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).

#rockbox log for 2005-04-10

00:05:14rasherI guess that's, like, a no.
00:08:32rasherOh well, I read the logs.. wasn't exactly hoping for much :)
00:11:17rasherShame linuxstb is busy (or whatever he's doing), he got a lot of work done
00:13:55 Join amiconn [0] (
00:14:25amiconnhi again
00:15:42amiconnBagder: Who is in charge of the fat driver (or knows it best)? I found a bug (or rather, my sister pointed me to the problem)
00:16:06Bagderzagor, Linus and Jörg have all done various things in there
00:16:11rasheryour sister appears to be an endless source of bugs
00:16:27rasher(or am I mis-remembering?)
00:16:45amiconnShe actually uses the thing much
00:17:26amiconnMaybe the problem only manifests with fat16, because the fat16 root dir has a fixed size, but the cause might be universal
00:18:03amiconnThe thing is that if a file or dir is deleted, the corresponding dir entry is not completely erased but rather only marked as deleted
00:18:35amiconnIf you copy and delete a lot of files, the dir gets eventually filled with deleted entries.
00:18:46rasherah.. fun
00:18:59amiconnIf this happens, rockbox is unable to save a file in the root, because the fat16 root dir is not extendable
00:19:01rasherit should start overwriting I guess?
00:19:02t0masmicrosoft designed the fat format right?
00:19:32rasherthey did indeed, but I'd think they had this one right afterall
00:19:36amiconnIt should start to overwrite deleted entries
00:20:03amiconnI made an image of the whole flash content for analysis, and also transferred it to my Ondio
00:21:10amiconnpreglow: Btw, that was the cause for the playlist problems I mentioned.
00:22:19rasherhaha, took me like an entire period to notice that I was watching _women's_ hockey
00:26:14preglowsomeone actually uses fat16?
00:26:44amiconnIt's the default file system for flash cards, and rockbox must support it
00:27:04amiconn...because the internal flash of the Ondio must be formatted with fat16
00:27:43amiconnThe archos firmware for Ondio can only boot from a fat16 medium
00:28:03pregloware they going for dos compatibility or what?
00:28:22amiconnAs I said, fat16 is the standard for flash media
00:28:27rasherthere's some reason, but I forget..
00:28:34rasheras always.
00:28:44amiconnfat32 would be overkill for small sizes
00:28:50preglowthere are patent issues with fat32, for one
00:29:42rasherprobably that
00:29:49preglowthe braindead long file name scheme is patented
00:30:18rasher haha
00:30:20 Join Rob- [0] (
00:30:21amiconnThat has nothing to do with fat16<->fat32, it is called vfat and used for both
00:30:27rasherATTENTION! Transfer speed is not infinite!
00:31:25rasherI was not entirely sure where vfat fitted in this scheme.. I was sortof under the impression that it was just another name for fat32
00:31:56rasherI guess because vfat is usually always fat32
00:32:04rashers/always //
00:32:46Rob-how's the iriver port going?
00:33:09rashernot many people are working on it
00:33:38amiconnrasher: It's not. vfat is used for both fat16 and fat32. In fact, vfat was invented (or at least published) before fat32
00:33:39CamiloI've got a camera that won't use fat32 but manages with fat12 fwiw
00:34:15amiconnThe first Version of Win95 (Win 95 and Win 95a) supports vfat, but not fat32
00:34:33rasherthat's nasty
00:34:39amiconnWin 95b added fat32
00:35:04Camilowin95b == win95 osr2?
00:35:45CamiloI'm about to obsolete that @ home.... last win box
00:35:52amiconnBut there were in fact some more versions of osr2
00:36:25amiconnI know at least of osr2, osr2.1 and osr2.5 (Win 95c)
00:36:29rasherI recently downgraded a laptop from 2000 to 98
00:36:34rasherfelt so dirty
00:36:47rasherWho's responsible for the tag db?
00:36:52preglowzagor, i think
00:37:09rasherit seems needlessly case-sensitive
00:37:20BagderI wrote the first perl shot of it
00:37:33Bagderzagor's done the rockbox parts completely
00:37:55rasherI guess this is the perl's fault
00:38:02rasheror at least, better handled there
00:38:22*rasher takes a look at it
00:38:25preglowperl's fault???
00:38:33Bagderthe perl script's fault
00:38:33preglowperl is without flaws
00:38:36rasherthe perl( script)'s
00:38:50rasherperl is naughty
00:39:15 Quit Ago_ ("Verlassend")
00:39:22rashermostly because I don't really know it well enough to like it
00:40:19preglowtaught myself it a couple of years back
00:40:24preglowbeen my favourite language since
00:41:24rasherI haven't really needed it
00:41:47preglowwell, you seldom need it
00:41:52preglowbut it's good for some stuff
00:42:01pregloweverything that has to do with parsing and data handling
00:42:07preglowand cpan is a bloody great resource
00:42:14pregloweverything has been done before and put there for you to use
00:42:18amiconnIt seems I found the fat bug...
