 (Quit: Leaving)
23:51:47 Join Judas_PhD [0] (~kevin@misterfluffy.dsl.xmission.com)
23:52:13 Join hobby16 [0] (~ubuntu@162.123.68.86.rev.sfr.net)
23:52:15 Quit Judas_PhD (Client Quit)
23:57:33 Join Judas_PhD [0] (~kevin@misterfluffy.dsl.xmission.com)
23:58:14 Quit krazykit (Ping timeout: 246 seconds)

Previous day | Next day