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 2002-10-06

00:04:34 Join Jet8810 [0] (
00:05:08 Nick seb-away is now known as _seb_ (
00:17:02 Quit rad[idle] ()
00:20:52 Quit Mighty_ ("[BX] Beavis & Butthead use BitchX. Huh-huh huh-huh huh-huh he said "Bitch".")
00:27:29 Join rad[idle] [0] (
00:41:22 Quit rad[idle] ()
00:41:47***Saving seen data "./dancer.seen"
00:51:25 Join rad[idle] [0] (
01:02:42 Join datazone [0] ([sSWkRrXLU@
01:06:30 Quit Jet8810 ("Client Exiting")
01:08:32 Quit adi|home ("Client Exiting")
01:12:30 Join ddb [0] (
01:12:42 Join breaker [0] (
01:13:08 Join Remo [0] (
01:18:20 Quit ddb ()
02:14:13 Quit rad[idle] ()
02:32:34 Quit breaker (Read error: 110 (Connection timed out))
02:41:50***Saving seen data "./dancer.seen"
02:43:38 Part PiotR
03:02:25 Join langhaarrocker [0] (
03:30:34 Join Synthe` [0] (
03:31:10 Quit Synthe (Read error: 104 (Connection reset by peer))
03:49:09 Quit langhaarrocker (Read error: 110 (Connection timed out))
04:11:15 Join rad[idle] [0] (
04:16:33 Quit rad[idle] ()
04:26:35 Join werty1st [0] (
04:26:47DBUGEnqueued KICK werty1st
04:26:47werty1stsomebody here4?
04:34:08 Part Remo
04:41:51***Saving seen data "./dancer.seen"
05:13:01 Nick _seb_ is now known as seb-sleep (
05:21:19 Join ddb [0] (
05:24:13 Part ddb
05:39:05 Join ddb69 [0] (
05:40:42 Part ddb69
06:28:41 Join hardeep [0] (
06:41:55***Saving seen data "./dancer.seen"
07:05:36 Join adiamas [0] (
07:06:16 Nick adiamas is now known as adi|home (
08:41:56***No seen item changed, no save performed.
09:26:00 Quit hardeep ("Leaving")
09:46:14 Quit werty1st ()
10:27:01PsycoXulwhy doesn't the "Supported" file filter show .mod files
10:29:26adi|homefor player i assume?
10:29:34adi|homeare you able to connect to the rockbox site
10:29:57PsycoXuli could a minute ago when i looked at the TODO...
10:30:10PsycoXulyeah still can...
10:30:22PsycoXuland yeah, player...
10:30:42adi|homelet me update my cvs files and ill look at the source
10:31:56PsycoXulthere's a bug report for low clicks when changing volume
10:32:12PsycoXulreminds me of a similar bug the c64 SID chip had
10:32:20PsycoXulwhich got exploited to playback low quality PCM audio...
10:32:25PsycoXuli wonder if we could do something similar :p
10:34:48PsycoXuloh well i'm going to bet now
10:39:51 Join rad[idle] [0] (
10:41:58***Saving seen data "./dancer.seen"
10:45:17adi|homedamnit i can't seem to reach the website
11:01:38rad[idle]works fine from here (germany)
11:24:08adi|homesomethings down between the site and east coast usa
11:24:12adi|homei can't even ping it
11:37:56adi|homePsycoXul: i show the .mod file showing in the sim.
11:50:25 Join LinusN [0] (
12:00:28 Join Electrocut [0] (
12:01:36Electrocuthas sb succeeded in using jukebox with 95 ?
12:01:45adi|homeLinusN: just thought you might want to know that i can't hit the rockbox site from eastern coast usa
12:01:51LinusNElectrocut: win95?
12:01:55adi|homeElectrocut: huh?
12:02:10LinusNadi|home: i can reach it from here
12:02:17Electrocutyes ... i want to use it with a P133
12:02:33adi|homeright.. but your not east coast usa ;)
12:02:35LinusNElectrocut: i don't think win95 has usb storage support at all
12:02:42adi|homeheheh nothing you can do about it.. just pointing it out.. :)
12:02:52Electrocutso I will use nux on it ... :)
12:03:00LinusNElectrocut: of course
12:03:36Electrocutdrivers are easy to set for a nux-newbe ?
12:03:50LinusNnothing is easy for a linux-newbie
12:04:02Electrocutok ;-/
12:04:11LinusNElectrocut: install the latest 2.4.19
12:04:18LinusNor even 2.5
12:04:43LinusNenable ush-storage as a module
12:04:46Electrocutok ... i'll do that
12:05:04adi|homeElectrocut: linux is a steep learning curve
12:05:08adi|homebut a continuious one
12:05:15adi|homeand when you reach the top.. nice view ;)
12:05:21Electrocutadi|home ;)
12:05:59Electrocuti'll just get a 200 Mo Hd, so i can't install 98se, that's why ;)
12:06:47Electrocutwhen you connect the archos, it's like windows ? (new driver letter ??)
12:07:15Electrocutnew drive letter ... sorry
12:07:28LinusNyou get a new SCSI device
12:07:40LinusNwhich you mount somewhere in the file system
12:07:48LinusNfor example /mnt/archos
12:07:58LinusNyou add a new entry in /etc/fstab
12:08:11LinusNor do mount -t vfat
12:08:29LinusNthere are a lot of howto's out there
12:08:30Electrocutok ... and you can directly browse in it ? like the explorer ?
12:08:47LinusNit becomes a new directory under /mnt/archos
12:09:06LinusNthen you can use wny tool you like to browse it
12:09:15LinusNi prefer 'cd' and 'ls'
12:09:26Electrocutok ... thanks :)
12:19:31 Join breaker [0] (
12:28:02breakerI have the problem that I have no admin access for the trackers on sourceforge anymore.
12:29:22LinusNreally? lemme check
12:29:24breakerI committed some lang files yesterday and cannot change the status to fixed/done/..
12:30:44LinusNi can't remember that you ever had admin access...
12:31:03LinusNactually, only 4 people have admin access
12:31:12breakerHm? I really thing that I changed some items.
12:31:18breakerSoem days ago.
12:31:21LinusNi think so too
12:31:32breakerAnd why could I do so?
12:32:07LinusNyou must have been admin by mistake somehow
12:32:18breakerSourceforge prob?
12:32:39LinusNprobably me :-)
12:33:03breakerBut then you must have changed that state to no access..
12:33:18LinusNthat must have been Zagor
12:34:10breakerIf I shouldn't have access, it's ok. But then I can only add the text "committed." when I have committed a patch (for ex. .lang file) that was on the tracker.
12:34:27LinusNthat would be ok
12:34:47breakerAnd then you four see it and change it to done.