00:43:19amiconn...argh, no :(
00:43:22*rasher creates a mental map of
00:44:02CamiloI love perl
00:44:14CamiloI used to work with perl for a living...
00:44:47preglowi'd survive doing that
00:45:08Camiloif I do perl too much it's like a love/hate relationship
00:45:09rasherargh.. this is going to be annoying..
00:45:31CamiloI always find myself doing a little bit of something else on the side
00:45:43rasherguess the easiest to do is to keep a lowercasename => name array (hash? I can never remember which's which)
00:46:02Camiloif I can help with any perl stuff, let me know :)
00:46:03rasherlots of lookups, and double the memory usage
00:46:17 Quit DMJC ("Leaving")
00:46:35rasherI'm sure you could do this faster than me.. it doesn't seem a very complicated problem
00:46:43rasherbut I'm going to be stubborn now
00:47:42Bagderthen you need to either do them lower or upper case
00:47:51Bagderwhich is ugly
00:48:08rasherWell not in the db
00:48:12rasherjust temporarily
00:48:19rasherto check if it's been used before
00:48:27Bagderright, a case-check-hash
00:48:37rasheryes.. with the real-cased version as value
00:48:41t0mas"ANSI C++ forbids implicit conversion from `void *' in assignment"
00:48:48t0mashelp! :)
00:48:59*rasher locks t0mas up
00:49:22Bagdert0mas: use explicit cast?
00:49:38*t0mas hit's his head against the wall :)
00:50:30Bagderbed time!
00:50:44t0masdoesn't work
00:50:47Camilorasher, what's the case problem then?
00:50:48t0masor I'm doing something wrong
00:51:00amiconnI have to correct myself again :-| I found the spot which causes the bug, but the fix isn't that simple. I'd need a fat expert...
00:51:19t0masBagder: I have a function
00:51:19Camiloamiconn, what's the fat query?
00:51:21t0mastypedef void* (*modmain_t)();
00:51:26t0masof the modmain_t type
00:51:30t0masand it return's a void*
00:51:37rasherCamilo: well say I have two files tagged "Album: Foo bar" and "Album: Foo Bar" .. I get two entries
00:51:42t0maswich is a pointer to a function
00:51:50amiconnCamilo: The problem is that adding an entry to the dir needs 2 scans of the dir
00:51:56t0masvoid (*ptr)();
00:52:00t0masptr = modmain();
00:52:04rasherCamilo: I'd like just one of them to be used (the first encountered, I guess)
00:52:05t0masgives that error
00:52:58amiconn(1) Find <n> consecutive free entries, where <n> is the number of entries needed to store the longname and the shortname (they need to be consecutive)
00:53:11amiconn(2) Make sure the shortname isn't already taken
00:53:14preglowdo any of you guys know when 'and' became a legal substitute to && in c?
00:53:41Camilorasher, you could keep a hash of lc($x) => $x then look for things that are the same by checking the hash first
00:53:54rasherYes indeed, that's what I'm planning
00:54:02rasherjust need to figure out where to put it
00:54:02amiconnIn order to save one read sweep, add_dir_entry() does both scans in a single loop
00:54:38Camilok rasher
00:54:59amiconnThe implementation does not work properly if there is a non-extendable dir which has no free slots
00:55:19rasherI think I have it now..
00:58:11rasherwhat's the difference between $$id3{'ARTIST'} and $id3->{'ARTIST'} ?
00:59:09t0mashmz... Segmentation fault == bed time
00:59:15*t0mas is going to bed...
00:59:31t0mashope I see it tomorrow :)
00:59:36 Quit t0mas ("good night")
01:01:04Camilorasher, the $$ one has an extra indirection
01:01:42Camiloin the example the -> is syntactic sugar and can be omitted (I think)
01:02:30rasherextra indirection?
01:02:58Camiloeg. $y = 'hello'
01:03:04 Join lostlogic_ [0] (
01:03:09Camilo$x = &$y
01:03:17Camiloprint Dumper($x)
01:03:38rashercrazy.. oh well.. let's see if this did it
01:04:09Camilo(Sorry \ not &)
01:04:45Camilothen compare print $x with print $$x
01:12:35 Quit lostlogic (Read error: 110 (Connection timed out))
01:13:31rasherhrm, it almost worked
01:14:08rasherit's using the right albumname, but created a new entry nonetheless
01:16:02rashernot sure I understand that :-\
01:18:17amiconnThere are more problems in add_dir_entry()...
01:18:35rasherMy code is wrong I guess
01:18:42rasherCamilo: want to have a look?
01:20:35rasherit's pretty simple.. I imagine I'm either missing something in the code, or it's just my lack of perl experience.. or both
01:23:29Camiloyeah let me have it
01:23:45Camilowhere can I get it?
