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 2009-11-11

00:01:03CIA-6New commit by kugel (r23604): Fall back to info vp from sbs when intersection fails (again, r23575 changed it despite it was agreed on info vp beforehand).
00:02:39Unhelpful"/home/chshrcat/build/rockbox/firmware/target/arm/s3c2440/crt0.S:558: Error: symbol vectors is in a different section" <- building eabi for beaft. would a reasonable solution be to make .vectors a subection of .text? can i still make sure that things go into their proper places that way?
00:04:05 Quit scorche (Nick collision from services.)
00:04:08 Join Riku [0] (
00:04:49 Quit Lss (Read error: 131 (Connection reset by peer))
00:04:50 Quit evilnick_B ("Page closed")
00:04:51 Join scorche [50] (n=scorche@rockbox/administrator/scorche)
00:05:28 Quit ender` (" Two things are infinite: the universe and human stupidity, even though I'm not yet sure about the universe. -- A. Einstein")
00:06:36 Join kugel [0] (n=kugel@rockbox/developer/kugel)
00:07:20kugelUnhelpful: I changed to to non-pc relative adressing, that worked on the mini2440 as far as i can tell
00:08:08Unhelpfulit doesn't seem that we should need to do that, though?
00:08:41Tornethe vector address is fixed, i dont' see why you should need a symbol reference for it tbh ;)
00:09:00kugelwell, it loads to lr, isn't that supposed to contain a absolute address anyone?
00:09:05Bob_Cwhy doesn't it just jump to start annyway?
00:10:17Unhelpfuldeclaring the vectors symbol in the .lds would also work, wouldn't it?
00:10:27Bob_Cor return to start in this case
00:11:34Bob_Cor maybe set lr to PC+4
00:13:06 Join Mike_lifeguard [0] (n=Mike@wikimedia/mikelifeguard)
00:14:49 Part Mike_lifeguard ("Mike is a four-letter word.")
00:17:25Bob_Cand I don't see why vectors is a separate section either
00:20:21TorneUnhelpful: yah, non-section-relative symbols shoudl be fine
00:20:39Torne(i haven't actually looked at this code)
00:21:40Bob_CThere already is a symbol "vectorstart" in the lds btw
00:21:42 Quit daggett ("Ex-Chat")
00:23:05 Join GeekShado_ [0] (
00:26:46Unhelpfulso using .text.vectors could be one solution, or vectorstart...
00:27:27 Quit domonoky1 (Read error: 131 (Connection reset by peer))
00:33:38 Quit BBBradley ("CGI:IRC (EOF)")
00:37:10 Quit mt (Read error: 113 (No route to host))
00:39:22 Quit kugel (Remote closed the connection)
00:39:57 Quit Grahack ("Leaving.")
00:40:43 Quit GeekShadow (Read error: 113 (No route to host))
00:52:06 Quit mikroflops (Remote closed the connection)
00:52:07 Quit antil33t (Read error: 104 (Connection reset by peer))
00:52:31 Join mikroflops [0] (
00:57:32 Quit mikroflops (Remote closed the connection)
00:57:53 Join mikroflops [0] (
00:58:37 Quit Omlet ("( :: NoNameScript 4.22 :: )")
00:59:25 Join antil33t [0] (
01:00:09 Join GeekShadow [0] (
01:02:42 Quit GeekShado_ (Read error: 60 (Operation timed out))
01:05:24 Join Strife1989 [0] (
01:09:12 Quit pixelma (Nick collision from services.)
01:09:12 Join pixelma_ [0] (i=quassel@rockbox/staff/pixelma)
01:09:16 Quit amiconn (Nick collision from services.)
01:09:18 Join amiconn_ [0] (i=quassel@rockbox/developer/amiconn)
01:09:32 Nick pixelma_ is now known as pixelma (i=quassel@rockbox/staff/pixelma)
01:09:38 Nick amiconn_ is now known as amiconn (i=quassel@rockbox/developer/amiconn)
01:13:03***Saving seen data "./dancer.seen"
01:14:17 Quit Tuplanolla (Read error: 110 (Connection timed out))
01:14:29 Quit Stephen__ ("Leaving")
01:16:40 Quit mikroflops (Read error: 54 (Connection reset by peer))
01:17:39 Join Blue_Dude [0] (n=chatzill@rockbox/developer/Blue-Dude)
01:18:39Blue_Dudeevilnick: still here?
01:18:46 Quit Rondom (Nick collision from services.)
01:18:57 Join Rondom [0] (
01:20:34 Join Addison__ [0] (
01:21:29Blue_DudeI was unhappy with the crossfade logic so I actually made some changes. There's a new option for crossfade: auto track skip only. It only crossfade for automatic track changes, but not for manual ones.
01:22:52 Quit Strife89 (Read error: 110 (Connection timed out))
01:22:56Blue_DudeI was also unhappy with the crossfade decision logic. It was a mess, so I changed it to a much clearer switch/case tree. The svn update is coming shortly, after fixing the manual.
01:22:59 Quit GeekShadow (Read error: 113 (No route to host))
01:31:01 Join uflops [0] (
01:31:16 Quit killan_ ("( :: NoNameScript 4.22 :: )")
01:32:11 Quit goffa (Read error: 113 (No route to host))
01:34:08 Join StealthyXIIGer [0] (
01:36:00 Nick Strife1989 is now known as Strife89 (
01:36:51rjgBlue_Dude: that sounds good, I have had some issues with the cross fade when skipping tracks a lot
01:37:33Blue_DudeWe'll see how well this works. I'm just finishing up the manual now, and I'll commit in a few minutes.
01:37:54 Quit addisonj12 (Read error: 110 (Connection timed out))
01:38:00rjgwhile you're here, any thoughts on this?
01:38:13rjgI'm unsure if I should be logging it as a bug on the tracker or not
01:38:47 Quit AaronM (Remote closed the connection)
01:39:44 Join MethoS- [0] (n=clemens@
01:39:58rjgsomething has changed between recent builds and the preivous one I was using from some months ago, something to do with what it does with the WPS while the screen is off - seems to hault it now causing that buffer of changes to spew out at once
01:41:03rjgI'm unsure if it's something which should be fixed in the DGT themes or if it's a rockbox issue
01:41:11 Join AaronM [0] (
01:42:09 Part toffe82
01:43:54rjgI don't want to waste developers time by posting rubbish in the tracker
01:44:39Blue_DudeThe skin code has undergone so many changes in the past few weeks, I can't even keep up. It's very actively in development right now. It broke legacy skins a while back so I'm not surprised that you're seeing issues. Make sure you're using a current theme.
01:45:39rjgyeah, DGT had an update fairly recently when there was a major change to the theme code which broke it completely
01:45:51rjgI'll have a hunt around for any updated versions
01:46:33 Quit StealthyXIIGer (Read error: 131 (Connection reset by peer))
01:46:39Blue_DudeThey redid a lot of the theme tags so unless a theme has been very recently updated, it's probably not going to work as you'd expect.
01:46:40rjgI guess if the worst comes to the worst I'll attempt to modify the old version to work with newer builds
01:46:43 Join StealthyXIIGer [0] (
01:46:51rjgBlue_Dude: ok, thanks
01:47:12Blue_DudeSure. Ran into that problem myself not long ago.
01:48:21CIA-6New commit by blue_dude (r23605): Crossfade: added a new option, rewrote decision logic, updated manual and menus. Translators please note that updated translations may be required ...
01:48:43Blue_DudeThis ought to shake the tree some. :)
01:49:05 Quit Thundercloud (Remote closed the connection)
01:50:47 Quit StealthyXIIGer (Read error: 104 (Connection reset by peer))
01:50:54rjgI'll try it out tomorrow
01:50:55 Join StealthyXIIGer [0] (
01:53:41Blue_DudeSome devs hate, hate, hate it when you modify the lang files, but I think I did it right this time. Everything was changed at once.
01:54:11rjg −− * Latest Syntax DGTV, 1.5 I guess ? - StephenCarroll - 22 Sep 2008
01:54:16rjgsep 2008 is definitely too old
01:56:59Blue_DudeHa, all green. I'm outta here. I'm going to hide until the smoke clears. :)
01:57:53 Quit Blue_Dude ("ChatZilla 0.9.85 [Firefox 3.5.5/20091102152451]")
01:59:41 Join robin0800 [0] (
02:00:26togeticcan you increase playback speed?
02:02:38 Quit Llorean ("Leaving.")
02:02:38evilnicktogetic: Yes you can
02:02:48evilnickIt's in the manual
02:13:45 Join JdGordon|_ [0] (
02:23:54togeticevilnick: the only thing i'm seeing in the manual is timestretch, but i don't see it anywhere in my sound settings
02:25:12evilnickWhat player do you have?
02:25:26togeticevilnick: ipon nano 1st gen
02:25:48togeticrockbox version 3.3
02:26:38evilnick and look for Pitch Screen
02:27:22evilnickHmmmm, it might be worth upgrading to 3.4 as I'm not 100% sure that timestretch is in 3.3
02:27:30gevaertsI don't think 3.3 has timestretch
02:28:12togeticcan i upgrade without removing any of my files?
02:28:18togeticand can i upgrade using linux?
02:28:29togetici've had to install rockbox using windows every time
02:28:43evilnickBut you can either simply speed up the song and the pitch will go up, or use timestretch and affect the speed but not the pitch.
02:29:18evilnicktogetic: Check out rbutil, once you have RB on your player, you can use any program that can write to your player
02:29:25togeticcan you speed up the audio without using timestretch, with a feature that is in 3.3?
02:29:47gevaertsupgrading should be easy
02:30:00evilnickgevaerts: Pitch will speed it up, I thought.
02:30:19gevaertsyes, but the pitch will go up, which I assume is not desired
02:30:37 Quit AaronM (Remote closed the connection)
02:32:42gevaertstogetic: if you have problems installing from linux, you're likely doing something wrong
02:33:01 Join AaronM [0] (
02:33:01 Join AEnima1577 [0] (
02:34:09 Join goffa [0] (n=goffa@
02:34:29togeticdo i select 'complete installation' to upgrade?
02:35:07gevaertsyou don't need a new bootloader, but it won't hurt either
02:36:17togeticgrr, it can't open the ipod, so i'm told
02:37:06gevaertsrbutil may need root rights
02:37:09togeticdo i need to mount it first and then select /mnt/ipod
02:37:20togeticor do i need to not mount it and select something like /dev/sdb2
02:37:51togetici already had it running as root
02:37:53gevaertsdoesn't the autodetection work?
02:37:58togeticstill do
02:38:19togeticyeah, but it doesn't write
02:38:56gevaertsdoes dmesg mention anything about FAT errors?
02:39:26 Join killan [0] (
02:40:18togeticgevaerts: yes it does
02:41:11gevaertsok, so it probably remounted the device read-only
02:41:34gevaertsyou'll need to fix that first
02:41:45togeticit is what?
02:42:18togetic│20:33:46 gevaerts | ok, so it probably remounted the device read-only
02:42:25gevaertsthis is not a rockbox problem. Your ipod has a corrupted filesystem, and when linux detects that it switches to read-only to avoid further damage
02:42:58togeticso to fix this, i wipe the drive with fdisk and then install rockbox?
02:43:48*TheSeven is observing a very weird phenomenon
02:43:56gevaertsuhm, no. (a) fdisk does not wipe drives, and (b) fsck.vfat sounds like a better option. First try to repair
02:44:16TheSevenif i bypass the HCLK=>PCLK factor 2 clock divider, SDRAM access slows down by a factor of 2
02:45:16TheSevensetting an undocumented bit just sped it up by a factor of 4!
02:45:44gevaertsimagine what could happen if you set a documented bit!
02:46:26TheSevenheh, the documented bit for the clock divider isn't even there
02:46:49TheSevenwe thought it was just shifted by one for some reason (both on 8700 and 8701, so probably just a bug in the preliminary ds)
02:47:06TheSevenoh wait a sec
02:47:13TheSeventhis is influencing my timer readings
02:47:30 Quit JdGordon|_ ("Miranda IM! Smaller, Faster, Easier.")
02:47:32 Join saratoga_ [0] (i=9803c264@gateway/web/freenode/x-uichamwndznxnaqn)
02:47:38TheSevenoh yes, these are based on pclk
02:48:12TheSevenbut nice to know that we can switch between clock divider factors of 1/2/4/8 while only 1/2 is documented
02:53:33TheSeveni can't make it work in synchronous mode though
02:54:08TheSevenasync and fastbus both work fine, but sync doesn't
03:00:43 Quit Riku (Read error: 104 (Connection reset by peer))
03:03:02 Quit MethoS- (Remote closed the connection)
03:07:22TheSevennorboot is running it in async mode when i take over, even though they run at 192/96/48MHz, so sync should be possible
03:10:41TheSevenhm, "FCLK must be HIGH whenever there is a BCLK transition."