12:34:56LinusNstill, the reason is that the project administrators decide which patches are accepted or not
12:35:10LinusNsmall lang patches are generally ok
12:35:29breakerThen I shouldn't commit the patches that are on the patch tracker?
12:35:38LinusNnot all of them
12:35:57LinusNfor example, the queue patch is something we'd like to discuss first
12:36:05breakerYes, I wouldn't do so generally.
12:36:11breakerYes. True.
12:36:58LinusNit's all a matter of common sense, really
12:37:17LinusNi think you know what you may commit and what not
12:37:18breakerAnd the things I can commit (like the lang things) are the minority. Then it's ok. I'll add a "committed and done" text to them.
12:37:47breakerYes, I think thats clear to me.
12:38:33breakerOK then, thanks for the help!
12:38:56LinusNany time
12:42:00***No seen item changed, no save performed.
12:58:38 Quit Electrocut ("Leaving")
13:09:44 Join Django [0] (
13:20:01 Join Lear [0] (
13:27:01 Join ahaaa [0] (
13:28:16 Part ahaaa ("Client Exiting")
13:28:54 Join langhaarrocker [0] (
13:29:17langhaarrockerI'm fed up with bindable keys
13:29:26LinusNwe have serious problems with handling the remote control keys
13:29:33LinusNlanghaarrocker: already? :-)
13:29:36langhaarrockerkey release?
13:29:50LinusNno, the WPS is so focused on the keypad
13:30:04LinusNand the remote is blocked when keylock is on
13:30:17langhaarrockerNo prob with key schemes, though
13:30:30langhaarrockerJust define a scheme without key pad but with remote :)
13:30:51LinusNyup, as soon as you rewrite thw wps for key shcemes
13:31:27langhaarrockerIt's easy to make a dummy rewrite that really simply maps new keys -> old keys
13:31:41langhaarrockerBut that's dirty
13:32:05LinusNit's not the key mappings that are the problem in the remote case
13:32:18LinusNit's all the states in the wps code
13:32:57langhaarrockerWasn't it that key locking was already in button.c?
13:33:13LinusNbutton.c has no kaylock handling
13:33:30LinusNwps is the only place
13:33:59langhaarrockerWhat is BUTTON_LOCKED good for, then?
13:34:00LinusNwps stays in the keylock() function as long as the keys are locked, ignoring all keys
13:34:06LinusNlanghaarrocker: beats me
13:34:32LinusNBUTON_LOCKED is not used
13:34:57langhaarrockerI hoped it was something like a mask that masks out all locked buttons.
13:35:15LinusNthere were plans for that once upon a time
13:35:15langhaarrockerMaybe it's easier to kick out key lockíng from wps and in fact use BUTTON_LOCKED.
13:35:35LinusNlanghaarrocker: maybe
13:36:06langhaarrockerWere there plans for double click events, too?
13:36:16LinusNnot that i know of
13:36:44langhaarrockerI think it should be possible to do because buttons are evaluated with a tick_thread.
13:36:58LinusNdouble clicks is a very limited feature
13:37:19langhaarrockerwhy not go n-klicks?
13:37:26LinusNstill very limited
13:37:35LinusNthe first click still has to be evaluated
13:38:19langhaarrockerIt's just an idea in the very back of my mind.
13:38:25langhaarrockerBut other things first.
13:38:40LinusNthink of it, picture the NEXT key
13:38:58LinusNwhat function would be assigned for single vs double click for that in wps?
13:39:19langhaarrockerdouble pitch
13:39:51langhaarrockeror preview id tag of next song for 5 secs
13:40:41LinusNlanghaarrocker: ok, let's say you do that
13:41:12LinusNwhen exactly will the button driver decide that the event is a double click and not a single click?
13:41:23LinusNa timeout, yes
13:41:47LinusNthat means that the NEXT function will feel very slow, non-responsive
13:42:16langhaarrockerMaybe it shoud be that when you recieve the first click you can ask button.c wether it might become a double click. That way you don't have delay if you're not interested in double clicks.
13:42:39LinusNmost, if not all, systems handle double clicks by handling both events
13:42:59LinusNin Windows: single click selects, and double click executes
13:43:46langhaarrockerHrgh. Not even I was thinking of Windows 4 Jukebox
13:44:18LinusNthe timeout for double click would be about 0.5s
13:44:42LinusNso all keys that have double click functions attached to them will have a 0.5s delay for the single click function
13:45:22langhaarrockerI might come to the same result if I tried.
13:45:31LinusNyou will, trust me :-)
13:46:36langhaarrockerI started to unify the screens.c functionality
13:46:59langhaarrockerI wanted to demo calling a screen with key schemes.
13:47:19langhaarrockerWhen I've done that I will post something
13:47:26LinusNif we find a good application for double clicks, we will of course implement them
13:48:04LinusNlanghaarrocker: the screens is a good application for configurable keys
13:49:04langhaarrockerAnd since I was on them I of course prepared them so that they are easier to become user definable, as well. :)
13:49:22LinusNthen the user can choose what functions to have on his function keys
13:49:38langhaarrockerWorks already for shuffle
13:50:36 Join MrDjango [0] (
13:50:42langhaarrockerBut there are some serious update probs: When calling screens is unbound from the caller - who is to redraw the lcd when the screen quits?
13:51:26LinusNlanghaarrocker: the function that called your custom button code
13:51:39LinusNthe wps main loop for example
13:51:52langhaarrockerThat's the way it works now.
13:54:39langhaarrockerI'd prefer a solution where we bind a key to show_screen(n) instead of tree_show_screen(n).
13:55:02LinusNyeah, but that's no problem, is it?
13:55:46LinusNyour code must still be called from somewhere
13:55:49langhaarrockerif we just have a method show_screen(n) that can be called from anywhere we don't know who is the caller that has to redraw
13:56:00LinusNthe caller redraws
13:56:09LinusNthe called function doesn't care
13:56:36langhaarrockermakes it more difficult do do partial redraws.
13:57:20LinusNlet the called function return true if a redraw must be done
13:57:27LinusNor something
13:57:39 Nick LinusN is now known as Linus|away (
14:07:34MrDjangoHi folks! Are you planning to include bindable keys with 1.4 ??
14:07:44langhaarrockerno way
14:07:55MrDjangoGlad to hear that :-)
14:07:58langhaarrockerTheres way to much work to be done for that
14:08:04 Quit Django (Read error: 110 (Connection timed out))
14:09:05MrDjangoI wonder who decides what goes into the code.
14:09:31langhaarrockerWen Zagor, LinusN and Bagder agree on something it goes in.
14:11:17MrDjangoIs there a scheduled release date for 1.4? xmas ?