01:23:58CamiloI can't DCC">
01:26:04Camilotry this
01:26:47Camiloif (exists($lcalbums{lc($id3->{'ALBUM'})}))...
01:27:10rasherwill do
01:28:02Camiloif (exists($lcalbums{lc($$id3->{'ALBUM'})}))...
01:28:19Camilosorry, if the original code uses $$ to access that hash, then you should, throughout
01:28:39rasherwell it uses both
01:28:41Camilo$$id3{'ALBUM'} = $lcalbums{lc($$id3->{'ALBUM'})};
01:28:48rasherwhich confuses me
01:29:10Camilo$$id3{'ALBUM'} = $lcalbums{lc($$id3{'ALBUM'})};
01:29:16Camiloyeah it's not the nicest
01:29:42Camilothey are definitely different - I recommend using $$ in this case because there are $$ references
01:29:54Camilobut in general I'd rarely use $$
01:30:16rasherlet me see..
01:32:44rasherweirder and weirder
01:32:57rasherit's not working for artists, but is for albums.. or maybe it's not working at all
01:33:11*rasher prints some more debug messages
01:33:32***Saving seen data "./dancer.seen"
01:34:35 Quit Sucka ("a bird in the bush is worth two in your house")
01:35:38CamiloI put up
01:36:43rasher200 Km/h In The Wrong Lane - exists, now: 200 km/h in the wrong lane
01:36:57rasherwell that looks right (not a word about me having tatu songs!)
01:37:33rashermaybe I need to do more..
01:37:48rasherI think that's what I have now
01:40:52CamiloI'm gonna go now but mail me if you're stuck
01:40:54rasherthey were by different artists
01:41:08Camiloso it's OK?
01:41:08rasherwell, differently tagged
01:41:12rashernot sure
01:41:17rasherI need to find another example
01:41:27rasherbut I'm pretty sure the original version wasn't sure
01:41:31rasherdidn't consider case at all
01:41:36rasheruh, wasn't ok
01:42:07rasherguess I'll just fix the tag instead
01:42:36CamiloI have nearly 5000 badly tagged files here to try it on :)
01:43:15*preglow hugs freedb masstagger
01:43:17Camiloanyway mail me if I can help more, I'm and I'm going to sleep now
01:43:56rasherI think I'll be able to fix the rest now
01:48:16rasherit's working!
01:51:45rasherwell that was fun
01:53:39rasherpreglow: that sortof assumes full albums :-\
01:53:56preglowwhy, yes
01:54:08preglowi always have full albums
01:54:14preglowmore or less always, at least
01:54:27rasheranyway, most of my files are tagged nicely
01:54:30rasherjust not album
01:54:34rasheronly artist/title
01:54:46rasherexcept for full albums
01:54:57rasherthose are nicely tagged
01:55:21*rasher head off to open a patch-thingy in the patch thing
02:00:53rasherwait a minute!
02:03:06rasherthis is murder for multi-artist cds!
02:03:12rasheryou get an album-entry for each artist :(
02:04:08rasheroh haha
02:04:47rasherI was getting the same error now.. but.. I was using the old version of the
02:05:38 Quit Camilo (Read error: 110 (Connection timed out))
02:07:00preglowsong dbs act that way, rasher
02:07:35rasherI built a new song db
02:07:42rasherbut with the old version of the script ^.^
02:07:45 Join izzy_ [0] (
02:08:56 Part amiconn
02:10:37rasheriripdb takes ages
02:10:43rashercompared to
02:10:53preglowiripdb is the iriver one?
02:10:54rasheroh.. could be some cache thing
02:10:56rasherprobably is
02:10:58preglowyes, it's shite
02:11:26rasherah, it was cache
02:11:35rasherran iripdb first, then
02:11:39 Join gromit^ [0] (
02:11:49rasherso naturally iripdb had to load from disk, and then, everything was in cache
02:11:55rasherthat sure sped it up
02:12:38rasherlooks like it's working
02:13:55*rasher ponders some automatic checking of the patches
02:14:07rasherhow many errors they spout when applied to the cvs version
02:14:21rasheras a way of seeing how relevant/up to date they are
02:15:41rasherBut now I sleep.
02:16:59 Quit izzy (Read error: 101 (Network is unreachable))
02:40:27rasherYup, I will.