03:10:52TheSeventhat's the only thing i could think of that may be violated
03:11:48TheSevenbut the penalty of async vs. sync shouldn't be high anyways
03:13:04***Saving seen data "./dancer.seen"
03:24:31saratoga_are codecs any faster with a faster memory bus?
03:24:53TheSevenwe're already driving the AHB at it's maximum speed
03:25:22TheSevenbut yes, memory access performance seems to be proportional to AHB speed
03:26:01TheSevenliar: ping
03:26:49TheSevensaratoga_: judging from the datasheet, I would expect HCLK = SDRAM_CLK
03:27:22saratoga_TheSeven: isn't there a divider in there somewhere? or do you mean thats disabled?
03:27:53TheSevenif the datasheet is right, there is an optional factor 2 divider which is off
03:28:13TheSevenhowever, the datasheet seems to be inaccurate in the MIU area
03:28:34TheSeventhere are some regs that don't match at all, but MIUCON seems to match
03:28:47TheSeven(and that's where that clock divider bit is)
03:30:12TheSevenbtw apple's bootloader is clocking the same way as rockbox
03:33:26 Quit robin0800 (Remote closed the connection)
03:34:03TheSevenliar: I finally have an idea where your USB problem could be coming from, or rather, I'm wondering why it works for me.
03:34:41 Quit DerPapst ("Leaving.")
03:35:29TheSevenUSB doesn't seem to like an AHB clock <66MHz at all. We are the only guys who are using boosting builds, right?
03:37:16TheSevenyou maybe want to have a shot at increasing the boosting min clock to 63897600Hz (CLKCON=0x32803280/CLKCON2=0x80)
03:38:36TheSeven47923200Hz doesn't work at all for me if I do it in iBugger, and I'm still wondering why it works within rockbox. Or is it just always running at full speed while USB is connected? Who knows...
03:40:31TheSevenbe aware that the 66MHz clock will result in timer inaccuracy (32% too fast when going to min speed)
03:41:19TheSevenso in the long term we may want to go for 95846400Hz instead, or just always run at full speed if USB is connected, as we don't need to care about power in that case anyway
03:46:25 Quit goffa (Read error: 104 (Connection reset by peer))
03:48:34 Quit uflops (Remote closed the connection)
03:48:57 Join uflops [0] (
03:50:12 Join goffa [0] (n=goffa@
03:54:46*TheSeven is going to sleep now
03:56:33 Join T44 [0] (
03:59:38 Quit uflops (Read error: 104 (Connection reset by peer))
04:01:07 Join uflops [0] (
04:04:56 Join BHSPitMonkey [0] (n=stephen@unaffiliated/bhspitmonkey)
04:07:58 Quit Rondom (Nick collision from services.)
04:08:08 Join Rondom [0] (
04:14:58 Quit Topy44 (Read error: 113 (No route to host))
04:31:49 Quit TheSeven (Nick collision from services.)
04:32:07 Join The_Seven [0] (n=theseven@rockbox/developer/TheSeven)
04:32:19 Nick The_Seven is now known as TheSeven (n=theseven@rockbox/developer/TheSeven)
04:35:56 Join Bob_C_ [0] (
04:35:57UnhelpfulTorne: so it seems to me the relevant thing would be DVFEN(bit 4) in PMCR0? and possibly FSVAIM (bit 15) "This bit masks the DVFS frequency adjustment interrupt."
04:36:23Unhelpfulit sure looks like you can say "manage voltage and frequency for me, and no, i don't really care to know when you change things"
04:46:06 Quit phanboy4 (Read error: 54 (Connection reset by peer))
04:54:45 Quit Bob_C (Read error: 113 (No route to host))
05:13:05***Saving seen data "./dancer.seen"
05:15:08 Quit AEnima1577 (Client Quit)
05:30:52 Quit Horscht ("Verlassend")
05:35:45 Join phanboy4 [0] (
05:36:34 Join Llorean [0] (
05:45:00 Quit StealthyXIIGer (Read error: 145 (Connection timed out))
05:45:06 Join AEnima1577 [0] (
05:55:51JdGordonbrain fart time....
05:56:08JdGordonwith the removal of the inbuilt statusbar there is no need for the statusbar config option
05:56:32JdGordonbut, what do we do for screens which want to force the bar?
05:56:57kkurbjunJdGordon, is it intended that the custom statusbar will show in plugins?
05:56:59JdGordonThe obvious answer is to supply a hardcoded one... but then how do we have the user tell us if he wants the hard coded one or none?
05:57:05JdGordonkkurbjun: its not
05:57:19kkurbjunJdGordon, I think some status is needed in the plugins
05:57:28Lloreankkurbjun: Few plugins have it now.
05:57:34JdGordonthey can enable it if they want
05:57:38kkurbjunthere needs to be a way you can get battery readouts and the like
05:57:46JdGordonalmost none have it.. and only in their lists/menus
05:57:49kkurbjunthey used to all have it in the menus
05:57:54JdGordonwhere it should magivally pop up anyway
05:57:57kkurbjunyeah, that's what I'm referring to
05:58:00LloreanCan't plugins implement their own status indicators, since it should be themed to match the plugin rather than the WPS anyway?
05:58:16JdGordonplugins will be fixed once the core is working properly
05:58:24kkurbjunthe plugins all use the standard menus I think they should look as integrated as possible to rockbox
05:58:55LloreanEven when they use the standard menus, they often use their own color scheme or what not.
05:59:10Lloreanif you preserved the backdrop image and custom status bar, a plugin choosing to use its on colors could be a disaster and unusable.
05:59:30*JdGordon really hates calling the skinned bar the statusbar :)
05:59:42kkurbjunThey should not choose their own colors in the menu, and the backdrop should be preserved in menu
05:59:57JdGordonhow do you guys feel about having a blank .[r]sbs named "rockbox-none" which would have the effect of disabling the sbs?
06:00:05Lloreankkurbjun: So theme authors can never have their menus match their plugin unless they want an entirely custom menu (which we frown on)?
06:00:39kkurbjuntheme authors plugins?
06:00:40 Quit panni__ (Client Quit)
06:01:07kkurbjunJdGordon instead of a menu option to disable the statusbar?
06:01:18JdGordona disable option is redundant
06:01:37Lloreankkurbjun: Sorry, plugin authors.
06:01:51JdGordoneither a sbs is loaded, or its not
06:01:54kkurbjunnone of the plugins have menus that are particluarly tailored to the plugin
06:02:09*JdGordon thinks even rockbox-none isnt needed
06:02:11Lloreankkurbjun: Jewels uses the same background and text color in the menu as it does in the game.
06:02:15LloreanThis looks rather nice
06:02:25JdGordonbut not the background image
06:02:39LloreanJdGordon: Yes, but using the theme's background image there could be horrible, if it's a light image.
06:02:55JdGordonbut thats up to the plugin authour to disable
06:03:05kkurbjunLlorean, which menu are you referring to in jewels?
06:03:18LloreanJdGordon: Yeah, but kkurbjun's position appears to be that plugin authors should not be allowed to choose - the plugin menus should always match core menus in appearance.
06:03:39JdGordonin the menus and in the plugins UI are two seperate things imo
06:04:03Lloreankkurbjun: Has it changed?
06:04:44LloreanI think plugin authors should be allowed to decide *not* to match the core UI theme, rather than being forced to be constrained to it if they wish to use a list.
06:05:04kkurbjunIf a plugin went to the effort to provide it's own backdrop in the menus I could see allowing them to pick their own colors, but unless they went all the way I don't think that plugins should just pick arbitrary colors on a solid background
06:05:09LloreanMaybe simply give them the tools to create their own "list + sbs" within the plugin itself.
06:05:27Lloreankkurbjun: So it's only an artistic choice if there's an image involved?
06:05:57kkurbjunLlorean, the menu's are hardly artistic now
06:06:02kkurbjunthey are just plain jane
06:06:22LloreanYou have a very narrow view of artistic if you think "having it be the same colors as the rest of the application" isn't an artistic consideration
06:06:27kkurbjuntake a look at the jewels menu in a recent build
06:06:35LloreanYou want the menus to "belong" to Rockbox's look, and the author might want them to "belong" to the plugin's look instead.
06:06:37JdGordoncan I but in here for a sec? :D
06:06:57kkurbjunwhat plugin author are you referring to?
06:06:57Lloreankkurbjun: simple isn't the same as "not an artistic choice"
06:07:02kkurbjunare you an author?
06:07:15JdGordoneither a sbs is loaded or it isnt.... would a user ever be expected to actually *want* to use the inbuilt skinned bar (assuming its uber crap and we supply good replacements?)
06:07:21Lloreankkurbjun: is this the "nobody's asking for it right now, so surely nobody could want it" argument then?
06:07:52LloreanJdGordon: Can we have plugins default to using the list and built in SBS, but be able to use their own alternative via some method?
06:08:13JdGordon(in theory)
06:08:42kkurbjunLlorean: the menus are not integrated in the plugins right now - it is a complete separation from the game or app or whatever as it currently stands
06:08:45LloreanThat seems the most ideal way. If they don't explicitly say "override settings" they get whatever the core settings are for their lists, if they do override they can (and must) provide their own alternative
06:08:48JdGordonthe skin buffer needs to be able to be told to use a different buffer to do that, but its doable
06:09:00kkurbjunmaintaining the rockbox look in those cases seems reasonable
06:09:38kkurbjuniff they went to the effort to make something more custom I don't see a reason not to allow that, but until that is done I don't see a reason to break the standard rockbox look
06:09:43Lloreankkurbjun: Forcing it seems unreasonable though.
06:09:44JdGordon*very* few plugins have a need to keep close to the rockbox look
06:10:18kkurbjunnearly all of the plugins have an almost uniform look
06:10:26kkurbjunteru did a good job with that
06:10:29Lloreankkurbjun: That's because the list widget is limited right now.
06:10:44kkurbjunthe plugin's menus that is
06:10:48LloreanThey have a uniform look because they aren't offered options for having a non-uniform look without ceasing to use the built in list.
06:11:08LloreanMany plugins *had* custom menus that looked much, much nicer than the existing ones.
06:11:09kkurbjuntehy could do a uniform look
06:11:16*JdGordon would like to see the list being more skinnable soon also
06:11:21kkurbjunit is completely possible for a plugin to set it's own backdrop
06:11:33kkurbjunI experimented with it in the doom menus
06:11:34LloreanA backdrop isn't a significant change in look.
06:12:39kkurbjunLlorean: the plugins old menus were barely usable across targets and quite a few did nasty things like force the builtin font
06:12:47Lloreankkurbjun: Yes.
06:12:50LloreanYou keep missing my point
06:13:05LloreanThey offered *artistic* options which were taken away for a good reason - so that they could function better.
06:13:15LloreanNow we can return the artistic options without removing the functionality (in theory) and should
06:14:16kkurbjunyou can't make universal skins that work on all targets currently. Forcing a new port to port all of the custom skins for a new screen size seems insane
06:14:36kkurbjunif we had something more relative for skins I would agree that more flexibility should be allowed
06:14:37JdGordonyou asctually can :)
06:14:40LloreanThey have to port all the plugins to a new screen size anyway.
06:14:53kkurbjunnegative values are somewhat relative
06:15:03LloreanSeriously, if you need to resize the bitmaps to 32% larger and then tweak a few numbers, resizing all the windows and fonts is hardly much harder.
06:15:11kkurbjunbut it's not relative enough that you could go across all targets
06:15:28kkurbjunLlorean: have you ever gone thorugh it on a port
06:15:39kkurbjunit's awful already and you want to make it worse
06:16:04Lloreankkurbjun: So shouldn't we get rid of themes and plugins entirely then?
06:16:07LloreanIt'd make it much, much easier.
06:16:20*JdGordon is going to make the inbuilt sbs stupidly simple.... just battery, playback status and RTC... anything else *really* required?
06:16:32LloreanJdGordon: Volume
06:16:40LloreanYou always need to know if pressing "play" is going to deafen you
06:16:42JdGordonbaring in mind this will almost never be shown
06:17:49kkurbjunLlorean: if you want to volunteer to change all the skins that plugin authors come come up with for a new screen size I'm all for it - but we should make porting a new target/screen harder than it already is
06:18:04kkurbjunshould not that is
06:18:10Lloreankkurbjun: Plugins are not required for a port to be usable.