14:11:45langhaarrockerI expect it to come up next week
14:11:45MrDjangoSanta Claus? :-)
14:12:05MrDjangoWill there be user documentation?
14:12:26langhaarrockerIf you want to write it you're welcome :)
14:12:28rad[idle]The daily builds work pretty well for me ;)
14:12:34MrDjangoI haven't found out yet what "play selected first" does.
14:12:40MrDjango.. just for example.
14:12:45langhaarrockerit in shuffle
14:13:26langhaarrockerWhen you have shuffle enabled and hit play on a song it either is played and then shuffled or a random song is played directly.
14:13:57MrDjangoJust read on the mailing list that there is a way to see ID3 info via the ON button.
14:14:14langhaarrockerhold f1 and press on
14:14:45MrDjangoI think we need a reference card as .txt file to look up all those 'secrets'
14:14:59langhaarrockerI didn't know that a few day either a few days ago. But than I stubled across it while reading the source. -> The best documentation ever.
14:15:14langhaarrockerWhy not have a full featured online help?
14:15:25MrDjangoTo much to read.
14:15:41MrDjangoNobody wants to read, they want to KNOW.
14:15:57MrDjangoLike playing piano with the learning effort :-)
14:16:07rad[idle]Hey, a reference card is a great idea!
14:16:10langhaarrockercontext sensitive. Holde some help key while pressing the button and you get info on what it does.
14:16:19 Nick rad[idle] is now known as radhard (
14:16:58langhaarrockerI want screen shots in the reference card. -> We must do a html browser first.
14:17:23MrDjangoEven more button presses? I wonder if the keypad is designed for heavy use. Maybe we will wear out those tiny switches soon.
14:17:48langhaarrockermy down button is already broken.
14:18:18langhaarrockerIf I was a better programmer I wouldn't have had to do that much testing while programming wormlet.... :)
14:18:52radhardI remember reference cards as small sheets of paper with all necessary information printed on it. Don't know if html Browser is really necessary... ;)
14:19:41langhaarrockerOk, what about a pdf reader then?
14:20:56adi|homeMrDjango: so.. what your suggesting is basically "hey guys.. offer documentation in _every_ format that _every_ user could ever want!"
14:20:56MrDjangoWe already have a text reader. Now find someone who is really familiar with rockbox, then make hime write the refcard.
14:21:09adi|homein other words.. "don't ever code, just document what you _want_ to code"
14:21:12CtcpIgnored 1 channel CTCP requests in 0 seconds at the last flood
14:21:12*adi|home shakes his head
14:21:45adi|homewe offer you a FAQ
14:21:48adi|homewe offer you a mailing list
14:21:52adi|homewe offer you irc
14:22:03adi|homewe offer you archived logs of all the above
14:22:13adi|homeand we offer you peacemeal docs off the web
14:22:18*adi|home sighs
14:22:25langhaarrockerSo why don't we offer the things users really want, then?
14:22:34radhardWell, i'm scared that once in a time i will find a directory named "Windows" in the root folder ;) That wont happen, no?
14:22:36MrDjangoadi: did I offend you in some way ?
14:22:50adi|homenah.. im just touchy about the document thing
14:23:03adi|homeim the one that wrote up 95% of the FAQ
14:23:13langhaarrocker(which is great)
14:23:21adi|homeso constantly getting questions that _are_ answered there has made me bitter
14:23:35adi|homenot that yours was ;)
14:23:42adi|homeits on my list of 'things to add'
14:23:49Linus|awaydocumentation is always a problem
14:24:11adi|homei blame Bjorn ;)
14:24:22adi|homeyou damn swedes and your damn fast coding
14:24:25Linus|awaywe all do :-)
14:24:53 Nick Linus|away is now known as LinusN (
14:24:58radhardYep, FAQ is very good. Oh, and while you are all here: Thank you very much for rockbox itself, it's great work!
14:25:05MrDjangoadi: I have have been thru the FAQ several times and it is great. Thanks for the work!
14:25:13langhaarrockerDocumentation will become worse when we really get key bindings. We'll have a list of hundreds of funcion snippets that can be assigned to keys.
14:25:16LinusNi code, and sometimes i find the time to document it
14:25:22adi|homewe were serious though guys... as you have questions and get them answered.. feel free to send them to the list or whatever as doc patches...
14:25:28adi|homeill jump on those pretty happily
14:25:35*adi|home smiles happily
14:25:44adi|homeit _really_ needs to be better organized.
14:25:59*adi|home smirks at LinusN
14:26:05adi|homei wouldn't call what you do 'code'
14:26:06*adi|home laughs
14:26:13langhaarrockerdocument structure or documentation process?
14:26:40LinusNlanghaarrocker: structure
14:27:14MrDjangoI understand the current doc is for 1.3. But the current CVS version has many additions and I haven't seen any infos on newly invented button combinations.
14:27:28LinusNMrDjango: exactly
14:27:37LinusNthe docs are lagging behind
14:27:55MrDjango.. which is normal for any project :-)
14:28:03LinusNthat's because we implement things we want, because it's fun
14:28:12LinusNand doumantation isn't fun :-)
14:28:15langhaarrockercool: I just found that the link to the wormlet game has been included.
14:28:29adi|homei dunno.. the humor in the FAQ has seemed to go over well ;)
14:28:41adi|homethough you guys keep editing out my biting sarcasm ;)
14:28:46MrDjangoBeing a software engineer I totally agree
14:29:04langhaarrockerDocs? read the code!
14:29:10*adi|home screams "I want a god damn job!"
14:29:12*adi|home sighs
14:29:19adi|homeoh.. i have a quote about that langhaarrocker .. hold on
14:29:39MrDjangoDo you think my wife should be able to get along with rockbox?
14:30:10langhaarrockerI just know rockbox, not your wife. (what was her name?)
14:30:33adi|homeI just know your wife, not rockbox. (what's it like?)
14:30:40*adi|home ducks
14:30:41MrDjangoAngelika. She is the perfect tester for any consumer product :-)
14:30:51adi|home<@Logan> I spent a minute looking at my own code by accident.
14:30:51adi|home<@Logan> I was thinking "What the hell is this guy doing?"
14:31:38 Part langhaarrocker
14:31:51MrDjangoI will show Rockbox to her and see what happens.
14:31:57 Join langhaarrocker [0] (
14:32:00 Nick MrDjango is now known as django|away (
14:32:07 Quit Lear ("ChatZilla 0.8.9 [Mozilla rv:1.1/20020826]")
14:34:01radhard;) nice quote! May i ask a short question on another topic? I didn't find anything in the doku...
14:34:21*langhaarrocker is giggling
14:35:19radhardHas there been any discussion about flashing the flash?
14:35:49langhaarrockerisn't the sh1 mask programmed?