02:40:30 Quit rasher ("CGI:IRC 0.5.4 (2004/01/29)")
02:49:20 Join DMJC [0] (
03:33:34***Saving seen data "./dancer.seen"
03:46:57 Join izzy [0] (
03:55:44 Quit izzy_ (Read error: 145 (Connection timed out))
04:02:56 Quit preglow ("wap")
04:04:13 Join ashridah [0] (
04:06:11 Join QT_ [0] (as@area51.users.madwifi)
04:20:13 Quit QT (Read error: 110 (Connection timed out))
04:37:17 Quit izzy (Read error: 110 (Connection timed out))
04:49:58 Quit DMJC ("Leaving")
05:26:48 Quit gromit^ ("Leaving")
05:33:37***Saving seen data "./dancer.seen"
05:37:33 Quit karim ("Leaving")
05:51:03 Join TCK| [0] (
05:53:21 Quit TCK (Read error: 104 (Connection reset by peer))
07:33:38***Saving seen data "./dancer.seen"
07:35:22 Join DMJC [0] (
07:58:21 Quit WireDDD (
07:58:21 Quit _Lucretia_ (
07:58:21 Quit Nibbler (
07:58:21 Quit bobdbob (
07:58:21 Quit courtc (
07:58:21 Quit webmind (
08:00:55NJoinWireDDD [0] (
08:00:55NJoin_Lucretia_ [0] (
08:00:55NJoinNibbler [0] (
08:00:55NJoinbobdbob [0] (
08:00:55NJoincourtc [0] (
08:00:55NJoinwebmind [0] (
08:02:54 Join asdsd_ [0] (
08:03:05 Join webmind_ [0] (
08:14:10 Join courtc_ [0] (
08:14:52 Join NibbIer [0] (
08:16:08 Join _Lucretia [0] (
08:16:39 Quit courtc (Nick collision from services.)
08:16:46 Nick courtc_ is now known as courtc (
08:17:13 Quit _Lucretia_ (Connection timed out)
08:17:31 Quit Nibbler (Connection timed out)
08:17:35 Quit WireDDD (Connection timed out)
08:18:10 Quit webmind (Connection timed out)
08:20:04 Quit bobdbob (Read error: 110 (Connection timed out))
09:21:41 Join amiconn [0] (
09:33:42***Saving seen data "./dancer.seen"
09:34:02 Quit DMJC ("Leaving")
09:39:28 Quit asdsd_ (Read error: 110 (Connection timed out))
09:42:38 Quit XShocK (" HydraIRC -> <- The professional IRC Client")
09:50:45 Join rasher [0] (
09:58:10 Join Wired [0] (
10:14:18amiconnboogss are us...
10:20:05rasher'tis tragic
10:42:55HCloops, capslock
10:53:28 Join muesli- [0] (
10:58:00 Nick QT_ is now known as QT (as@area51.users.madwifi)
11:01:31*HCl obviously isn't awake yet...
11:01:34HClnightmares suck..
11:01:36HClso whats new?
11:02:16muesli-charles and camillia got married
11:05:21HClso did lots of other people, most likely
11:07:15muesli-i didnt, did you? :D
11:13:49 Join Sucka [0] (
11:13:49 Quit rasher ("CGI:IRC (Ping timeout)")
11:19:30 Quit amiconn (" bbl")
11:29:30 Nick Sucka is now known as Sucka`away (
11:33:45***Saving seen data "./dancer.seen"
11:35:31 Quit muesli- (Read error: 113 (No route to host))
11:47:10 Quit ashridah ("Leaving")
11:47:25 Join ashridah [0] (
12:02:47 Join t0mas [0] (
12:47:54 Join xen` [0] (
12:52:05 Quit Ka (Read error: 145 (Connection timed out))
13:03:38 Join Ka [0] (
13:19:01 Join matsl [0] (
13:33:49***Saving seen data "./dancer.seen"
13:49:24 Join amiconn [0] (
13:54:20 Join DMJC [0] (
14:18:17 Join preglow [0] (
14:25:25 Quit Wired (Read error: 104 (Connection reset by peer))
14:59:51 Join webguest05 [0] (
15:00:08webguest05can anyone tell me how to install rockbox on my H120
15:02:02webguest05anyone here...?
15:04:07 Quit webguest05 (Client Quit)
15:12:04preglowmethinks he doesn't really want it
15:13:27 Join gromit^ [0] (
15:15:42DJ_Dooms_Day|AFGod you guys are ass's :P
15:16:11elinenbeDJ_Dooms_Day|AF: what's the deal with your nick?
15:18:18 Join pfavr [0] (
15:29:51 Quit DMJC ("Leaving")
15:33:54***Saving seen data "./dancer.seen"
16:19:55 Join einhirn [0] (
16:21:21 Nick DJ_Dooms_Day|AF is now known as DJ_Dooms_Day (DJ_Dooms_D@
16:23:35 Join bagawk [0] (~Lee@bagawk.user)
16:51:19 Nick Sucka`away is now known as Sucka (
16:53:19 Join wacky_ [0] (
16:53:25wacky_Anyone tested the 1.65 firmware yet ?!
16:53:35 Join karim [0] (
16:53:49wacky_I would have liked to have these recording options.. but I can't leave the on-the-fly playlists !
16:53:52ashridah1.65 eh? what's in it? i don't recommend trying it with the rockbox bootloader :)
16:54:07wacky_well.. a couple of new recording features..
16:54:12*ashridah stopped going to iriver's site when it became a horrible mess of flash-based tosh
16:55:07ashridahthat's it/
16:55:07wacky_LinusNielsenFeltzing 2004-04-10: It has now been tested with 1.65 without problems.