06:18:16LloreanSo, frankly, if they're harder it's not a big deal.
06:18:46LloreanPlugins on the other hand are a "nice bonus" and should be allowed to look as nice as their authors or maintainers want them to look.
06:19:21LloreanIf you don't make the default list flexible, as the core starts looking nicer and becoming more flexible with its menus, people will start simply coding custom menus for the plugins again so they can get that same sort of flexibility
06:19:38kkurbjunLlorean: that's a ridiculous argument, seriously try it some time, pick an arbitrary screen size and look at the amount of work required already.
06:20:35kkurbjunit's a little better with some plugins now than it used to be - it could be alot better if good decisions and planning was done
06:20:47LloreanWhat, drop bitmap support in plugins entirely?
06:20:55LloreanRemove absolute positioning, making all drawing relative?
06:21:10LloreanIt's always going to be a lot of work, but that's not a reason to restrict authors arbitrarily.
06:21:12kkurbjunLlorean: I don't mind bitmap rescaling, its all the other hard coded numbers you have to change
06:21:20kkurbjunand understand what they are being used for
06:21:22LloreanDo you honestly believe people will NOT create custom menus, given that they did in the past?
06:21:58LloreanThemes are easy to resize. Very easy. It's a couple minute's work after the bitmaps are done, especially if it's an actual skin and sbs file.
06:22:14LloreanIt's not like you need to "figure out what the numbers are used for", the theme syntax isn't terribly bad.
06:23:00kkurbjunLlorean: till you start contributing code I have a hard time seriously considering your argument for making development harder.
06:23:22LloreanThat's frankly a bullshit excuse.
06:23:36LloreanI'm allowed to have an opinion whether or not I've contributed any code recently (or at all for that matter)
06:23:55LloreanYou on the other hand seem to be thinking that creating skins is some enormously difficult task.
06:24:11LloreanHow many themes have you actually made or resized? You could probably automate it and then correct it if it looks funny.
06:25:20LloreanAm I allowed to say "I won't take you seriously until you show me some themes you've resized" or does that argument only work when you say it (even though this isn't about adding new C or ASM tasks)
06:26:05kkurbjunyeah, you can say that just fine and I will point you to the mr500 theme
06:26:11LloreanSo you've resized one?
06:26:15LloreanAnd how long did it take you?
06:26:26LloreanWas it your first theme? How long would it take you to do a second one?
06:26:28kkurbjunyes, I made scripts to automate making icons of all different size
06:26:34kkurbjunfrom svgs
06:26:48kkurbjunbut it's not something that I would argue for lightly
06:26:58 Quit Strife89 (Read error: 110 (Connection timed out))
06:27:04LloreanHow hard would you say it is to take an existing WPS and adjust it to a new target size?
06:27:31LloreanBecause that's what we're talking about doing here basically. An SBS and a list viewport.
06:28:52kkurbjunif you are talking the scale of all the plugins I would say a big pain - especially when you consider how complex some themes can get
06:29:01kkurbjunif you had relative positioning I would be all for it
06:29:13LloreanDo you think the plugin is going to be displaying the WPS or something?
06:29:19kkurbjunwith a well thought theme you could hit all the targets
06:29:22LloreanSBS and list viewport.
06:29:25kkurbjunwhy not?
06:29:42LloreanIf you can do relative positioning, then you can also do automatic scaling
06:29:44kkurbjunyou could have an sbs displaying everything the wps has
06:29:44LloreanSo problem solved
06:29:49LloreanA simple script, no work for the plugin authors at all
06:29:57LloreanIt's the exact same math.
06:29:58JdGordonis "char mytext[NB_SCREENS][] = {"blaaa", "fdslakhfsalkjh"};" correct?
06:30:02kkurbjun:-D agreed
06:30:10LloreanSo you don't need relative positioning
06:30:29kkurbjunsorry, thought you were speaking relative positioning
06:30:30LloreanAnd you can't do relative positioning for all targets anyway
06:30:39Llorean1:1, 4:3, 3:4, 16:9 and other aspect ratios exist
06:30:53LloreanYou'd need to create new variants as new screen dimensions existed.
06:31:08LloreanYou'd also need to manually tweak to deal with rounding issues.
06:31:18kkurbjundimensions are alot less work than all the different screen sizes
06:31:23LloreanSo it's better to use a conversion script to create the new SBS then tweak the output after looking at it.
06:31:25kkurbjunor ratios I mean
06:31:28LloreanSo use scripts
06:31:37LloreanIt's the exact same math as relative positioning
06:31:44LloreanAnd you get nicer looking themes.
06:31:57LloreanBecause you can fix them, rather than permanently being at the mercy of rounding.
06:32:14kkurbjunsomeone has to write and maintain the scripts you are decoupling the plugin author from the responsibility of making their screens work for more than one target
06:32:51Lloreankkurbjun: And?
06:32:58JdGordonarg... I hate pointers.... how do i do a static array or 2 strings?
06:33:34Lloreankkurbjun: I thought the idea was to remove the responsibility from the plugin author so as not to make it "too hard" to convert these non-essential parts of Rockbox.
06:33:59kkurbjunit is alot simpler if it "just works"
06:34:09LloreanRelative positioning will never "just work"
06:34:20kkurbjunlook at brickmania now
06:34:28LloreanAny time you have a new ratio, the people porting to the target will need to readjust *every* plugin to deal with the new dimensions.
06:34:41LloreanIt's the same problem as the scripts.
06:35:07kkurbjunit requires custome bitmaps, but you don't have to do anything beside that - before there were a ton of defines for each screen saying how fast things should b emoving, how fast the screen should update, etc
06:36:18LloreanSo basically you'd rather a way that is technically more interesting, but creates harder to fix errors, than the one that requires a bit more work to maintain, and allows the authors to easily create the exact screen they want?
06:36:47kkurbjunthere are no errors that I am currently aware of - it just works
06:36:55LloreanFor relative positioning, not brickmania
06:37:18LloreanRelative positioning of elements - if the screen gets only slightly larger, elements that are evenly spaced may not be any more due to rounding differences
06:37:21kkurbjunhell, doom can scale it's screen to any size target, you don't have to rescale all the sprites it uses and make the maps work for each target, etc
06:37:28LloreanThe author may prefer to keep them evenly space and just add padding on the outside edges
06:37:54LloreanYou're refusing to address the actual problems with your method, instead citing examples where it does work.
06:38:01LloreanI've admitted in some cases it does work
06:38:04LloreanBut there are cases where it doesn't, too.
06:38:12LloreanIf you don't care about them, that's fine. But out and say it. Don't keep ignoring htem.
06:39:54kkurbjunand I think in most (if not all) cases it can work, I agree there might be an exception, but if you put some thought into it it ends up being alot less maintenance and work in the long run. You might have to make a minor sacrifice, but it's better than sacraficing your time and energy on tasks that do nothing to further real development
06:40:11LloreanThere's no "if not all" about it.
06:40:18LloreanThere are cases where it will, and must, fail.
06:40:39LloreanScaling everything is not always going to give the best possible result.
06:40:50kkurbjunyes, there are no absolutes, but in the situation we are discussing I would be hard pressed to find one in the current sampling we have
06:41:03LloreanAh yes, the "we don't need to care about the future" argument.
06:41:04LloreanFair enough
06:41:20LloreanIn the existing case, relative scaling is probably fine, especially now that most or all menus have been stripped of any artistry.
06:41:35kkurbjunLlorean: if that future comes then so be it, it can be dealt with then, but making bad decisions now doesn't help anything
06:41:46Lloreankkurbjun: But it's not a bad decision. It's simply a decision you don't like.
06:41:53LloreanBecause "what if the scripts break"
06:41:59LloreanIf the scripts *don't* break however, it works fine.
06:42:14kkurbjunwhat happened to checkwps not too long ago
06:42:16LloreanAnd while it's possible to avoid the scripts breaking, relative positioning is fundamentally broken from the get-go
06:42:24Lloreancheckwps is far more complex.
06:42:33 Nick fxb is now known as fxb__ (
06:43:20LloreanAs well, how would your relative positioning deal with different font sizes?
06:43:30LloreanA screen with 1.2 times the height is unlikely to use a font that's exactly 1.2 times the height
06:44:23LloreanYou're basically thinking entirely in graphical senses from the samples you've given (brickmania and doom) were everything displayed can be scaled uniformly. But we don't scale fonts in lists, and are quite unlikely to any time soon.
06:45:06kkurbjunso plugin authors should be able to make skins that could very likely not work with the user selected font
06:45:24kkurbjunwhich would break just as much as a relative positioning
06:45:28kkurbjunif not worse
06:46:40 Join Tuplanolla [0] (n=jani@unaffiliated/tuplanolla)
06:48:00LloreanTrue, without some way to deal with fonts it's mostly useless for both cases.
06:48:26LloreanBut even for scaling the core theme, relative positioning is going to have more troubles with fonts than allowing the author of the theme to simply tweak it after an automated pass.
06:53:24 Join sutefani [0] (
06:54:23 Quit sutefani (Client Quit)
06:54:26 Join sutefani [0] (
06:55:16kkurbjunLlorean: I'm not arguing to take away absolute positioning, but I think that relative positioning would be a good option for alot of problems
06:56:17kkurbjunand I could see plugin specific themes being a good place for them if the decision was made to allow them to be themed
06:56:47kkurbjunor if you took the theming out of the port and make it something like a universal theme that could cover plugin screens too then I wouldn't mind that
06:57:10kkurbjunI just would not want to have to work with each plugin's theme when doing a port
06:57:38kkurbjunIn that case I would rather leave it to the themer than the plugin author
06:57:58kkurbjunand the plugin author could still create a theme/skin if they wanted to external to the main builds
06:59:30 Quit sutefani ("CGI:IRC (Ping timeout)")
07:01:12 Quit liar (Read error: 113 (No route to host))
07:05:56 Join adiroiban [0] (
07:13:09***Saving seen data "./dancer.seen"
07:14:32*JdGordon has everything working corectly :)
07:14:56JdGordonalmost :p
07:21:52JdGordonyay! actually working
07:24:09 Join sutefani [0] (n=spancrat@
07:25:07*sutefani waves. i'm reading the ircguidelines before sounding foolish. hope everyone's having a good night :-D
07:30:29kkurbjunsutefani: no worries, there's not too many people active right now though
07:31:15kkurbjun8 hours earlier is usually more active
07:31:18sutefaniright-o; kkurbjun. i just got my first rockboxable sansa from the wootoff, and i'm trying to figure it out. thanks for your reply!
07:31:38sutefaniroger; perhaps if i have specific questions after this thorough read, i'll come back tomorrow
07:32:23sutefani(i haven't been in an irc chat in about ... 13 years. i'm feeling a bit nostalgic) ha!
07:34:08 Quit bluebrother (Nick collision from services.)
07:34:11 Join bluebroth3r [0] (n=dom@rockbox/developer/bluebrother)
07:35:17pixelmausing relative positioning etc. for plugin( menu)s and having this work right on all screen size and aspect ratio is a dream and will probably look degraded on half of the screen sizes or lead to exceptions all the way. I rather have absolute positioning and more control about it - you have to do it once to adapt to new sizes, you need to do the work with the bitmaps anyways - and yes I did a lot of this for the c200 port and ported a few other
07:35:17pixelma plugins to different screen sizes
07:36:44pixelmain general. Surely some things could and should be improved but very carefully
07:37:41 Quit saratoga_ ("Page closed")
07:39:25CIA-6New commit by jdgordon (r23606): remove 3 bad viewport functions: ...
07:39:38pixelmaand I think the troubles with PLA proves my point that uniformity isn't always good
07:42:45JdGordon" Batt:%bl %bc%ac%?lh<*|> %mh%ar%?ca<%?St|time format|<%cH|%cI>:%cM|−−:−−> " is the inbuilt sbs I'm planning on having (i.e the fallback when I screen forces it on but none are loaded)... any objections?
07:43:31 Join teru [0] (
07:48:57CIA-6New commit by jdgordon (r23607): fix build, ? buttonbar....
07:56:43pixelmawouldn't −−:−− show up on non-RTC targets?
07:56:58JdGordonno, its #if-ed out for rtc
07:57:04JdGordonnon rtc
07:57:22pixelmaand what about volume?
07:58:14JdGordonthis bar will almost never be displayed... its the absolute most crucial info only
07:59:45pixelmawell, I found Llorean's argument earlier that you should know if pressing the "resume" button will deafen you very logical so I would say volume info is crucial too
07:59:56JdGordonthe only real screens which it will be visible in right now is the rec and fm screens... and both shuold be fixed to use the real sbs
08:01:51teruJdGordon: change for draw_slider() in onplay.c in r23606 doesn't looks correct for me. i think vp.x and vp.y should be replaced by 0?