14:35:58langhaarrockerI think there's no way to flash it.
14:36:15radhardi don't think so 'cause there is a flash part in the recorder....
14:37:39langhaarrockerI think I'd better shut up on that topic due to my severly limited knowledge.
14:41:45radhardWell i did not think about the controller having a mask rom. I'll check this out. But i don't think they use a 256K flash part for just sitting around ;)
14:41:52LinusNyes, the flash can be reprogrammed
14:42:01***Saving seen data "./dancer.seen"
14:42:02LinusNand yes, the controller has a mask rom
14:42:22radhardSo, did anybody try this?
14:43:01LinusNno, why?
14:43:14langhaarrockerrolo in flash? cool
14:43:16radhardFaster booting...
14:43:53LinusNand serious risk of screwing things up badly
14:44:03langhaarrockerWell, you should try it first.
14:44:16radhardOr maybe even rolo. With a little hardware mode (that i possibly could do) programming the flash should be possible.
14:44:25LinusNfirst of all, there is lots of additional code that needs to be written
14:45:07LinusNthe ROM code is tailored for that particular version of the hardware
14:45:12radhardWell, maybe i could do it. I think it's worth the risk. But then i'll need a little support.
14:45:44LinusNrockbox depends on the ROM code setting up the hardware correctly
14:46:23radhardAh, ok. But that should be possible. I think it's the refresh and the MAS-Parts. Anything else?
14:46:42langhaarrockerLinusN: have you thought about loop recording? Something like 1.) enter the loop duration 2.) starting the recording loop 3.) wait until the band has played the song right after 19 tries 4.) stop the recording.
14:46:42LinusNthe ATA too
14:47:22radhardHmm, ATA should not need much setup, does it?
14:47:31LinusNwhat are the loop and stop conditions?
14:47:31 Nick django|away is now known as Django (
14:47:56langhaarrockerloop is just like a ring buffer. Stop is when the user hits the stop button.
14:47:57LinusNradhard: the address encoding is dofferent between models
14:48:18LinusNlanghaarrocker: that is a file system problem
14:48:54LinusNthe USB detection is different between hardware versions
14:49:13LinusNthe recording DMA interface is different between hardware versions
14:49:40LinusNradhard: i'm not saying that it's impossible, just that is isn't all that easy
14:50:04DjangoJust did a user-interface-acceptance test with Angelika. The result is not encouraging :-)
14:50:16LinusNwe collect hardware information on all available models and versions
14:50:37LinusNDjango: not surprising. player or recorder?
14:50:43radhardWell, but that's why you already do different versions of rockbox. So it should be possible for all players. Sadly i only have a recorder here. Ok, if we talk about rev1,2,3... How many revsions are there?
14:51:00langhaarrockerIf there was a socket for the flash I'd still consider trying that. But I don't think there's enough space to solder in a socket.
14:51:30radhardThere are socket in that size. We have a lab that could solder them.
14:52:24langhaarrockerBut they're still higher than normal smds.
14:52:24DjangoThe good news is she likes Rockbox better than the Archos firmware :-)
14:52:28radhardDjango: When i first tried rockbox, i did an acceptance test on myself. It encouraged me very much! ;)
14:53:09radhardYepp, they are higher, but there is pretty much free space in my recorder...
14:53:37DjangoSo far she is only familiar with real jukeboxes. We have five of them, 2 Seeburg, 2 Rock_o_la's and 1 AMI.
14:53:48langhaarrockerI'm still buffled that there are enought jukebox owners in the world so that something like rockbox could grow.
14:54:27LinusNlanghaarrocker: we know of people whoo purchased Jukebox because of rockbox
14:54:34DjangoBut so far the natural jukebox functionality has been missed.
14:54:43langhaarrockerI thought it would be a small niche market.
14:55:09DjangoWith a real jukebox you drop in a coin and then select up to 6 songs to play.
14:55:21DjangoNeither Archos nor Rockbox do that!
14:56:18DjangoThat was the reason why I started working on the queue feature.
14:56:49radhardWhat, putting in a coin, type a number from the list and here a song? You'll need a booklet that is bigger than the player... ;)
14:57:04langhaarrockerLinusN: about loop recording: We might get around the file system problem if we specify the size of the ring in clusters instead of time.
14:57:27langhaarrockerAre there any cons that fit into the usb slot?
14:57:50 Join thebreaker [0] (
14:58:48LinusNlanghaarrocker: sure, we can redesign the file API for ring buffering, but i don't think it's that simple
14:59:12langhaarrockerI never have simple things in mind :)
15:00:04 Join elinenbe [0] (
15:00:12Djangolanghaar: do you have any plans for a logrithmic peak meter?
15:00:46elinenbeDjango: did you see my suggestion for the queue patch?
15:00:46langhaarrockerlogarithmics are no fun if you don't like math...
15:01:20Djangoelinenbe: yes, but the patch tracker is not the right place for discussions.
15:01:31Djangolanghaar: agree :-)
15:01:45elinenbeDjango: That is true.
15:01:59elinenbeDjango: How difficult would that be to implement
15:02:52langhaarrockerI hoped I could use the code that Paul Suade posted on the list. But it still uses floating point math.
15:03:01langhaarrocker(for logarithms)
15:03:18Djangoelinenbe: what do you mean? the F1/F2 stuff?
15:03:19LinusNlanghaarrocker: only for creating the tables
15:03:33langhaarrockernot only there.
15:03:33LinusNthe tables can be created compile time
15:03:42langhaarrockerI know that.
15:04:03langhaarrockerBut there was other fp stuff around.
15:04:04LinusNDjango: we will probably not accept the queue patch unless it can add to the already playing playlist
15:04:42radhardMaybe a log approximation function would do the job.
15:05:11elinenbeDjango: yeah
15:05:20langhaarrockerDo it and I'll use it happily :)
15:06:09DjangoLinus: That doesn't make sense to me.
15:06:09LinusNDjango: makes sense to a lot of other people
15:06:09LinusNand that is how the original firmware does it
15:06:34Djangoelinenbe: sorry, don't like that. It would drive me up the wall if I had to press another button.
15:06:54LinusNwhen people speak about the queue function, they speak about inserting song into the playlist
15:07:06DjangoLinus: what exactly do you want? Should the entire directory go on the list and then the just selected song??
15:07:46LinusNDjango: have you used the Archos queue function?
15:08:03DjangoLinus: the original firmware builds a playlist that can't be used immediately.
15:08:12LinusNit inserts the selected song after the currently playing one
15:08:20radhardlang: How much %error would go ok and what range are we talking about?
15:08:24DjangoLinus: give me file writing and I will do that.