16:55:09wacky_oh yeah!!
16:55:24ashridahso only the recording stuff? no other features?
16:55:56wacky_I'm goign to check out right away :)
16:56:58karima question
16:57:08karimis there an input volume control ?
16:57:14karimfor recording
16:57:24karimthat's really missing in the iriver firmware
17:02:10wacky_dunno.. I'm going to check out
17:05:26preglowi think i'll just leave mu 1.63 in
17:09:55 Quit bagawk (Read error: 110 (Connection timed out))
17:23:27 Quit ashridah ("Leaving")
17:28:22elinenbeamiconn: did you ever get your iriver yet?
17:31:33amiconnStill no iriver :(
17:33:58***Saving seen data "./dancer.seen"
17:36:28 Join asdsd_ [0] (
17:43:30 Join [IDC]Dragon [0] (
17:43:45amiconnhi Jörg :)
17:43:58[IDC]DragonHi Jens, long time no IRC
17:44:14[IDC]Dragonbut I've read the logs, partially
17:44:28amiconnDid you also read about the fat problem?
17:44:35[IDC]Dragonam too busy for "active" rockboxing
17:44:59[IDC]Dragonbut I've never been in the directory code
17:45:08amiconnLooks like I need to (partially) rewrite add_dir_entry()
17:45:15amiconnIt has at least 3 problems
17:45:30[IDC]Dragonwould be much better, to recycle the entries
17:45:49[IDC]Dragonand it's not time critical code
17:46:13karimanybody here avec experience with USB 2 ubs ?
17:46:31amiconn(1) It fails on non-extensible dirs, e.g. the fat16 root, but also if an ordinary dir isn't extendable because there is no free space
17:48:00*[IDC]Dragon waits for (2) and (3)
17:48:03amiconn(2) It may fail if a dir is extendable, but not extendable enough, e.g. if extending by one cluster is possible, but not by two. With 512 byte clusters, a longname entry may take more than 1 cluster
17:48:52amiconn(3) It would never ever reuse the very first deleted entry, because it uses index 0 as an indication of "not yet found"
17:49:30DJ_Dooms_DayHey, any of you guys interested in helping create a RPG/inventory system in doom 3?
17:49:33[IDC]Dragon(3) is probably neglible, or it could return -1 if signed
17:49:36amiconnThen I have another problem, which you also might have read about. My 2 GB MMC arrived, but I can't use it with rockbox yet
17:49:47[IDC]Dragonyes, interesting
17:50:04[IDC]DragonI'm surprised it works with Archos
17:50:31amiconnWell, it doesn't work correctly. Archos can read the dirs, but playing music isn't possible
17:50:50amiconn(silent play and slower-than-usual scrolling)
17:51:05 Join milkclub [0] (
17:51:09[IDC]Dragonsilet play, haha
17:51:32[IDC]Dragonmaybe the MMC is from Hitachi?
17:52:10amiconnNo it isn't. It's branded 'magu'. Afaik these use samsung components
17:52:27milkclubHey guys. Do you mind if I ask a total n00b question about the rockbox firmware for iRiver H1x0?
17:54:05[IDC]Dragonamiconn: btw, how about using my MMC USB mode powerup patch?
17:54:16preglowmilkclub: go ahaead
17:54:35amiconn[IDC]Dragon: I still did not look into that :(
17:55:13[IDC]DragonI thought it might be appropriate while you're on the MMC subject
17:55:39[IDC]Dragonit seems to uncover a problem
17:56:15milkclubIs there any plan to have 'most played' style playlists built in, that monitors what you play most? It's pretty much the only thing on the iPod I envy, just a silly little thing.
17:56:48preglowwell, no plans, but if anyone wants to code it, it's just a matter of doing it
17:56:52preglowsounds like a pretty simple thing
17:57:10milkclubthat's what i thought. i haven't got a clue how, though.
17:57:30milkclubi thought as there were going to be games, it might be a nice little feature. just an idea.
17:57:38[IDC]Dragonthere's a patch which logs what you've played, iirc
17:57:51amiconn[IDC]Dragon: Maybe I'll do that, but I have a somewhat higher priority to get the new card up and running
17:58:26[IDC]Dragonamiconn: sure. Once you're an MMC user, the desire to stat wit it will come... ;-)
17:58:32amiconnThis will be quite some work, adding handling of variable block size
17:58:57[IDC]Dragonfor a start, you could encapsulate it
17:59:26[IDC]Dragonalthough inefficient 4 times slower
17:59:31amiconnThat needs to be done anyway. The fat driver needs to work on 512 byte sectors, always
17:59:46amiconnI need to implement a block caching system
17:59:53*[IDC]Dragon wonders why the MMC does so
18:00:10[IDC]Dragonare you sure it has to be addressed this way?