08:02:20CIA-6New commit by blue_dude (r23608): pcmbuf: bug fix with pcmbuf flush, code cleanup, added comments
08:02:38LloreanI think volume is far more crucial than RTC.
08:03:23JdGordonteru: no?
08:03:45JdGordonok, I'll stick the volume in... just starts being low on space
08:03:59 Join kugel [0] (i=kugel@rockbox/developer/kugel)
08:04:05LloreanSo drop the clock.
08:04:20JdGordonbut I think you guys are not understanding just how impossibly hard it will be to see this...
08:04:29LloreanNothing bad's going to happen if your MP3 player isn't showing the time, several of the players simply can't, and most of the ones that can't don't in a lot of themes
08:07:40 Join Rob2223 [0] (
08:09:37 Join Bagder [0] (n=dast@
08:11:16 Join KBH [0] (
08:11:19 Quit Dgby714 (Nick collision from services.)
08:11:20 Join DG-2 [0] (
08:11:28 Nick DG-2 is now known as Dgby714 (
08:11:30saratogaBagder: can you put these test files on the download server? and
08:11:41 Join Unhelpful_ [0] (
08:13:37 Join kadoban_ [0] (
08:14:05 Quit teru (
08:14:05 Quit Tuplanolla (
08:14:05 Quit droidcore (
08:14:05 Quit HBK (
08:14:05 Quit lol3 (
08:14:05 Quit Unhelpful (
08:14:05 Quit kadoban (
08:14:13 Nick KBH is now known as HBK (
08:14:59NJoindroidcore [0] (
08:15:02NJoinTuplanolla [0] (n=jani@unaffiliated/tuplanolla)
08:15:10 Quit Rob2222 (Read error: 60 (Operation timed out))
08:18:22 Join Grahack [0] (
08:18:42 Quit sutefani ()
08:20:09kugelJdGordon: why do you introduce bugs without having a proper replacement?
08:20:32JdGordonwhat bugs?
08:20:47kugelI told you that set_*() need the current viewport, the buttonbar change is rather buggy too
08:21:14JdGordonget the current viewport from the damn set_defaults()
08:21:36kugeland I told you that that doesn't work always
08:21:51kugeleverywhere where you pass a non-NULL vp to do_menu()
08:22:29JdGordonin that case the function i just removed wont help...
08:23:09 Join ender` [0] (
08:23:12 Join bertrik [0] (
08:23:48JdGordonI would have had to remove code from menu.c or list.c or anywhere... but i didnt... set_current was never set except for plugins
08:24:20kugelso you broke several plugins
08:24:49JdGordonthey were broken befoe
08:25:37*kugel will just file a bug report later
08:26:16JdGordonif any plugins broke because of tha chaneg, the build would have broken
08:27:22kugeltry setting the zoom level in pictureflow with a theme that uses ui vp/%Vi
08:28:18kugelor is it even worse, all plugins use the ui vp now even if they are on a black background?
08:30:02JdGordon??? the loader disables the sbs/bar which means its always going to use the fullscreen
08:31:08kugelit doesn't disable the ui vp
08:32:35 Join AndyIL [0] (n=pasha_in@
08:33:41 Quit kugel ("exit(0);")
08:36:23 Join Tomis [0] (n=Tomis@
08:37:54*JdGordon cant see the problem
08:39:03 Quit goffa (Read error: 104 (Connection reset by peer))
08:41:20 Join LinusN [0] (n=linus@rockbox/developer/LinusN)
08:43:28 Join goffa [0] (n=goffa@
08:43:56 Quit AndyI (Read error: 110 (Connection timed out))
08:46:35 Join kugel [0] (i=kugel@rockbox/developer/kugel)
08:47:08kugelJdGordon: I don't think plugins should use the ui vp if they don't draw lists or draw lists without backdrop
08:47:58JdGordonthen they will be using the fullscreen viewport and doing everything themselves
08:48:35kugelyou fixed no plugin to do that
08:48:58JdGordonfunny... all the ones i tried work exaclty like that
08:49:20kugelwith a ui vp enabled theme?
08:50:37 Quit phanboy4 ("Leaving")
08:53:16 Join flydutch [0] (
08:59:33 Join liar [0] (
09:05:21 Quit bertrik ("De groeten")
09:06:07 Join lol3 [0] (
09:06:13liarTheSeven: pong
09:06:20liarTheSeven: what have you found out?
09:07:52 Join dmb [0] (n=Dmb@unaffiliated/dmb)
09:09:32 Quit dmb (Read error: 104 (Connection reset by peer))
09:09:58 Join dmb [0] (n=Dmb@unaffiliated/dmb)
09:10:24kugelkkurbjun: ping
09:12:17kugelJdGordon: r23607 seems buggy too
09:13:12***Saving seen data "./dancer.seen"
09:14:06 Join gregorovius [0] (n=diego@
09:17:20gregoroviusHi, one small thing that's bothering me: since upgrading my ipod 5.5g to 3.4, the song title/year and artist in the AAfreak theme get hidden behind the album art
09:18:01gregoroviusanybody seen this? I don't really know how wps works
09:21:23gregoroviusfound something, will try
09:21:26 Quit gregorovius ("Leaving")
09:24:15 Join liar_ [0] (
09:25:08 Join Thundercloud [0] (
09:25:18 Quit liar (Operation timed out)
09:39:49 Quit kugel (Read error: 60 (Operation timed out))
09:48:55 Join kugel [0] (i=kugel@rockbox/developer/kugel)
09:52:44 Nick liar_ is now known as liar (
09:52:58 Join petur [0] (n=peter@rockbox/developer/petur)
09:59:30 Quit Thundercloud (Remote closed the connection)
10:02:34 Join dionoea_ [0] (n=dionoea@
10:05:22 Quit dionoea (Read error: 104 (Connection reset by peer))
10:11:55 Quit reid05 (Read error: 110 (Connection timed out))
10:12:03 Join reid05 [0] (
10:23:20kugelBlue_Dude: ping (for the logs)
10:29:56 Quit Llorean (Read error: 104 (Connection reset by peer))
10:34:37 Join uflops_ [0] (
10:34:38 Quit uflops (Read error: 54 (Connection reset by peer))
10:34:43 Join einhirn [0] (
10:37:52 Join Llorean [0] (
10:38:19 Join kugel_ [0] (i=kugel@
10:44:42 Quit kugel (Read error: 60 (Operation timed out))
10:54:34 Join robin0800 [0] (
10:57:57 Quit robin0800 (Remote closed the connection)
11:02:13 Join robin0800 [0] (
11:04:44 Quit robin0800 (Remote closed the connection)
11:11:00 Quit Grahack ("Leaving.")
11:12:04 Join robin0800 [0] (
11:13:15***Saving seen data "./dancer.seen"
11:18:26 Join Highlander [0] (
11:25:56 Quit freqmod_qu (Read error: 104 (Connection reset by peer))
11:26:17 Join freqmod_qu [0] (n=quassel@
11:29:11 Quit Highlander ("Quitte")
11:30:44 Quit BHSPitMonkey (Read error: 113 (No route to host))
11:35:51 Join DerPapst [0] (n=DerPapst@
11:35:54 Quit kugel_ (Read error: 110 (Connection timed out))
11:37:08 Quit DerPapst (Read error: 131 (Connection reset by peer))
11:37:26 Join DerPapst [0] (
11:42:20 Quit bubsy ()
11:47:01 Join bubsy [0] (n=bubsy@
11:49:33 Join Highlander [0] (
11:54:11 Quit shaggy-h (Read error: 104 (Connection reset by peer))
11:54:34 Quit killan (Read error: 104 (Connection reset by peer))
11:54:35 Join killan_ [0] (
11:59:20 Nick Unhelpful_ is now known as Unhelpful (n=quassel@rockbox/developer/Unhelpful)
12:07:30 Join robin0800_ [0] (
12:08:19 Quit robin0800 (Remote closed the connection)
12:10:07 Join kugel_ [0] (i=kugel@
12:10:56 Nick kugel_ is now known as kugel (i=kugel@
12:14:37 Join robin0800 [0] (
12:14:39 Quit AndyIL (Read error: 104 (Connection reset by peer))
12:16:19 Join Lynx_ [0] (
12:17:35 Quit JdGordon| (Read error: 104 (Connection reset by peer))
12:19:20TorneUnhelpful: indeed, you'd also probably need to tune the actual levels it went up and down at a bit, but hey
12:22:45gevaertsTheSeven: yes, the USB driver boosts for MSC. Something that we probably need to finetune a bit, as probably not all targets need it
12:22:46UnhelpfulTorne: yes, i saw about the complicated and long list of adjustable scaling triggers
12:24:11TorneUnhelpful: the tradeoff is hilariously complicated
12:24:29Torneyou start with the assumption that you want to run as *fast* as possible so you get to being idle quicker
12:24:41Tornethen you go "wait shit but power changes with a square relationship to voltage"
12:25:00Tornethen you tear all your hair out and guess
12:30:37 Join kugel_ [0] (i=kugel@
12:30:48 Quit kugel_ (Read error: 104 (Connection reset by peer))
12:38:25 Part LinusN
12:39:15UnhelpfulTorne: in the case where the relationship is linear, and idle cycles still have high cost at high speed, yes, run-fas-then-idle is the way to go
12:39:47Unhelpfulwhich i believe is about where things are with some other targets that have clock-only scaling and less efficient idle
12:40:06 Quit kugel (Read error: 104 (Connection reset by peer))
12:40:14Unhelpfuli think for beast the best we can manage is "try to run just about as fast as we need to to keep up"
12:40:59TorneUnhelpful: Well, it depends
12:41:17TorneARM11 should idle pretty effiiently, as I believe we've already observed
12:41:53Tornebut the nonlinear relationship between speed and consumption is the trick
12:48:23 Join kugel [0] (i=kugel@rockbox/developer/kugel)
12:52:45Unhelpfulif the voltage/clock relationship is roughly linear with auto voltage/clock scaling, and power has a square relationship to voltage, i don't think it *ever* makes sense to go faster than you "need to". if you finish the work in half the time, at four times the cost, it will hurt you even if idling the other half of the time is completely free.
12:53:21 Quit kugel (Read error: 131 (Connection reset by peer))
12:53:52 Join Zagor [242] (n=bjorn@rockbox/developer/Zagor)
12:58:31 Join kugel [0] (i=kugel@rockbox/developer/kugel)
12:59:23 Join dfkt [0] (i=dfkt@unaffiliated/dfkt)
13:03:34TorneUnhelpful: The problem is working out how fast you actually need to, though
13:03:48TorneUnhelpful: to actually get the "perfect" solution you need an oracle function
13:04:08Tornewhich you can only implement perfectly in a statically scheduled hard-RT system :)
13:04:34Torneit's one thing having the hardware to measure the pipeline busy/idle state and ramp up and down based on that
13:04:50Tornebut what that hardware doesn't know is what's coming next..
13:05:19Torneif we know "we need to do exactly X amount of work every Y milliseconds to continue to refill pcmbuf" then you could calculate the freq you need to be at to make that work just barely in time (static scheduling)
13:06:04Tornethe pipeline analyser doesn't know that although you are at 100% load right now, you are going to be idle in 10,000 instructions time at this rate, and thus you can afford to slow down.
13:07:39Tornethe automatic scaling in the hardware is more or less predicated on the assumption that a good approximation is to run at the slowest speed which doesn't cause the core to be fully utilised.
13:08:15 Join evilnick_ [0] (
13:08:22Tornethis is pretty reasonable for an interactive system, but not a realtime one
13:08:29Tornewhich is *kinda* what we are, when just doing playback ;)
13:09:08 Join Strife89 [0] (
13:09:12 Quit Strife89 (Read error: 104 (Connection reset by peer))
13:09:13TorneSo, yah.
13:09:19TorneIt's definately worth experimenting with
13:09:49TorneBut it's far from certain that letting it scale freq automatically will come out "better" than if we just implemented boosting using the same hardware and *also* turned on the appropriate voltage scaling to go with a boost/unboost freq
13:10:16Tornethe voltage scaling is basically an automatic win whatever we do though, so that should go on first :)
13:13:17***Saving seen data "./dancer.seen"
13:25:33 Quit seani ("Leaving")
13:25:54 Quit evilnick (Read error: 110 (Connection timed out))
13:25:58 Quit petur ("reboot")
13:26:19 Nick evilnick_ is now known as evilnick (
13:26:41Tornealso i doubt that freq/voltage is linear
13:31:17 Quit gevaerts (Nick collision from services.)