15:08:34elinenbejango: but it should be at least an option to insert or append the song to the current playlist
15:08:50LinusNDjango: i'm not talking about the technical solution, i am talking about the feature in general
15:08:55DjangoLinus: queue playmode with archos also does not the playlist, no?
15:09:02LinusNi don't know
15:09:18DjangoLinus: what IS the current playlist?
15:09:35DjangoLinus: an .m3u file?
15:09:42LinusNeh, it's a buffer with filenames
15:09:44DjangoLinus: the current dir?
15:09:57LinusNm3u format
15:10:00DjangoLinus: where do the filenames come from?
15:10:14LinusNeither the m3u file or the current dir contents
15:10:27LinusNdepending on if you play an m3u file or the current dir
15:10:53DjangoI to leave for lunch now before Angelika gets mad at me :-)
15:10:54 Nick Django is now known as django|away (
15:11:46radhardlanghaar: Well, 5% should be ok because of the display resolution. Do you have the formula at hand? Alternatively the input and output range would be ok.
15:12:01langhaarrockerredhard: we're talking about converting a range of [0 - 0x800] -> [-90dBfs 0dBfs]. We need to calculate log for quite big values with low precision but with higher precision for small values. Two digit precison of the fractional part are perfect
15:12:32 Quit breaker (Read error: 111 (Connection refused))
15:13:09langhaarrockerformula is dBfs = 20 * log( sample / SAMPLE_RANGE)
15:13:39langhaarrockerSMPLE_RANGE is 0x800 and 0 <= sample < SAMPLE_RANGE
15:14:11radhardOk. I'll sit back and think about it....
15:14:12LinusNlanghaarrocker: 0x8000
15:14:53LinusNor rather 0-7ffff
15:14:56radhardOk, that sounds like 16 Bit smapling ;)
15:15:01langhaarrocker15 bit
15:15:22radhardAren't 8000-ffff the negative values?
15:15:33langhaarrockerno, all positive
15:15:47langhaarrockerIt's peak volume infos, not samples
15:16:16radhardSo i'll use 0-0x7fff with a SAMPLE_RANGE of 0x8000, right?
15:17:34 Nick radhard is now known as rad[thinking] (
15:18:17langhaarrockerthe formula was in log10 (base 10) You can refomulate that by
15:18:17langhaarrockerdBfs = 20 * ( log(sample) - log(SAMPLE_RANGE) )
15:18:17langhaarrocker = 20 * ( log2(sample) - log2(SAMPLE_RANGE) ) / ( log2(10) )
15:19:07rad[thinking]Ok, but i'll use no log at all...
15:20:25langhaarrockerJust because I have that link that I don't want to have to understand:
15:21:37rad[thinking]Yeah, logs can be very weird... ;)
15:25:19 Nick LinusN is now known as LinusN|cooking (
15:31:36rad[thinking]I'll think on....
15:32:12 Quit langhaarrocker (Read error: 104 (Connection reset by peer))
16:18:07 Quit elinenbe ()
16:42:05***Saving seen data "./dancer.seen"
16:42:41 Join ddb69 [0] (
16:43:57 Quit ddb69 (Client Quit)
16:56:27 Nick rad[thinking] is now known as radhard (
16:58:11radhardOk, i've got a solution that works with linear appoximation. The log-function can be splitted in 8-10 small linear function (only one needed at a time). It works. Now i have to figure out, how to do this in fixed point integer. Should be straight forward, but not today ;)
17:00:24radhardLinusN: Would that be ok to implement? Only integer, 3-4 comparisons and ohne linear function (y=a+n*x) that also works with int?
17:00:48radhardand one linear...
17:10:30 Quit datazone (Remote closed the connection)
17:15:50 Join Jet8810 [0] (
17:17:37 Nick seb-sleep is now known as _seb_ (
17:17:53 Quit thebreaker ()
17:20:59 Nick radhard is now known as rad[hard] (
17:21:17 Nick rad[hard] is now known as rad[idle] (
17:50:24LinusN|cookingrad[idle]: you there?
17:51:04 Nick LinusN|cooking is now known as LinusN (
17:51:06 Join langhaarrocker [0] (
17:51:17 Nick rad[idle] is now known as radhard (
17:51:18 Nick django|away is now known as Django (
17:51:28LinusNexplain it a little further
17:51:49langhaarrockerradhard: log solved?
17:52:27radhardI used am approximation algorithm (little VB program) on the LOG-function. So i figured out, that 8-10 linear Functions could be used.
17:53:23radhardNot fully. I converted it into 8 linear functions, that use float. Next task is to to the approximation with fixed point fload (int).
17:53:40LinusNradhard: do you have some code i could see?
17:54:14radhardNP. Where post?
17:54:21LinusNlinus at
17:54:31radhardDo you have the VB6 IDE?
17:55:01LinusNnope, i'll settle with the source
17:57:13radhardOk, it's out. But might look strange to you, if you don't know VB ;)
17:59:11 Quit langhaarrocker (
17:59:11 Quit Django (
17:59:11 Quit Schnueff (
17:59:11 Quit mem (
17:59:11 Quit LinusN (
17:59:35NJoinlanghaarrocker [0] (
17:59:35NJoinDjango [0] (
17:59:35 Join LinusN|walkdog [0] (
17:59:35NJoinSchnueff [0] (
17:59:35NJoinmem [0] (
17:59:53radhardWell, it's a fast hack, that might need some explanation. First, i shiftet the -90-0dB to 0-90dB. Also i left out 0, 'cause log(0) is undefined. Rockbox could assume 0.
18:01:02radhardAnd since VB uses log with base e, i hab to do a conversion to log base 10.
18:05:11radhardThe results are collected via RecordRange(). The Rockbox code would have to use a 10*4 int lookup table and finde the correct range by simple for (max 10 compares) or sukzessive approximation (like an A/D converter) with max 4 compares.
18:07:44DjangoI understand the input range is 0 to 0x7FFF. What is the output range?
18:08:11radhardOriginal: -90 to 0dB, me: 0-90dB
18:09:42DjangoThe output goes to the display. What pixel range is that?
18:09:42langhaarrocker-0,684159081 dBFs per pixel
18:09:42langhaarrockerLCD_WIDTH = 112
18:09:53DjangoOk, I try again :-) How many pixel do you use horizontally?
18:09:59radhardfrmmain.heigt and frmmain.width
18:10:28radhardi use 0x8000 "virtual" pixels ;)
18:10:29DjangoThe size of the peak meter can be changed?
18:11:06langhaarrockerThe range of peak meter must be changable, too.
18:11:31radhardDjango: Do you have the code? Wonder why you are asking, not linusn..
18:11:51langhaarrockerpricipially the size of the peak meter can be changed although I doubt that it will become relevant, soon.