18:00:25amiconnIn fact MMC can have all power-of-two block sizes up to 2048 bytes
18:00:59[IDC]DragonI remember you told me about commands addressing on byte level, hence a 4GB limit
18:01:03amiconnYes I am sure. Both the MBR and the partition boot sector are based on 512 byte/sector
18:01:52amiconnYes the commands always address byte wise, but you can't easily address partial blocks
18:02:25 Quit milkclub ("thanks for the help guys")
18:02:48amiconnThe card has 2 bits that tell whether it supports partial block reading (they usually do) and partial block writing (they usually don't)
18:03:38amiconnSo it's easier to assume partial blocks are completely impossible
18:06:07[IDC]Dragonyes, although the flags won't be too hard to handle, once you have the code
18:06:58amiconnIt would mean more code, and given the fact that all cards I've checked so far don't support partial block writing, I need the block caching anyway
18:07:57 Join Stryke` [0] (~Chairman8@resnet-241-86.resnet.UMBC.EDU)
18:08:13amiconnWriting will be significantly slower (2 times for 1024 byte/block, 4 times for 2048 byt/block) if ata_write_sectors is called (sequentially) with count==1
18:08:52amiconnCalling it with a larger count won't be much slower than for the default 512 byte/block
18:08:58[IDC]Dragonwe could have a "cache" for 1 block
18:09:07amiconnI intend to cache 2 blocks
18:09:24[IDC]Dragonor that
18:09:48amiconn..for swap-copying the next block while one block is transferring on write, like it is now
18:10:11*[IDC]Dragon remembers vaguely
18:10:46[IDC]Dragonand I drool about your 2 GB card, in general
18:11:08amiconnThe really ugly thing is that for >512 byte/block, writing a single sector needs to read the whole block first (if it's not already cached), replace the sector, then write back the whole block
18:12:35[IDC]Dragonand with "simple" caching, you don't know if there's more to come
18:12:55[IDC]Dragonbut for multi-sector transfers you do
18:14:01[IDC]Dragonless than twice as much as my 1GB
18:14:21amiconnThe cache will work similar to the fat cache, i.e. remember which block(s) it holds, so subsequent writes to the same block saves the read
18:15:11[IDC]Dragonyes, but if the whole block is written, the read was unnecessary
18:15:12amiconnI also consider to use deferred write-back, i.e. only write back (1) if the cache is needed for a different block or (2) a timeout expires
18:16:01amiconnThat would solve the write slowdown and reduce the flash wear for large blocks (by factor of 2/4)
18:16:55amiconnThis is probably the second step, but needs handling of special cases (unit shutdown...)
18:17:08preglowhow many write cycles do those cards stand?
18:17:10[IDC]DragonMMC extraction
18:17:24[IDC]DragonI'd be careful about write caching
18:18:20amiconnThe timeout would be small, I think 0.1 s would be more than enough to catch sequential single-sector writes
18:19:01[IDC]DragonI see it's all in your mind already
18:19:25amiconnAnother bad thing is that the block cache needs to be moved out of IRAM (not enough space)
18:19:59[IDC]Dragonwas that useful?
18:20:11[IDC]DragonMMC transfer is slow anyway
18:20:43[IDC]Dragonthanks to our SH, the MMC is not to blame
18:20:46amiconnIt speeds up transfer to/(from) the cache
18:23:29amiconnI didn't measure...
18:24:23amiconnI remember that you managed to run the fat test. Could you tell me something about it?
18:24:48amiconnI once looked at it, but didn't really understand how it's supposed to be used
18:25:43[IDC]Dragonit probably needs to be updates, since multivolume
18:25:57[IDC]Dragonwhat don't you understand?
18:26:23[IDC]Dragon(afk for a minute)
18:29:08amiconnWell, I don't understand how I am supposed to actually run this or that test
18:30:05amiconnMounting the test image to access it from the PC isn't the problem. I have at least 2 options (now)
18:31:01[IDC]Dragonyou need linux
18:31:08amiconnNot really
18:31:35[IDC]Dragonfor the shell script chkdsk
18:32:08[IDC]Dragonfor my test, I stripped that off the script
18:32:28amiconnWell, I can use my linux installation (debian under vmware)
18:32:37[IDC]DragonZagor than later ran the full thing for me
18:33:19[IDC]DragonI have a little freeware program that can mount an image file under Win32
18:33:45amiconnI could also use the image as a file disk for a virtual machine, then run windows in it
18:35:28[IDC]Dragonthe progam is called VDK, and is originally made for VMware images
18:35:35amiconnThe shell script isn't of much use for testing my directory fixes, since it creates its own test images. I need to test on an image with a filled-up (with most entries marked as deleted) root dir
18:36:07amiconnOf course the test script should also run without error
18:36:09[IDC]Dragonyou can adapt it
18:36:14 Join DJ_Dooms_Day|AF [0] (
18:36:23[IDC]Dragonmaybe I did so, too
18:37:29[IDC]DragonI think I even made a plain batch file
18:46:05[IDC]Dragonamiconn: for your 2 GB card, the cluster size should be 32KB
18:46:36[IDC]Dragonso you won't see too many single sector transfers, right?