13:31:22 Join seani [0] (n=seani@
13:31:26 Join gevaerts [0] (n=fg@rockbox/developer/gevaerts)
13:33:37 Join draft [0] (
13:35:21Unhelpfulhrm. yes, that will be a problem. :/
13:35:32draftgreat that the second generation iPod Nano is now cracked as well. has anyone here installed it?
13:36:08drafti might have some questions about Manually installing it from
13:40:39 Join uflops__ [0] (n=yogurt@
13:44:00 Join Grahack [0] (
13:52:02 Quit seani (Remote closed the connection)
13:52:16 Join seani [0] (n=seani@
13:53:47 Quit kugel (Read error: 110 (Connection timed out))
13:56:24TheSevengevaerts: It does make sense for nano2g though
13:56:42gevaertsTheSeven: that's what I assumed after what you said :)
13:57:28gevaertsTheSeven: Are you sure the work-threshold is 66MHz though? I'd expect 60
13:57:50 Join uflops__1 [0] (
13:58:28 Quit Highlander (Read error: 104 (Connection reset by peer))
13:58:44TheSevenwell, 47923200 fails while 63897600 works
13:58:56gevaertsTheSeven: one thing that may be wrong though is that we only boost when going to MSC mode, not when connecting. That's fine on PP, because the issue there is that packets larger than 96 bytes don't work without boosting, but if even small packets are unreliable that's not good enough
13:59:02TheSevenso 60 may be true, even though I don't see any reason why it should fail at lower freqs
13:59:02 Quit uflops__ (Read error: 131 (Connection reset by peer))
13:59:10 Quit uflops_ (Read error: 110 (Connection timed out))
13:59:23TheSevenit has huge fifos and an auxillary clock source for the whole utmi+ domain
14:00:12TheSevengevaerts: USB seems to instantly break even for smallish (8/16 byte) packets on nn2g at 47923200
14:02:16gevaertsTheSeven: maybe a stupid question, but does it also break if you keep the voltage within spec? :)
14:02:29 Join kugel [0] (i=kugel@rockbox/developer/kugel)
14:03:05UnhelpfulTorne: so, what we probably need is to set frequency manually when "only" decoding... unless there's some way we can manually feed dvfs our own measure of how "busy" we are?
14:04:07CtcpPing from gevaerts!n=fg@rockbox/developer/gevaerts
14:07:16 Part Bagder
14:07:31 Join kugel_ [0] (i=kugel@
14:08:17Unhelpfulbut then if some interactive load needs more, we need to be able to boost. hrm. perhaps a cpu_boost for beast that uncaps DVFS, and a manually set frequency the rest of the time?
14:10:34 Quit kugel_ (Client Quit)
14:10:50 Join Horscht [0] (n=Horscht2@xbmc/user/horscht)
14:11:34 Quit z35 ("Leaving")
14:15:55 Join Biont [0] (
14:19:21 Join Bagder [0] (n=dast@
14:21:46BiontHi!the WPS I made is showing an ugly redraw bug after some changes I made, but I just can't find the mistake. I spent 2 hours looking over the code now and nothing seems to solve it. Is it possible that it just happens?
14:22:40BiontI get the same glitch every time anmd it wasn't there before, so there must be a bug *somewhere* but I'm pretty desperate right now
14:25:50 Join angelwolf71885 [0] (
14:30:43 Nick uflops__1 is now known as mikroflops_ (
14:32:25 Quit Grahack ("Leaving.")
14:36:28seaniOT, but how do I suppress join / leave / quite messages in IRC? Are they client specific, server specific, or some mysterious combo?
14:36:29 Quit DerPapst ("Leaving.")
14:40:21robin0800seani: which client?
14:42:51seaniThe client is "muxi", although I can see someone confirming the error of my ways :-)
14:43:05seaniOops "Smuxi" that is
14:44:10seaniYou can create a set of "On Connect" commands, I thought it might be one of those but...
14:46:37TorneUnhelpful: haven't much of a clue, to be honest. I suspect we need to implement the code to control the hardware and then just fiddle about with it :)
14:49:44 Join drafti [0] (n=draft@
14:50:05draftiShouldn't Rockbox support 2nd Generation Nano?
14:50:17draftii just tried to run ipodpatcher and it says like this:
14:52:18TheSevendrafti: this is an ipodpatcher bug that was fixed in svn a while ago
14:54:24 Join fdinel [0] (
14:55:59 Quit kugel ("exit(0);")
14:56:00 Join petur [0] (
15:04:23 Quit fdinel ("Miranda IM! Smaller, Faster, Easier.")
15:04:50 Join uflops_ [0] (
15:06:46 Join kugel [0] (i=kugel@rockbox/developer/kugel)
15:07:24 Quit kugel (Client Quit)
15:08:26 Quit angelwolf71885 (Remote closed the connection)
15:08:40 Join uflops__ [0] (
15:09:10 Join kugel [0] (i=kugel@rockbox/developer/kugel)
15:09:31 Quit kugel (Read error: 104 (Connection reset by peer))
15:13:19***Saving seen data "./dancer.seen"
15:13:38 Join kugel [0] (i=kugel@rockbox/developer/kugel)
15:14:18 Join fdinel [0] (
15:16:46 Quit mikroflops_ (Read error: 60 (Operation timed out))
15:18:50 Quit reid05 (Success)
15:19:21 Quit uflops__ (""Maailmanpyoramuualta"")
15:19:29 Join reid05 [0] (
15:19:58 Quit drafti (Remote closed the connection)
15:22:00 Quit liar (Read error: 104 (Connection reset by peer))
15:25:40BiontCan someone help me with my WPS code? I just can't get rid of a redraw glitch that's normally caused by overlapping viewports. But I can't find any mistake in it. Is it possible that it just happens when a WPS gets too complex?
15:25:51 Quit uflops_ (Read error: 110 (Connection timed out))
15:26:41BiontIt's nothing severe, it's just ugly, because it cuts of 15-20px from the album art whenever playback is paused
15:27:51BiontMaybe I just got blind for mistakes after spendig so much time with the WPS, if someone wants to have a look at it, I can post it on pastebin
15:28:01 Join Blue_Dude [0] (n=chatzill@rockbox/developer/Blue-Dude)
15:28:32Blue_Dudekugel: pong!
15:31:13 Join n1s [0] (n=n1s@rockbox/developer/n1s)
15:34:19gevaertsBiont: is your wps available somewhere? I'm not an expert, but I imagine that maybe more people can help if they can actually see it :)
15:34:28 Quit Hadaka (Read error: 60 (Operation timed out))
15:37:21Biontgevaerts: The thing is, its a touch enabled theme for Cowon D2 and the simulator is outdated, so I'd had to upload some 30mb somewhere just to show it (including the new simulator which is pretty bloaty)
15:37:49 Quit kugel (Read error: 110 (Connection timed out))
15:38:33 Join Naked [0] (
15:38:34 Nick Naked is now known as Hadaka (
15:39:01 Quit AEnima1577 ("Leaving.")
15:41:28BiontUnless you have a D2, of course
15:42:31gevaertsBiont: lots of people can build their own simulator :)
15:44:52 Join panni_ [0] (
15:47:29 Join liar [0] (
15:47:34 Join Lss [0] (
15:48:31Biontgevaerts: Okay then, give a minute
15:49:13Blue_DudeDid the last crossfade update break the %xf tag? If so, where should I go to fix it?
15:49:37 Join evilnick_B [0] (i=0c140464@rockbox/staff/evilnick)
15:50:14Blue_DudeAlso, the changelog in the manual seems very out of date...
15:50:27 Join DerPapst [0] (
15:57:20*gevaerts tries to find how to pause the sim
15:57:37Biontgevaerts: long click on album art/middle of the screen
15:58:54gevaertsBiont: in absolute mode?
15:59:03*gevaerts isn't really used to touchscreens in rockbox yet
15:59:31BiontYes. Sorry, I forgot to mention that
16:02:24gevaertsBiont: so the problem is, when you long click and it shows this "Options" box, the AA gets clipped?
16:03:39 Join AEnima1577 [0] (
16:03:59Biontgevearts: exactly. The problem showed up while fiddling around with the conditional viewports for the volume change display and I have no idea how to get rid of it
16:05:19 Quit Lss (Read error: 104 (Connection reset by peer))
16:05:48gevaertsBiont: there's more. If the screen turns off and I make it wake up again, the AA is gone
16:06:44 Join Lss [0] (
16:07:43Biontgevaerts: I know. But I think that's just the way the skin engine works. It's not designed for that kind of stuff, after all
16:08:23 Quit polobricolo ()
16:11:52 Join Strife89 [0] (n=michael@
16:16:14 Quit antil33t (Read error: 131 (Connection reset by peer))
16:16:22 Join antil33t [0] (
16:16:22Biontgevaerts: maybe that can be solved anyway, but I noticed the parser is very very picky about that
16:16:28gevaertsBiont: I have to admit that this WPS is too complex for me right now :)
16:17:20Biontgevearts: Okay no problem, maybe I'll find out myself. (I hope you like the wps anyway :) )
16:17:48Biontgevaerts: Thanks you for your time :)
16:17:51gevaertsBiont: only it looks to me like you ask the AA not to be shown while paused
16:18:37gevaertsViewport c shows the AA, viewport d doesn't.
16:18:45evilnick_BBlue_Dude: The new crossfading works great - thanks for adding that extra option as that's precisely what I've always wanted
16:19:07gevaertsSo I think it looks like a bug that there are still bits of AA being shown when paused, not that there isn't enough of the AA :)
16:19:31evilnick_BIf you're manually skipping tracks then it makes sense (to me) that you'd have the instant feedback of a non-crossfaded trackchange, for pocket operation etc.
16:19:41Biontgevaerts: You're right. I'm working on that right now, but I fear that I did it that way for a reason "back then"
16:20:04Blue_Dudeevilnick_b: Welcome! It might need a bit of work, but it's 95% of the way there. I'm not really certain what the shuffle and man track skip option is for, but it was already there, so...
16:20:37evilnick_BThe options might need some work on the naming system
16:21:01evilnick_BEr... I mean: The options could be named better/more obviously
16:21:48Blue_DudeI tried to do some of that with the latest update. I thought there would be more screaming from the .lang purist crowd though.
16:22:38evilnick_BShuffle now means that you need to have shuffle enabled, but that crossfades "auto"-changes AND manual changes, which to me seemed unexpected (from the name)
16:22:47Blue_DudeBut since I don't know what good the shuffle and man skip option is, I had a hard time coming up with a new name for it.
16:23:03evilnick_BAnd that's precisely where I'm now stuck :)
16:23:58Blue_DudeMaybe we ought to ditch the second shuffle option and just fix the plain vanilla shuffle one.
16:24:38Blue_DudeWhat would "expected" behavior be? Since I redid the decision tree, fixing crossfade behavior is trivial now.
16:27:43 Quit Strife89 ("Gone to class.")
16:27:53 Quit liar (Read error: 104 (Connection reset by peer))
16:31:38 Join toffe82 [0] (n=chatzill@
16:33:55Biont*sigh*....I don't get this theme myself any more. I guess I'll just live with that small issue
16:34:39Biontasking the AA not to be shown when playback's paused seems to be the only way to do it
16:35:16gevaertsBiont: can't you show the images option images on the same VP as the AA?
16:37:50Biontgevaerts: Maybe, but I need a viewport for the different touch areas so that they don't interfere with each other, so I can as well show the images in their respective viewport
16:38:04Biontgevaerts: Or did I miss something?
16:38:40gevaertsI don't know. I do simple wpses :)
16:39:58gevaertsit's a nice looking theme though. I spent a while trying to find your sbs until I saw that it was where I'd been looking all along, so it looks very well integrated :)
16:41:04 Join StealthyXIIGer [0] (
16:41:07Biontgevaerts: thank you :)
16:42:03Biontgevaerts: But making touchscreen WPS that actually have some functionality is still quite a pain atm. I think it doesn't get bigger thatn this anytime soon
16:43:00CtcpPing from gevaerts!n=fg@rockbox/developer/gevaerts
16:43:33Biontgevaerts: That's why I called for custom tags/actions in the forums :/ That would make many thing much more awesome^^
16:46:26Blue_Dudeevilnick_B: for shuffle, maybe it should be much the same as Auto Track Skip, and only crossfade at the end of the track with shuffle enabled. Man track skips would still occur immediately.