18:12:03radhardIt could be changed by simple int-multiplication.
18:12:30DjangoI don't see why Rockbox has to calc logs. All you need is a predefined table
18:13:08langhaarrockerI'll modify the meter so that the range can be adjusted. Otherwise we'll have all the lower bits always on while at the interesting top end the resolution is insufficiant
18:13:14DjangoIs the size of the peak meter on screen fixed or can it be changed by the user?
18:13:28radhardAs said before, i would have to port the code to fixed point (int) calculation. 0-9000 for 0-90dB would fit, i think.
18:13:34langhaarrockerDjango: We don't want a table with 0x8000 entries
18:13:43 Join Remo [0] (
18:14:03Djangolanghaar: do you really think that was my suggestion? :-)
18:14:07langhaarrockerDjango: the size now is fixed. But the meter principially can be scaled.
18:14:58DjangoOk, how many pixels does the peak meter use for the output range? 112?
18:14:59langhaarrockerDjango: the values are spread so oddly that a simple map would affect the precision so heavily that is unusable for half the value range.
18:15:24DjangoThe precision is already limited by the number of horizontal pixels.
18:15:37DjangoNo need to be more precise.
18:16:15 Quit Jet8810 ("Client Exiting")
18:16:24radhardThere is a need. Just drae the log-function and you'll see.
18:16:39langhaarrockerThe peak meter reserves 3 pixel for clip display: meter width = 112 -3
18:17:39DjangoOk. So the actual output range is pixel position 0 to 108.
18:17:39 Quit mem (Read error: 104 (Connection reset by peer))
18:17:56langhaarrockerBut we mustn't rely on that. draw_meter takes width as parameter.
18:18:15langhaarrocker(no, it's called peak_meter_draw - but who cares)
18:19:04DjangoWhat is the parameter good for? Did Archos use LCDs with different resolution?
18:19:18langhaarrockerIf we want the peak meter not to use a complete line
18:20:03DjangoAll your problems will go away when you decide on a fixed size.
18:20:04langhaarrockerWe might want a small meter for the remote with lcd that seems to be about to be developed...
18:20:40langhaarrockerWe have to map linear values to logarithmic values. A simple table can't help us here.
18:21:03langhaarrockerOr only at the cost of an unacceptably large table.
18:22:49radhardlanghaar: What do you think, is 0-9000(int) appropriate for 0-90dB? Otherwise i have to use long.
18:23:20langhaarrockerIt is perfect. I'd be confident even with 0-900
18:23:54radhardI need higher resolution for the increment. y=n*x+m
18:24:54radhardn and m comes from a table with about 10 entries. Do you know how to implement sukzessive approimation to find the right pair to use?
18:25:31langhaarrockerI remember ad converter use that technique...
18:26:48radhardExactly. I'll give you a table with Xmin values and you would have to find out what n and m to use. Also possible to hard-code, then it's faster.
18:28:18radhardAlternatively you could help me to get the compiler up and running. Then i can implement it here. The short manual did not do the trick on my machine, sorry.
18:28:38langhaarrockerdo you use cygwin?
18:29:15radhardNo, i downloaded the executables and installers...
18:30:11langhaarrockerIt makes things much easier when you use cygwin because there are many rockbox scripts and makefiles that work with it right away
18:31:26radhardI'll check
18:31:34langhaarrockerTogether with the sh1 complier from it worked
18:32:15langhaarrockerYou don't have to install complete cygwin. Just make sure you get all the building tools.
18:34:55radhardDoes not seen they are offering a building tools packt. I'll have to d/l the whole stuff.
18:35:31langhaarrockerWhile installing you can choose what to install
18:35:33 Join mem [0] (
18:39:39radhardHmm, i don't like theese Internet Installers. But it offers d/l. Should be ok. Quite a lot of big packages ;)
18:39:46 Join hardeep [0] (
18:39:56langhaarrockerI know :(
18:41:18radhardlanghaarrocker: located in germany?
18:41:30langhaarrockeryes. Why?
18:41:46radhardJust interested. Me too.
18:42:08***Saving seen data "./dancer.seen"
18:42:28langhaarrockerJukeboxes spread all over the world...
18:43:20langhaarrockerradhard: so you have the same dull, gray, rainy weather?
18:43:55radhardYeah, and it wouldn't have been so widespread without rockbox ;) Yepp, and it's getting dark slowly. Clouds all over the sky :(
18:45:24langhaarrockerI only can imagine that people by an archos because of rockbox if they know linusn or someone else here personally
18:45:42 Join ddb69 [0] (
18:46:06langhaarrockerby := buy
18:46:13radhardI've been out running a few kilometers this afternoon. We got a little wet ;). Nobody i know uses the original OS.
18:46:38ddb69hi guys, anyone know when the official release of 1.3 will be?
18:46:47ddb69oops, I mean 1.4
18:46:57radhardHehe. No.
18:47:18radhardWas already looking for the site.. ;)
18:47:37 Join datazone [0] ([K6vBDd2eH@
18:47:42langhaarrockerI wouldn't be surprised if it happened by next weekend, though. But that's no promise and It's not me who is involved in that, anyway
18:48:05ddb69ok, thanks
18:48:53 Part ddb69
18:50:14radhardOk, i'll continue with coding. Maybe i could finish the problem this evening. If you really plan to implement it this way. Cygwin is at 30%, will take another while...
18:51:21langhaarrockerJust gimme some some log function and you'll have dbFS the next day...
18:58:14langhaarrockercu later.
19:49:10 Part Remo
19:54:52 Join panthar [0] (
19:55:29 Quit langhaarrocker (Read error: 104 (Connection reset by peer))
20:40:25 Quit hardeep (Remote closed the connection)
20:42:10***Saving seen data "./dancer.seen"
21:21:41 Quit datazone (Read error: 104 (Connection reset by peer))
21:22:12 Join elinenbe [0] (
21:25:20 Nick Django is now known as Django|away (
21:35:49 Join Phantom [0] (
21:52:15Phantomsomebody ?
21:53:03 Nick LinusN|walkdog is now known as LinusN (
21:54:40 Quit elinenbe (Read error: 110 (Connection timed out))
21:58:39Phantomwhat's about Star ?
22:01:23Phantomthe new game
22:01:52PhantomI love games
22:01:58Phantomlot,lot of games
22:02:04LinusNah, it isn't in yet, is it?
22:02:37LinusNit has issues
22:02:54PhantomI why "snake" was removed ???
22:03:00LinusNso the author has things to fox before we can use it
22:03:17LinusNZagor thought one snake gem was enough
22:03:37LinusNnobody continued developing Snake, so we removed it
22:03:37Phantomthat isnt true
22:03:48Phantomsnake was pretty good
22:04:01Phantomwhat can you add ?