18:46:44amiconnyes and no.
18:47:02amiconnThe cluster size is indeed 32KB, because it is fat16 formatted
18:47:40amiconnBut there is a number of functions that uses single sector buffers. add_dir_entry() is an example....
18:48:10amiconnThe data transfers should be at least cluster sized
18:48:22[IDC]Dragonyes, that's what I meant
18:48:52[IDC]Dragonand if the start sector is block aligned, you don't need to read before write
18:50:06*[IDC]Dragon remembers peple with problems on 1 GB cards, perhaps these were != 512 byte/block ?
18:54:03 Quit DJ_Dooms_Day (Read error: 110 (Connection timed out))
18:56:43 Quit asdsd_ ("Trillian (")
18:57:34 Join asdsd_ [0] (
19:08:47*[IDC]Dragon waves goodby
19:08:52 Quit [IDC]Dragon ()
19:16:08 Join amiconn_ [0] (
19:17:01 Quit amiconn (Nick collision from services.)
19:17:01 Nick amiconn_ is now known as amiconn (
19:17:19 Quit elinenbe (Read error: 60 (Operation timed out))
19:26:27 Quit DJ_Dooms_Day|AF ()
19:30:21 Nick _Lucretia is now known as _Lucretia_ (
19:31:49 Join amiconn_ [0] (
19:31:53 Quit amiconn (Nick collision from services.)
19:31:55 Nick amiconn_ is now known as amiconn (
19:34:00***Saving seen data "./dancer.seen"
19:35:20 Join mecraw [0] (
19:35:39 Part mecraw
20:22:34 Quit karim (Remote closed the connection)
20:26:51 Join rasher [0] (
20:40:11 Quit amiconn (Read error: 110 (Connection timed out))
21:00:03 Quit rasher (Read error: 110 (Connection timed out))
21:00:38 Join rasher [0] (
21:01:13rasherFlashed with 1.65EU now.. working fine :)
21:04:56rasherBagder: why did you stop the Rockbox Digest? Lack of time/interest or something else?
21:06:36thegeekwas the loader injected?
21:07:28rashernot much -
21:13:06 Quit Rick ("I… don't need to be here.")
21:21:22 Join Rick [0] (
21:26:32 Join TCK [0] (
21:29:42 Join izzy [0] (
21:34:01***Saving seen data "./dancer.seen"
21:44:56 Quit TCK| (Read error: 110 (Connection timed out))
21:47:09 Nick webmind_ is now known as webmind (
21:50:12 Join _aLF [0] (
22:17:26 Join kergoth [0] (
22:22:32 Join bill2or3 [0] (
22:23:30bill2or3can anyone point me to a comparison of the different models features?
22:23:38bill2or3(of the archos)
22:24:27 Join TCK| [0] (
22:25:41bill2or3I'm looking at a Jukebox20, and dont want to end up getting the wrong model..
22:31:54 Join Aison [0] (
22:31:54 Quit TCK| (Read error: 104 (Connection reset by peer))
22:32:09 Join TCK| [0] (
22:35:24 Join TCK- [0] (
22:39:46 Quit asdsd_ ("Trillian (")
22:39:46 Quit xen` (Read error: 110 (Connection timed out))
22:40:15rasherbill2or3: there's a feature comparison page on
22:42:06bill2or3thanks rasher
22:43:22 Quit TCK (Read error: 110 (Connection timed out))
22:44:38bill2or3so the 'FM Recorder' is the best one..
22:45:16rasherUnless you're prepared to wait for the iriver port
22:45:21 Join elinenbe [0] (
22:46:12bill2or3I think I'd like to have the fm tuner
22:46:31rasherthe iRiver has fm
22:46:38bill2or3ahh.. a lot smaller too.
22:47:00rashernot sure how the price compares though
22:47:09rasherplus, they're out of production and hard to get hold of
22:47:16 Quit gromit^ ("Leaving")
22:48:00bill2or3that looks like it'll go pretty cheap..
22:48:53rashersure does
22:49:04bill2or3how big of a drive can I stick in there?
22:49:16rasherNo idea
22:49:20rasherpretty big I think
22:50:03rasheras long as it's 9.5mm high, most will work I think
22:51:40bill2or3I found an old irc log that says up to 120g
22:53:21elinenberasher: is it true that you have a version of rockbox for the iriver that can rolo into the iriver firmware?
22:53:47rasherNo, that was not me
22:53:59 Quit TCK| (Read error: 110 (Connection timed out))
22:54:19bill2or3the chart says no drive spindown on the player, that's gotta suck batteries.
22:55:14elinenberasher: thanks... no problem.