16:47:04 Join Grahack [0] (
16:48:33evilnick_BBlue_Dude: That'd get my vote, but it is changing an option that's been the other way round for quite a long time.
16:49:09evilnick_BProvided that it's a justifiable change then it probably should be changed, but perhaps worth sending a dev-ml email round?
16:49:44Blue_DudeYeah, I could see that. Right now it's really an Always if Shuffle option.
16:49:53 Quit AEnima1577 ("Leaving.")
16:50:23Blue_DudeI'll send one off...
16:50:42gevaertsBiont: I guess you have to talk to the experts for that
16:52:54Biontgevaerts: I'll just keep on hoping that some of them comment on the thread :)
16:56:07 Join AEnima1577 [0] (
17:02:04 Quit Zagor ("Don't panic")
17:10:00 Join liar [0] (n=liar@
17:11:42Blue_DudeAny WPS tag gurus here?
17:11:53 Quit DerPapst ("Leaving.")
17:13:23***Saving seen data "./dancer.seen"
17:13:32n1swill gcc let you pass litteral numbers (or ints with the value of the enum) to a function that takes an enum as a type?
17:13:45*n1s guesses no
17:14:45TheSevenmaybe if you cast it, but this isn't clean at all... why do you need to do this?
17:14:50Blue_DudeFor low memory targets, crossfade doesn't exist. So for those targets, should the crossfade WPS tag be removed entirely or just return 0?
17:16:44 Quit adiroiban ("Leaving.")
17:17:01 Quit seani (Remote closed the connection)
17:17:07saratogaI would have it still exist but return 0 so that WPSs can be used on low and high mem targets with the same screen size
17:17:30 Join seani [0] (n=seani@
17:17:43Blue_DudeSo keep the token but make it a dummy function?
17:17:44saratogaotherwise the theme site will have to sort targets by memory as well as screen which is bound to be confusing
17:18:24saratogai don't know how wps works, but IMO it should be implemented in a way that makes the token still work
17:19:54 Part draft
17:20:10Blue_DudeIt's supposed to return the crossfade enable state, but that setting doesn't exist for some targets. I made it return NULL just to get the thing to compile, but I'm thinking that just returning 0 (disabled) all the time might be a better approach.
17:20:38saratogamaybe having the crossfade state variable still exist and always be zero would be cleaner?
17:20:55Blue_DudeOr get rid of the token entirely if it doesn't apply. That will cause syntax errors in the WPS script though.
17:20:57saratogaunless theres a lot of overhead to having it around
17:20:59n1sTheSeven: some functions in sound.c take ints as args but are really passed the values of an enum, but i assume these values are saved in the settings so if i start converting these to take enum args i will have problems with the settings reading
17:21:06 Part toffe82
17:21:52saratogaif its just a few bytes having the variable still exist makes sense to me, asking if crossfade is enabled is perfectly valid even if you don't have it, the answer logically should be "no"
17:22:00Blue_Dudesaratoga: Nah, I killed off all related variables. It would only be needed in this one spot, and returning 0 is easy.
17:22:02TheSeventhen you should cast them at a point as near as possible to the settings code, i think
17:22:24 Quit seani ("Leaving")
17:23:06n1sTheSeven: i guess so, not sure i will mess with it today though
17:23:11 Join polobricolo [0] (
17:24:19Blue_Dudesaratoga: I'll keep the token. I didn't think about sharing themes.
17:24:54 Nick dionoea_ is now known as dionoea (n=dionoea@
17:26:06 Join jgarvey [0] (
17:28:14 Join Strife89 [0] (n=michael@
17:31:22 Quit panni_ ("( :: NoNameScript 3.81 :: )")
17:33:41 Quit Dgby714 (Nick collision from services.)
17:33:41 Join DG-2 [0] (
17:33:50 Nick DG-2 is now known as Dgby714 (
17:36:47Blue_DudeHm. Where can you find a list of the \opt{} settings for the manual?
17:37:32 Join Strife1989 [0] (n=nds@
17:38:36 Join seani [0] (n=seani@
17:41:42 Join toffe82 [0] (n=chatzill@
17:43:28n1sBlue_Dude: most are autogenerated from /apps/features.txt and the rest are in the platform files
17:43:52n1smaybe not most
17:43:55Blue_DudeThanks very much, I'll look there.
17:44:22Blue_DudeI'm trying to use the HAVE_CROSSFADE tag in the manual, but it ignores it...
17:45:30Blue_DudeOh I see. I'll add the tag in features.txt and it should behave...
17:47:01 Quit krazykit (Read error: 60 (Operation timed out))
17:47:17 Join Omlet [0] (i=omlet05@
17:47:39 Quit StealthyXIIGer (Read error: 54 (Connection reset by peer))
17:48:02 Join StealthyXIIGer [0] (
17:48:54Blue_DudeIf I add:#if defined(HAVE_CROSSFADE) \ crossfade \ #endif at the end of features.txt will it screw everything else up?
17:49:38 Join krazykit [0] (n=kkit@
17:52:46 Quit Biont ("ChatZilla 0.9.85 [Firefox 3.5.5/20091102152451]")
17:54:19CIA-6New commit by nls (r23609): Comstify a function pointer array
17:55:37n1sBlue_Dude: that should not screw anything up, but iirc it's alphabetically(ish) sorted so you might not want to add it to the end
17:56:12n1sfunctionally it doesn't matter where in the file you add it
17:56:42Blue_DudeThere's a note in the beginning about breaking backward compatibility. I really, really don't want to go there.
18:00:08 Part Bagder
18:02:38 Quit Strife89 ("My number of files is OVER 9000!")
18:05:33 Nick Strife1989 is now known as Strife89 (n=nds@
18:10:53 Quit Strife89 ("Going, going, gone.")
18:19:11 Join kugel [0] (i=kugel@rockbox/developer/kugel)
18:23:05 Join phanboy4 [0] (
18:25:58Blue_DudeNow that I've added a tag in features.txt, should I use it in the lang files? Would that save memory if I did?
18:28:13n1sah, but changing the order of the features will not break anything, changing the conditions so that a feature that was present on one device is no longer *will* and that breaks voice files
18:28:43 Join gtkspert [0] (
18:29:05 Quit einhirn (Read error: 104 (Connection reset by peer))
18:29:23 Join einhirn [0] (
18:29:36n1sBlue_Dude: it could save memory on devices that don't have te feature if they previously had the string but it will break voicefiles on those devices, btu now that we have pretty regular releases i don't know how much we care about the voicefiles
18:29:40Blue_DudeRight now the crossfade lang tags are keyed to swcodec. You're saying that changing them to crossfade would be a bad idea?
18:29:42n1s... breaking
18:29:53n1sBlue_Dude: no
18:30:27n1sBlue_Dude: i'm saying that it will break compatibility with older vociefiles on the targets which have swcode but not crossfade
18:30:45n1swhich are those targets btw?
18:31:50Blue_DudeThere are only three that have swcodec but not crossfade: Sansa Clip, Sansa c200v2 and Sansa m200v4
18:32:30n1syou should be fine then
18:32:41Blue_DudeAt least, those are the only ones that showed any memory savings when I carved out crossfade code a day or two ago.
18:32:43n1sas those are unstable targets with serious bugs anyway
18:33:08Blue_DudeSo just do it and get it over with? :)
18:35:20Blue_DudeI'm going to commit what I've got and tweak the lang files later...
18:35:25gevaertsBlue_Dude: none of those actually works properly anyway
18:35:54Blue_DudeOK. It's going to be cut and paste for a while, but better to do it now than later.
18:36:14Blue_DudeI'll post lang changes tonight.
18:36:22Blue_DudeManual changes coming right up though.
18:36:37n1sthe reason for this somewhat unfriendly behaviour of the voice system is that it's very simple and lacks proper version management
18:38:01CIA-6New commit by blue_dude (r23610): Add crossfade feature tag, update manual, fix crossfade WPS tag behavior
18:38:12 Quit phanboy4 (Read error: 104 (Connection reset by peer))
18:38:21 Quit kugel (Connection timed out)
18:39:15 Join funman [0] (n=fun@rockbox/developer/funman)
18:39:28 Quit fyre^OS (Read error: 60 (Operation timed out))
18:44:17 Quit gtkspert_ (Read error: 101 (Network is unreachable))
18:47:41 Quit Blue_Dude ("ChatZilla 0.9.85 [Firefox 3.5.5/20091102152451]")
18:47:43 Join DerPapst [0] (
18:47:48 Quit Sajber^ (Read error: 131 (Connection reset by peer))
18:52:01 Join fyrestorm [0] (
18:56:18 Quit funman ("free(random());")
18:58:57 Quit AEnima1577 (Read error: 104 (Connection reset by peer))
18:59:22seaniI may say, despite the fact there may be changes in the works, messing around with themeing / wps is fun. Any plans for multiline / text wrapping? I can always scroll of course but on a C240 lots of ghosting
18:59:29 Quit fyrestorm (Read error: 60 (Operation timed out))
19:00:02 Quit petur ("later")
19:00:15 Join antil33t1 [0] (
19:00:21 Quit antil33t (Nick collision from services.)
19:00:43 Join Tomis2 [0] (n=Tomis@
19:01:03 Quit ender` (" Replication with a nonconfigured partner is not allowed. -- net helpmsg 4006")
19:01:44pixelmaI played around a lot with colour schemes (basically settled for black on white), contrast, brightness and scrolling settings to get less blur on my c200's display. Values most probably vary per actual target (screen)
19:01:59 Join AEnima1577 [0] (
19:02:56 Join fyrestorm [0] (
19:03:04seanipixelma: Yes. I've gone for essentially a few shades of amber on black (for nighttime use) and scrolling that's jerkier and slower rather than smoother and faster
19:03:35seanipixelma: Smooth and fast looks great in the sim. F40, very different on a C240.
19:04:20 Quit fyrestorm (Read error: 104 (Connection reset by peer))
19:04:37 Join fyrestorm [0] (
19:05:32 Join gtkspert_ [0] (
19:05:44 Join ChelseaToner [0] (
19:05:44 Join shaggy-h [0] (
19:07:09 Join funman [0] (n=fun@rockbox/developer/funman)
19:07:13 Quit Tomis (Read error: 110 (Connection timed out))
19:07:13 Nick Tomis2 is now known as Tomis (n=Tomis@
19:08:29ChelseaTonerHey - Having problems putting music on to my 2nd gen ipod nano now that I've put Rockbox on it. None of my music applications (gtkpod, banshee, rythmbox etc) will recognise it. I can put music on it if i change to the apple firmware, but the music still doesn't appear on Rockbox when i switch back. I am aware that the 2nd Gen ipod nano isn't fully supported yet, is this a common problem with generation of devices? Or am i doing someth
19:08:29ChelseaTonering wrong?
19:09:19saratogaeither drag and drop your music onto the nano, get software that can add music to an MSC device, or use the rockbox database so you can see music added by the Apple software
19:10:05robin0800ChelseaToner: also check file view is set to all
19:10:29gevaertsrobin0800: why?
19:11:37robin0800gevaerts: to make sure you can see hidden files etc.
19:12:41ChelseaToneri don't think you can use and ipod as a mass storage device either way.. I'll try anyway
19:12:55gevaertssure you can
19:13:19robin0800ChelseaToner: set it to disc mode in i-tunes
19:13:24***Saving seen data "./dancer.seen"
19:14:38ChelseaTonerI don't have i-tunes. Is there an alternative way?
19:15:38n1sif you don't have itunes it should show up as an msc device
19:15:53 Quit shai ("Leaving")
19:15:58n1situnes hides this fac on systems where its installed
19:16:18n1sfact and it's
19:17:08 Quit gtkspert (Read error: 101 (Network is unreachable))
19:19:46 Join domonoky [0] (n=Domonoky@rockbox/developer/domonoky)
19:20:39polobricolois there a way i can build rockbox using cygwin
19:20:50evilnick_Bpolobricolo: Yep, that's on the Wiki
19:21:33 Join funman_ [0] (n=fun@rockbox/developer/funman)
19:22:36 Nick funman_ is now known as namnuf (n=fun@rockbox/developer/funman)
19:23:40polobricolothe thing is i don't have any problems to use the cross-compiler (i build my own arm code succesufully) the thing is i get a "No rule to make target '/home/Paul/rockboc/tools/root.make'"
19:23:46 Join iHPUser [0] (
19:23:49polobricolobut i checked the file exists
19:25:02polobricoloi get that error when running make
19:25:50 Join bertrik [0] (
19:26:12CtcpPing from gevaerts!n=fg@rockbox/developer/gevaerts
19:31:03iHPUserIs it possible to install a larger hdd on an iRiver iHP-140 using rockbox? The standard is 40g, if I put anything larger in it will it be readable?