22:04:14Phantomthis is a snake ./
22:04:26 Nick Synthe` is now known as Synthe (
22:04:59radhardaffirmative, it's one of the most addictive games also on Nokia phones
22:05:00Phantomis it possible to start a .lang on the Jukebox ?
22:05:19LinusNPhantom: no, you run .lng files
22:05:34Phantomwhat is the =/= ?
22:05:44LinusNi found Snake boring, dead slow.
22:06:07Phantomwhat is the =/= betwenn lang and lng ?
22:07:01LinusNlang are the source files, lng are the compiled language files
22:07:26Phantomok..and how to compile a lang file into lng ?
22:07:46LinusNnever done it myself
22:11:19PhantomIs there a place where I can find .lng ?
22:11:35radhardGot workink dB code. Uses integers and one long.
22:12:00LinusNintegers and long are the same on SH1
22:12:35radhardIf only i knew that before... No Problem. Works also with all longs.
22:14:01 Nick LinusN is now known as LinusN|away (
22:15:22 Join breaker [0] (
22:15:55radhardMaximum error is below 2%, that's pretty good for an integer appoximation. And it's already trimmed to high speed ;)
22:20:15PhantomIs there a place where I can find .lng ?
22:24:18PhantomIs there a place where I can find .lng ?
22:24:37radharddont know, soory.
22:24:38PsycoXultools/binlang apps/lang/english.lang apps/lang/<whatever>.lang <whatever>.lng
22:24:45PsycoXulright there :p
22:24:53Phantomand under windows ?
22:25:02PsycoXulinstall perl? :p
22:25:22PhantomI ve Cygwin
22:26:01PsycoXulperl tools/binlang apps/lang/english.lang apps/lang/<whatever>.lang <whatever>.lng
22:26:26PsycoXulmaybe \ instead of / ...
22:26:29PhantompsycoCul, You know very well the thing, so...can you send me a français.lng ? ;-)
22:26:43PsycoXuli've never actualy used it before...
22:26:46PsycoXulbut uh i suppose
22:27:13Phantomtell me if that work please
22:27:26PsycoXulMissing ID in apps/lang/francais.lang: LANG_REPEAT_ONE
22:27:26PsycoXulMissing ID in apps/lang/francais.lang: LANG_REPEAT
22:27:26PsycoXulMissing ID in apps/lang/francais.lang: LANG_REPEAT_ALL
22:27:27DBUGEnqueued KICK PsycoXul
22:27:28Phantomi will give you my email
22:27:39PsycoXulit needs to be updated
22:27:46PsycoXulbut those are all thats missing
22:27:46Phantomok, no prob
22:27:51PsycoXulit worked anyways
22:27:54PsycoXulit seems
22:28:05PhantomIwill send you what you must add
22:29:54Phantomid: LANG_REPEAT
22:29:54Phantomdesc: in settings_menu
22:29:54Phantomeng: "Repeat"
22:29:54DBUGEnqueued KICK Phantom
22:29:54Phantomnew: "Répéter"
22:29:54Phantomid: LANG_REPEAT_ALL
22:29:55***Alert Mode level 1
22:29:55Phantomdesc: repeat playlist once all songs have completed
22:29:57Phantomeng: "All"
22:29:59Phantomnew: "Tout"
22:30:01Phantomid: LANG_REPEAT_ONE
22:30:03Phantomdesc: repeat one song
22:30:05Phantomeng: "One"
22:30:07Phantomnew: "Un seul"
22:30:21Phantomcopy/paste that at the end of "français.lang"
22:30:28Phantomthat should work
22:31:30PsycoXulseems to
22:31:38PsycoXulcan you get dcc's?
22:31:55Phantommay bee, sometime that work
22:32:58Phantomthat should work
22:33:53PhantomI try...
22:35:01Phantomwe are good
22:35:58radhardHi! Has somebody set up the compiler up and working?
22:36:38PsycoXulradhard: most of us, afaik heh
22:36:52Phantom ? : why jukebox didn't remind lng and .wps after a reboot ?
22:37:09PhantomI must restart each..
22:37:47PsycoXulyou can name them default.wps and (i think) default.lng in the /.rockbox/ dir on the archos
22:38:11PsycoXulremembering the last selected ones is something on the TODO list
22:38:43radhardI wrote some code for rockbox and want to test it. I read the howto, installed complete Cygwin, installed compiler, tried FAQ and Howto again, but can't make it run. Never used gcc compiler befor, only embedded systems with own ide. Does anybody have a few minutes to help?
22:39:46PhantomI ve do the same thing
22:39:56Phantombut that is to hard to use
22:39:56***Alert Mode OFF
22:42:11***Saving seen data "./dancer.seen"
22:42:15radhardThat's no problem, if i only get it running the first time. i think.
22:44:07PhantomI don't understand : /.rockbox/
22:44:25Phantomhow to create a such folder ?
22:44:47Phantomwindows don't support "/" in folder name
22:45:12radhardPhantom: Also got the code. Maybe you could point me to important stuff to read? Windows also uses xxx.yyy in Folder names.
22:45:14PsycoXul/ is the directory seperator
22:45:21PsycoXulin windows terms that's \.rockbox
22:45:36breakerPsychoXul: Are you Psyco Dedman <>?
22:45:57PsycoXuljust open a commandline prompt, go to your archos drive (if it's x, type x: and enter), and then "mkdir .rockbox"
22:46:07PsycoXulbreaker: yeah
22:46:20breakerPsychoXul: I've just read your mail in the mailinglist.
22:46:27breakerI had the prob with the settings.
22:46:47breakerI tested it again and the prob is still there.
22:46:59breakerI have not default.cfg.
22:47:22breakerOnly a default.fnt and a default.wps.
22:47:43breakerAnd I selected "Reset settings" and made all my settings again.
22:47:59Phantomanother ? : where can I find new fonts ? "pre-compiled"
22:48:12breakerAll other values were saved correctly (contrast,...), but the sound is "flat" again after a reboot.
22:48:20PsycoXulhmm weird
22:48:48PsycoXulplayer or recorder?
22:48:58PsycoXulPhantom: i know nothing about fonts, i've got a player :/
22:49:44Phantombreaker : Use .EQ files to make a "good sound"
22:50:10PsycoXulmine saves just fine
22:50:10breakerI have soe .eq files. I just saw that bass was set to -12 !?! Wait a moment...
22:51:06PsycoXulon my player, if i don't do something that makes the disk access after changing the settings they don't get saved
22:51:20PsycoXuli.e. if i change them and then immediately shut it off, it stays at what it had saved previously
22:51:33PsycoXulwhich is why i asked if you have a player or recorder
22:51:44PsycoXulrecorder's save to the RTC instantly afaik
22:51:44breakerThat's because the player has no rtc and settings are written to disk only.#
22:52:58breakerIsn't the format "bass: 5
22:53:07breaker"bass: 5" and so on?