23:00:53rashergoogling to find out who it is... I'm also interested in that
23:07:46 Join amiconn [0] (
23:08:52amiconnbill2or3: It's not that the player can't spin down the disk (the would really suck battery), but it can'
23:09:02amiconnt physically power down the disk
23:09:48bill2or3ahh, so it's not a big deal battery-wise?
23:09:55amiconnIn fact, almost all players except the very oldest ones can do that, but it's not enabled in the official rockbox builds because a few players crash if they try
23:10:17bill2or3is that Jukebox10 I linked on ebay a model worth getting?
23:10:18amiconnPhysically powering down the disk just saves a little more battery
23:10:34bill2or3I see that it has the character display only.
23:11:32amiconnI wouldn't really recommend the player, because of 3 main disadvantages
23:11:49amiconn(1) No graphic display (2) USB 1.1 only (3) No recording
23:12:00bill2or3yeah, (2) is a killer
23:12:01 Quit rasher ("CGI:IRC (EOF)")
23:12:29bill2or3it looks like all the fun plugins need the graphic display.
23:12:52amiconnIt would become a real issue with a larger HD. Btw, as you asked: The jukeboxes use standard 2.5" laptop drives, height 9.5 mm.
23:13:17 Join rasher [0] (
23:13:34t0mashi rasher
23:13:37t0masand bye everyone
23:13:39 Quit t0mas ("Goodnight")
23:13:42amiconnUp to 128 GB (or 137 GB as the hd manufacturers count) are possible; up to 100 GB are available
23:14:49amiconnbill2or3: Depending on the features you want, and if you don't want to wait for the iriver port to finish, I'd recommend either a v1 recorder or an fm recorder
23:15:46amiconnThe fm recorder has an fm tuner (obviously), but reception isn't all that good afaik (I don't have one myself). It's also a little lighter because it uses a LiIon battery
23:15:52bill2or3what's the model name for the V1 recorder?
23:16:06amiconnThe v1 recorder has the advantage of (simply exchangeable) NiMH AA batteries
23:16:46amiconn'Archos Jukebox Recorder'. If you go for one, note that only the 20 GB models have USB2.0
23:16:57bill2or3is that one I just linked to supported by rockbox, or is it one of the newer video ones?
23:17:36amiconnThis is a Jukebox Multimedia and doesn't support rockbox
23:19:35 Quit pfavr ("ChatZilla 0.9.61 [Mozilla rv:1.7.6/20050324]")
23:19:36amiconnA v1 recorder:
23:21:32bill2or3AS IS - Selling for Parts or Repair. Item powers on but does not function.
23:21:43bill2or3what are the odds that that is just a crashed drive?
23:22:06amiconnOoops, overlooked that :(
23:22:28amiconnAn FM recorder, somewhat expensive already:
23:22:43bill2or3do the drives fail often if they get dropped?
23:23:04bill2or3because if I could get that first one for <$50 and put a drive in, it'd be a deal.
23:24:20amiconnThe most observed problem (from looking at the ml and the forum) is bad contacts (battery, grounding)
23:24:48bill2or3that one says it powers on.
23:24:59bill2or3I'd guess it would be a failed drive.
23:25:07bill2or3if it stays cheap maybe I'll bid
23:25:55amiconnAnother FM recorder:
23:26:12bill2or3which model do you have?
23:26:21amiconnIt seems to me there is more junk on ebay usa than on ebay germany...
23:26:44bill2or3I always look at for scooter parts.
23:26:50amiconnI have a recorder v1 (USB2, upgraded to 80 GB), a Studio 10 and an Ondio SP
23:30:31amiconnv1 recorder:
23:30:49rasherdammit, I found discussion about wanting "boot iriver firmware" menu entry
23:31:03rasherbut not the actual thing
23:31:36bill2or3so the difference between the V1 and V2 recorders is the just the batteries?
23:34:02***Saving seen data "./dancer.seen"
23:34:35amiconnThe batteries is the main difference. There are some more slight differences: (1) The v2 and fm don't have the option to connect a remote. (2) The v1 has both digital in and out, the v2/fm have digital in only
23:34:52amiconnBasically, the v2 recorder is the fm recorder without the radio
23:35:34bill2or3ah. so you can hook up a mic to the v1?
23:35:38bill2or3that'd be nice.
23:37:15amiconnBoth types have a builtin mic (which is only really usable for voice notes because it will pick up the hd noise). In order to hook up an external mic you'll need a preamp
23:37:28 Join charlieboy [0] (
23:37:49bill2or3well I dont really care about digital in or the remote.
23:40:39charlieboyhello :)
23:40:52 Quit matsl ("Leaving")
23:41:38 Nick TCK- is now known as TCK (
23:42:27 Nick TCK is now known as TCK- (
23:47:11bill2or3what rate does the fm recorder record at?
23:52:45*rasher goes insane
23:57:10 Quit _aLF ("Leaving")
23:57:27 Quit charlieboy ("CGI:IRC")

Previous day | Next day