19:31:59evilnick_BiHPUser: Yes
19:32:13polobricoloiHPUser: i would say yes by rockbox but maybe not with the original firmware
19:32:46evilnick_BThere's the usual 137GB limit to using the regular builds, but if you can compile your own then it's a simple couple of changes.
19:33:19 Join freddyb [0] (
19:33:25evilnick_BiHPUser: I've used a 60GB drive in an H140 successfully
19:33:53iHPUserAs long as rockbox is as stable as the original firmware, I don't see any reason to go back to the OF.
19:34:05gevaertsiHPUser: may help in choosing a working disk
19:35:13polobricoloanyone has an idea about the problem in cygwin with root.make not found
19:36:01domonokypolobricolo: did configure run without problems ?
19:36:14iHPUserIs there any top limit that the hardware can't handle, I don't want to drop a 500G hdd in and not be able use it all (after doing the custom compile)?
19:36:52domonokyiHPUser: for ihp1x0 staying under 137gb is preferable.. (it gets complicated if you want to use more)
19:37:23gevaertsiHPUser: I don't think there are 50 pin drives bigger than 80GB
19:37:25evilnick_BYou're also limited by the dimensions of the case, unless you want to start using desktop drives
19:38:14 Join Thundercloud [0] (
19:38:14gevaertsthere are adapters for ZIF-40 drives though, but then you run into fun space issues in the case
19:39:11iHPUser80G should be fine at this point −− I've seen 320G Samsung 1.8" drives though
19:39:34gevaerts1.8", yes, but probably not PATA, and definitely not 50 pin
19:39:35iHPUserI didn't see the pins on them though
19:41:08polobricolodomonoky: as on my linux box a warning becausemy arm-elf-gcc is too recent
19:41:14polobricolo*because my
19:41:43 Quit funman (Read error: 110 (Connection timed out))
19:41:51 Nick namnuf is now known as funman (n=fun@rockbox/developer/funman)
19:41:54gevaertspolobricolo: maybe line ending issues?
19:42:21domonokypolobricolo: are you trying now on linux or cygwin ?
19:43:03polobricoloi'm on cygwin (i can't use linux right now)
19:43:52domonokybut did configure in cygwin give any error ?
19:44:46polobricolono just the gcc version warning (but i had this warning on linux too)
19:45:23polobricolois there an ipodpatcher zip i could download but with nano2g support
19:45:42domonokygenerally you should use the recomended gcc/binutils versions.. (can be installed with (at least on linux))
19:46:52*domonoky points to the wiki for ipodpatcher
19:47:12domonokytake a look at the IpodNano2GPort page
19:49:44 Join stoffel [0] (
19:50:13 Join Strife89 [0] (n=michael@
19:52:33Strife89I can't find any information on how to tell a v1 Fuze from a v2 without opening it.
19:52:42iHPUserI was just looking at modding the iHP with a compact flash card, looks good but gives less space than I have currently. Would it be possible just to put a solid state drive in it?
19:53:24 Part polobricolo
19:53:43bertrikStrife89, I think it has a different original firmware version
19:54:55Strife89This one I just got has firmware version 01.01.07A
19:56:37iHPUserNevermind, they're too big
19:56:58 Join JdGordon| [0] (n=Miranda@nat/microsoft/x-mvmsqwoxxasrczry)
19:57:00gevaertsiHPUser: there are SSDs that fit
19:57:21Strife89I'm going to assume that that's a v1.
19:58:40bertrikStrife89, I think that's correct, see for example
20:00:34Strife89bertrik: Would it be better to update the firmware to 01.02.28, or leave it? (I intend to put Rockbox on it when the port is finished, naturally.
20:00:51 Part ChelseaToner ("Leaving")
20:03:25domonokyStrife89: no need to update it, rockbox already works pretty good on the v1 fuze :-)
20:03:32iHPUsergevaerts: I just found some, they're incredibly expensive.
20:03:35Strife89Come to think of it, I'd be willing to test it, since it's in the Unstable section.
20:05:04 Nick YPSY is now known as Ypsy (
20:05:14gevaertsiHPUser: I think some people have used photofast ssds in irivers
20:05:39bertrikStrife89, whether to update is up to you (I would do it, I'm an update-junkie). It will get upgraded during rockbox installation anyway (during installation the OF is replaced with a patched OF + rockbox bootloader)
20:08:05 Quit JdGordon| (
20:08:05 Quit Strife89 (
20:08:05 Quit bertrik (
20:08:05 Quit DerPapst (
20:08:05 Quit Dgby714 (
20:08:05 Quit reid05 (
20:08:05 Quit fdinel (
20:08:05 Quit Horscht (
20:08:05 Quit robin0800 (
20:08:05 Quit robin0800_ (
20:08:05 Quit Bob_C_ (
20:08:05 Quit AaronM (
20:08:05 Quit amiconn (
20:08:05 Quit kkurbjun (
20:08:05 Quit thegeek (
20:08:05 Quit blithe (
20:08:05 Quit FOAD (
20:08:05 Quit togetic (
20:08:05 Quit BlakeJohnson86 (
20:08:05 Quit topik (
20:08:05 Quit ej0rge (
20:08:05 Quit Tristan (
20:08:05 Quit GodEater (
20:08:05 Quit jordan`` (
20:08:05 Quit jhulst (
20:08:05 Quit ps-auxw (
20:08:05 Quit markun (
20:08:05 Quit rjg (
20:08:05 Quit fish_ (
20:08:05 Quit Overand (
20:08:05 Quit elcan (
20:08:05 Quit fxb__ (
20:08:05 Quit feisar- (
20:08:05 Quit knittl (
20:12:26iHPUsergevaerts: Those are nice looking but still quite expensive. I did find an adapter to make SDHC to CF, I think that will work.
20:12:58gevaertsiHPUser: maybe, maybe not. I tried one of those in my h300, and it didn't work
20:13:04 Join hebz0rl [0] (
20:13:15gevaertsthey're cheap enough to just try of course
20:14:13iHPUserThe one I found has a 16G limit though, smaller than what I currently have
20:15:12 Join Sajber^ [0] (
20:17:47 Quit einhirn ("Miranda IM! Smaller, Faster, Easier.")
20:18:08 Join domonoky1 [0] (
20:18:15 Join antil33t [0] (
20:20:43iHPUserThanks all
20:21:24 Quit funman ("free(random());")
20:22:10NJoinbertrik [0] (
20:22:42 Quit iHPUser ("CGI:IRC")
20:22:51 Join JdGordon| [0] (n=Miranda@nat/microsoft/x-sewrjeiecjqibdwo)
20:23:19 Quit domonoky (Read error: 104 (Connection reset by peer))
20:23:29 Quit flydutch ("/* empty */")
20:24:28 Join ender` [0] (
20:31:04 Quit antil33t1 (Read error: 110 (Connection timed out))
20:35:10NJoinStrife89 [0] (n=michael@
20:39:13 Join kugel [0] (n=kugel@rockbox/developer/kugel)
20:42:32 Quit antil33t (Read error: 145 (Connection timed out))
20:43:07 Quit HellDragon (Read error: 104 (Connection reset by peer))
20:43:19 Join HellDragon [0] (
20:47:26NJoinmarkun [50] (n=markun@rockbox/developer/markun)
20:47:26NJoinOverand [0] (
20:47:26NJoinelcan [0] (
20:47:26NJoinfeisar- [0] (
20:47:26NJoinps-auxw [0] (
20:47:26NJoinfish_ [0] (
20:47:26NJoinfxb__ [0] (
20:47:26NJoinknittl [0] (n=knittl@unaffiliated/knittl)
20:47:26NJoinrjg [0] (
20:48:01NJoinDerPapst [0] (
20:48:01NJoinDgby714 [0] (
20:48:01NJoinreid05 [0] (
20:48:01NJoinfdinel [0] (
20:48:01NJoinHorscht [0] (n=Horscht2@xbmc/user/horscht)
20:48:01NJoinrobin0800 [0] (
20:48:01NJoinrobin0800_ [0] (
20:48:01NJoinBob_C_ [0] (
20:48:01NJoinAaronM [0] (
20:48:01NJoinamiconn [0] (i=quassel@rockbox/developer/amiconn)
20:48:01NJoinkkurbjun [0] (
20:48:01NJointhegeek [0] (
20:48:01NJoinblithe [0] (
20:48:01NJoinFOAD [0] (
20:48:01NJointogetic [0] (n=togetic@unaffiliated/ibuffy)
20:48:01NJoinBlakeJohnson86 [0] (
20:48:01NJoinjhulst [0] (
20:48:01NJoinjordan`` [0] (n=jordan@
20:48:01NJoinGodEater [0] (n=bibble@rockbox/staff/GodEater)
20:48:01NJoinTristan [0] (
20:48:01NJoinej0rge [0] (
20:48:01NJointopik [0] (i=awesome@
20:52:56seaniA theme / language question: if I'm including plain text - "Artist" - in a conditional - %?ia<%ia|Artist Unknown> how do I "nest" this with the new language neutral format?
20:54:01 Quit crwl (
20:54:01 Quit crashd (
20:54:01 Quit tha (
20:54:01 Quit ThomasAH (
20:54:01 Quit Kohlrabi (
20:54:01 Quit bzed (
20:54:01 Quit Dhraakellian (
20:54:01 Quit jds2001 (
20:54:01 Quit CIA-6 (
20:54:02 Join bzed_ [0] (
20:54:03 Join Kohlrabi_ [0] (
20:54:04 Join tah [0] (
20:54:05seaniActually I'll pop that in the forum may be general and fiddly to discuss here
20:54:10NJointha [0] (
20:54:11 Join Dhraakellian [0] (n=ntryon@
20:54:16 Nick bzed_ is now known as bzed (
20:54:19 Join crashd [0] (i=foobar@
20:54:31NJoincrwl [0] (
20:54:37 Join CIA-5 [0] (n=CIA@
20:55:39 Nick tah is now known as ThomasAH (
20:56:25 Join phanboy4 [0] (
20:59:22pixelmaI'm not sure you can have the "unknown" part in this. As I understood you can reuse language strings that exist in the rest of Rockbox (for the menus, info screens, splashes etc), would have to look up in a language file what exists
21:00:03pixelmahaven't tried this tag myself yet though
21:01:22pixelmais this documented yet (CustomWPS or manual)?
21:02:11 Quit goffa (Read error: 113 (No route to host))
21:04:34 Join goffa [0] (n=goffa@
21:05:20 Quit StealthyXIIGer (Connection timed out)
21:06:12seanipixelma: Quite right, JDG is correcting me :-) It's documented in a sticky in the forum.
21:07:15 Join dfkt_ [0] (i=dfkt@unaffiliated/dfkt)
21:07:17pixelmaI wouldn't call that documentation
21:07:29Strife89I'm trying to make the bootloader for the Fize.
21:07:37 Join barrywardell [0] (n=barrywar@rockbox/developer/barrywardell)
21:07:53gevaertsStrife89: the Fize isn't supported
21:08:00Strife89mkamsboot says that it is "unable to open bootloader-fuze.bin".
21:09:17Strife89gevaerts: Good point, but it would appear that you might be hosting a broken piece.
21:09:52 Quit Addison__ ("Leaving")
21:10:26*bertrik should enable SDRAM for meizu m3 tonight
21:11:09gevaertsStrife89: IIRC you need bootloader-fuze.sansa (if you indeed have a fuze, and not a fize like you claimed)
21:11:13pixelmaseani: there is also a generic "unknown" string so you could combine the two but the resulting combination might not work (nicely) in all languages
21:11:37Strife89gevaerts: Ah, I didn't notice the typo. Yes, I have a *Fuze*. :)
21:11:38 Join Stephen_ [0] (
21:12:19Strife89gevaerts: I downloaded that file, using the link in the Fuze manual, and followed the instructions to combine it with the Fuze OF with mkamsboot.
21:12:54gevaertsStrife89: yes, I mean you need to run "mkamsboot fuzea.bin bootloader-fuze.sansa patched.bin "
21:13:25***Saving seen data "./dancer.seen"
21:13:46Strife89Ah, whoops.
21:14:02Strife89I had typed "bootloader-fuze.bin" instead.
21:14:15Strife89The idiot strikes again!
21:14:31gevaertsthat's what the manual says :)
21:15:10pixelmaseani: for a reference
21:15:12 Quit kugel (Read error: 60 (Operation timed out))
21:15:17Strife89That I'm an idiot, or what?