22:53:37breakerI have a .eq file with that and when I "execute" it the bass is set to this, but when I go to the sound menu, the bass is at -12.
22:56:15breakerSo now I see what's the prob. Loading the .eq file sets the bass to +5 (which is ok) and sets the bass value (accessible in sound menu) to -12. And -12 is saved to RTC and loaded the newt time.
22:56:31breakerSO I'll have a look at the .eq file loading code.
22:56:59PsycoXulhmm odd
23:00:26breakerYan you please make an .eq file with "bass: 5" in it, load it and tell me what the sound menu says?
23:00:44 Quit Phantom ()
23:00:55radhardHelp! I desperately want to install the compiler an get it working to test some code that i worked on half the day. Already read the Howto and FAQ and downloded and installed all necessary stuff, but can't make the tools. Anybody got some time to help?
23:01:35breakerPsychoXul: "treble: 5" works!
23:01:36radhardPointers to good doku also very appreciated!
23:01:53breakerradhard: Win or linux?
23:02:12 Join langhaarrocker [0] (
23:02:17breakerHm, I never tried to install the compiler under win. Sorry.
23:02:58radhardMaybe should work under Cygwin? Got it completely installed.
23:03:30breakerI don't know, but is the Cygwin only a compiler for the SH1?
23:03:40breakerAnd you need the tools compiler for Win/intel?
23:04:05langhaarrockerwith cygwin it is the easiest way
23:04:30radhardHi langhaar! Maybe you can help. The code is ready to get ported...
23:04:41radhardThank you breaker!
23:05:41radhard2% max error and uses only int/long. Should be very fast!
23:06:22langhaarrockergood. fast is even more important than accuracy
23:08:35 Quit Django|away ("User pushed the X - because it's Xtra, baby")
23:12:27breakerWhen "executing" a .eq file, the read values are not displayed anymore. Is this normal?
23:14:15breakerCan please someone with a recorder make an .eq file with "bass: 5" (value is not important), execute it and tell me what the sound setting (bass) shows? Mine always is set to -12!
23:15:14breakerAnd I can't see an obvious error in settings.c.
23:17:44langhaarrockerbreaker: I have the same result.
23:17:57breakerThat's interesting.
23:18:12breakerSo now the question is why this happenes.
23:18:27breakerThis is happening since I'd say three days or so.
23:18:34langhaarrockerBut I thought it was me because I just have been fiddling with persistence of global_settings
23:19:02breakerIn your version or did you commit something in cvs?
23:19:50langhaarrockerBut recently the repeat thing was expanded. That goes into global_settings which is stored to rtc / disk
23:20:15breakerBut that's not a problem of saving/loading from RTC I think.
23:20:22langhaarrockerWhen booting a crc is performed and if it fails it loads defaults
23:20:40langhaarrockerMaybe the defaults for the bass setting are faulty and the crc must be recalculated
23:21:16breakerLoading the .eq should set the bass value to the value that's in the .eq file. There is no CRC done there, is it?
23:21:26breakerAnd no save/load to RTC.
23:21:37langhaarrockerthe format of the .eq file changed a while ago.
23:22:19breakerNow it's "bass: 5" instead of "bass=5", but that shouldn't be the point. And for treble, it works.
23:26:55 Nick LinusN|away is now known as LinusN (
23:27:35LinusNrecorders save to RTC immediately, but it is also saved to disk
23:27:50LinusNand the disk info seems to take precedence
23:28:01langhaarrockerI think the mistake is in mpeg.c in the 1769
23:28:15LinusNthus, settings aren't properly saved until disk access
23:28:54LinusNlanghaarrocker: not, that is not the error
23:32:31langhaarrockeryes. I was mixing ups the functions _again_
23:34:47LinusNi just tried it on my recorder. it worked. is this a Player issue?
23:36:15langhaarrockerno. recorder. But my environment is to unstable to be reference here. I'm messing it all up.
23:37:36LinusNif atoi() fails, bass will be -12
23:39:05langhaarrockerAnd I thought atoi was improved
23:39:11LinusNit was
23:39:19langhaarrockermaybe double negation?
23:39:34langhaarrockersince the old didn't handle negative numbers?
23:39:57LinusNdid you try with a file with only "bass: 5" in it?
23:41:23LinusNand does the line end with CR/LF or only LF?
23:47:18LinusNi can't get it to fail
23:47:37LinusNcan you get it to happen with the daily build?
23:48:20breakerLinusN: I made a file with bass: 5 in it and it works. The same with a file with additionally treble: 5 in it.
23:48:40breakerPlease try That file does result in the -12 bass setting!
23:49:30breakerI just saw that I have a bass boost value of 50. Is it from 0 to 100 or 0 to 10?
23:50:29LinusNbass boost should be 0 - 10, afaics
23:51:45breakerI have also a flat.eq with all values set to 0. And there also the bass value (in menu) is set to -12.
23:52:15LinusNbass: 0 will give you -12
23:52:33LinusNbass: 0 _is_ -12dB
23:52:51PsycoXulis there any reason not to auto power off a player when idle while charging?
23:52:53LinusNhmmm, no
23:53:02LinusNPsycoXul: what?
23:53:24PsycoXuland powering off while paused seems kinda strange, might be annoying if you don't use resume?
23:53:50LinusNbreaker: your file doesn't set bass to -12
23:54:09LinusNPsycoXul: i can agree on the second point
23:54:23langhaarrockerWhen I load bass with set to 6 I hear that it is applied. When I then enter the sound menu it displays -12. When changing from -12 to -11 I hear that the bass is reduced.
23:54:25breakerLinusN: Hm? Here it is doing so. Please wait a minute.
23:54:28LinusNbut turning off while charging? why?
23:54:48breakerlanghaarrocker: Exactly the same here.
23:54:55PsycoXuli dunno, it just seems strange
23:55:02PsycoXuli guess its no big deal
23:55:11LinusNPsycoXul: why is it strange?
23:56:22PsycoXuli dunno... why's it need to stay on any time while idle for longer than the autoshutoff idle time? heh
23:57:22langhaarrockermy build is 12 bigger than the daily. :)
23:57:24LinusNPsycoXul: because we want it that way?
23:57:29PsycoXuloh ok
23:58:09LinusNi don't know, people tend to want it to stay on, because the auto-shutdown is a battery saving feature
23:58:26PsycoXulthats true...
23:58:32LinusNno need to save batteries when charging
23:58:51PsycoXuli was just kinda suprised when i woke up and the unit was still on this morning

Previous day | Next day