21:15:39gevaertsno, the manual says bootloader-fuze.bin, but that's wrong
21:15:47Strife89I see now.
21:16:49seanipixelma: Thanks, JDG makes a good point that if I restyle it as "Artist: Unknown" it's would probably work with two separate translated string, unlike "Unknown Artist" as a "phrase"
21:17:32gevaertsStrife89: please tell us if it works :)
21:17:51Strife89gevaerts: The Fuze is in the Firmware Upgrade process now.
21:18:53pixelmaseani: more likely yes
21:19:02 Join Nz17 [0] (
21:19:35Nz17Which is better for a Rockbox user, Sansa e200 or c200?
21:20:40Strife89Most likely an e200, but as gevaerts said, it depends.
21:20:48Nz17Well I want long battery life, FM usage, and Ogg/FLAC.
21:20:49bertrike200 has bigger and higher quality screen, but is also bigger and heavier
21:21:03pixelmaand has a scroll wheel
21:21:03Nz17Weight and size doesn't matter.
21:21:12Nz17Ooh, scrollable. :)
21:21:23gevaertsI think the e200 has a bigger battery than the c200
21:21:37Nz17That's what the table seems to indicate.
21:21:46Strife89Nz17: The e200 lasts about 4-6 hours longer than a c200, I think.
21:22:02Nz17I see.
21:22:14pixelmabattery life with the c200 is really ok too though
21:22:26Nz17I guess what I'm also trying to do is find something cheap too. ;)
21:22:46gevaertsalso, while I think the v1 is probably still preferable for normal users, the e200v2 should work without problems, while the c200v2 port isn't as far yet
21:23:15Nz17That's important to remember.
21:23:23pixelmaone other advantage of the e200 would be that there are ones with a bigger internal memory. Although with a microSDHC that gets a little less important
21:24:02Nz17Well I'm going to see if eBay has any listings for wall and car adapters for the e200 then. :D
21:24:21 Quit dfkt (Nick collision from services.)
21:24:29 Nick dfkt_ is now known as dfkt (i=dfkt@unaffiliated/dfkt)
21:24:33pixelmae200s exist with or with radio, while c200 seem to be sold with a label "with or without radio" but still both have the hardware (at least all reports I read suggest that)
21:25:11Strife89gevaerts: It boots nicely - in 1.5 seconds flat!
21:25:57Nz17Nice. But there is a warning about FM on the table at .
21:26:03CIA-5New commit by gevaerts (r23611): The downloaded bootloader is bootloader-e200v2.sansa or bootloader-fuze.sansa, not bootloader-e200v2.bin or bootloader-fuze.bin
21:26:15bertrikTheSeven, I think you implemented pcm_play_dma_get_peak_buffer, right? there's still a comment saying it isn't implemented
21:27:35bertrikshall I remove it (the comment)?
21:28:04Strife89gevaerts: I don't see anything on that page about the Fuze. Perhaps you meant this?
21:28:47gevaertsStrife89: huh?
21:28:59TheSevenbertrik: you may do so :-)
21:29:06Strife89gevaerts: Whoops, wrong person. ^^;
21:29:15TheSeventhis is (of course) still true for the corresponding rec function though
21:29:38Strife89gevaerts: I thought Nz17 was speaking to me back there. "Nice, but there's a warning on the radio."
21:29:48Nz17I was.
21:29:58Nz17It says, "Some" for the e200 series.
21:30:55CIA-5New commit by bertrik (r23612): pcm-s5l8700: add missing header file, remove out-of-date comment
21:31:13TheSevenwhich header was missing?
21:31:33Strife89Nz17: Ah. 1) I thought you were speaking about the Fuze I just set up. 2) Right; Rockbox fails to detect the FM Radio on some e200s.
21:32:25gevaertsStrife89: does it?
21:32:28Nz17I wish these hardware companies would take a lesson in consistency and external markings.
21:32:34Strife89gevaerts: Does it what?
21:32:41gevaertsThat table means that they don't all have FM
21:32:42 Join shotofadds [0] (n=rob@rockbox/developer/shotofadds)
21:33:01gevaertsStrife89: fail to detect FM
21:33:02Strife89gevaerts, Nz17 : Ah; my mistake.
21:33:34Nz17I guess I'd better be careful about which one I buy from eBay.
21:33:43Nz17Unless you guys have a favorite you'd like to recommend.
21:34:10seaniNz17: Likely not in their interests. If customers differentiate on an unadvertised / inadvertent / unsupported feature or side effect, they could find themselves with more stock on their hands
21:34:44Nz17I guess it's just not fair by design.
21:35:17 Quit robin0800 (Remote closed the connection)
21:36:02seaniNz17: Life in general, then.
21:36:23Stephen_I ordered from here and got v2 with FM -
21:37:17shotofaddswhat's the policy on bug reports in Flyspray for unstable targets? is that ok?
21:37:27gevaertsI think so
21:37:31Nz17I can't decide if this car adapter is brilliant or stupid:
21:37:46gevaertsThe rule was about supported targets, which unstable targets are
21:38:18 Join bmbl [0] (n=Miranda@unaffiliated/bmbl)
21:38:20*shotofadds has a look
21:39:23 Quit barrywardell ()
21:39:30shotofaddswell, the FlySpray doesn't have any rules at all ;)
21:39:39shotofadds*wiki page
21:41:47TheSevenNz17: this adapter looks very similar to one from belkin
21:42:13TheSevenjust a white case instead of a black one and the led in a slightly different position (and red instead of blue)
21:42:36Nz17Gah, I hate blue LEDs - they are so bright!
21:43:08TheSeventhe belkin one has the led internally and just a translucent piece of plastic to the outside where it shines through - looks nice and isn't too bright
21:43:18 Join adiroiban [0] (
21:43:30TheSevenit's a blueish-glowing light strip instead of a red point
21:43:59Nz17Oh, maybe I should check it out.
21:44:09Nz17Either one, I'm starting to think the adapter is brilliant - it can be used with any PMP that charges via USB. The only problem is you need your device's USB cable to use it.
21:44:27pixelmathe e200 (and c200) has blue LEDs too ;)
21:44:30TorneI suspect your claim there is mistaken ;)
21:45:01TorneUSB charging is, to my great disappoitnment, way more complicated than you might think
21:45:28Nz17Yeah... wait, huh?
21:45:45TheSeventhe belkin one is pulling the data pins to iirc 2.8 and 2.0 volts to make at least the zen and the ipods even recognize it
21:46:00Tornethat still doesn't cover everything, no?
21:46:03CIA-5New commit by bluebrother (r23613): Fix missing space caused by incorrect usage of the \dap macro (FS #10774 by Diego Herranz).
21:46:15Nz17I thought USB is supposed to be 5v?
21:46:20Tornethere's also D+/D- shorting for very recent things that follow the ccharging spec
21:46:25TheSevenNz17: yes, at the power pins
21:46:29TorneNz17: Charging from USB is generally a spec violation.
21:46:42TorneNz17: people make this less bad by trying to detect the charger specifically
21:46:46TheSevenTorne: not neccessarily
21:46:52TorneTheSeven: well yes. we've been through this one ;)
21:46:57TorneThe way many devices do it is a spec violation.
21:47:17Tornesome people are not quite that brave and do soemthing that's still wrong but less likely to be horrible :)
21:47:21TheSevenactually, what the ipods are doing is rejecting non-apple chargers while *still* violating the spec...
21:47:36TorneWell, they don't *now*
21:47:49Tornethe current devices use the real charging spec
21:48:01TheSevenreally? so they don't work with old apple chargers? nice one.
21:48:47Tornesee iphone
21:48:51Nz17Ooh that Apple - if it didn't have the "cool" image it wouldn't get away with half of what it does.
21:48:53Tornethere are also some inbetween ones which do both
21:49:06Torneand some apple chargers *implement* both, with a crazy resistor network
21:49:16Tornei've had great fun finding all this out, and trying to decide what we should do ;)
21:49:47Tornei can't say i've decided.
21:49:57Tornei mean not tha tit's down to me but i've not decided what my opinion is ;)
21:51:51TheSevenhm, my opinion is to allow the user to choose either the standards compliant way or the "drain whatever you get" way
21:51:59 Quit robin0800_ (Connection timed out)
21:52:28TheSevenand maybe a third choice that still tries to obey the 100mA limit if it enumerates that way but drains 500mA if the host isn't responding at all
21:52:42Torneyeah i was gonna try for some compromise
21:52:47Tornebut haven't ahd time/motivation to implement any of the above yet :)
21:52:53TheSeventhings get even worse if we also take into account that some devices can do 100/500/1000mA nowadays
21:53:23TheSeven(with 1000 being only allowed from charging-spec compliant chargers and USB3)
21:53:26TorneTheSeven: drawing more than 500 without the thing on the other end shorting D+/D- or ossibly being an apple charger is probably a very bad plan
21:53:37Torneeven if the user tells us to :)
21:53:49TheSevenmost chargers support 1000
21:54:11Torneyes, but if you don't know it is one, then that's quite a lot worse than just pulling 500 without knowing, no?
21:54:24Tornebecause loads of devices do that already so most of hte bad usb ports have been weeded out by natural selection ;)
21:54:33Nz17You guys have to deal with way too much trouble. ;)
21:54:41Torne(clearly darwinian evolution is the way to decide specs)
21:55:34Nz17Don't say that or IE will get you.
21:56:19TheSevenhm, if you do 1000 on a charger that doesnt support it, you will hit it's inner resistance and thus drop the voltage a little, making it in fact draw around those 500mA the charger supports but moving heat dissipation from the PMU to the charger :-)
21:56:28 Join kugel [0] (n=kugel@rockbox/developer/kugel)
21:56:44TorneTheSeven: Hooray :)
21:56:56TorneTheSeven: This is one of those "no good answer" things.
21:57:20TorneEspecially givent hat *even* being spec compliant can still cause bad things to happen. Its not *your fault* then, but hte bad things can still happen while your device is plugged in :)
21:57:43Torneso *someone* will blame you.
21:57:56TheSevenbasically the worst that can happen is the charger getting blown :-P
21:58:07TorneNot at all :)
21:58:13TorneThink of dodgy ports on laptops, etc
21:58:14TheSevenwhich is some way of darwinian evolution as you said above :-D
21:58:20TorneSee if you can mmelt the entire motehrboard :)
21:58:40Torneyou can do way worse than crisping a charger if you have a sufficiently pathological combination of stupid broken hardware.
21:59:02TheSevenports on laptops will enumerate properly, and even if they don't, they can either supply 1000mA (lots of external hdds need this) or will just shut off the port if we're sucking too much. we may re-try drawing 500mA then. :-P
22:00:08TheSeventhe worst thing i've seen (besides on acer motherboards that indeed blow up sometimes when you plug an innocent pendrive) is non-resetting fuses :-D
22:06:49 Quit freddyb (Remote closed the connection)
22:08:25bertrikit looks like sdram now works on the meizu m3 too
22:09:27TorneTheSeven: Oh sure, it has to be *really broken* hardware, to the point that it's not really in any way the devices' fault
22:09:31Tornebut people still blame the devices :)
22:09:39Tornewhatever was connected at the time is *obviously* the cause!
22:09:58Tornei guess we can offer them a refund :)
22:10:47*gevaerts thinks that those refund jokes could backfire
22:11:03gevaertsWhat will you do if someone actually paid for rockbox?
22:11:23TorneTHey didn't pay *us*, presumably
22:13:50gevaertswell, no
22:14:04Nz17But what if they did?
22:14:28Nz17Maybe someone is nutty enough to appreciate your efforts and in a fit of madness, MADNESS, gave $5?
22:14:38gevaertsthat's a donation
22:14:40Tornethat's not paying for it, that's a donation
22:15:18Nz17What do you want? Them to "donate" and then say in the comment section, "Software please."?
22:16:15gevaertsthe point is that someone can sell rockbox to people, and if they then ask for support here and someone (jokingly) proposes to refund what this person paid, it will be awkward
22:16:30 Quit Grahack ("Leaving.")
22:16:40Nz17I know.
22:16:42Nz17Though... it would be cool if, as a fund-raising endeavor the Rockbox group sold pre-modded players.
22:17:08Bob_C_fascinating legal issues there!
22:17:13bertrikwe don't really need money I think
22:17:16Nz17Then at least people wouldn't need to hunt on eBay and such, just find them centrally and make the group a few dinero.
22:17:16TheSevenNz17: which may imply warranties etc.
22:17:34TorneTher eare hilarious legal issues with that one
22:17:45TorneSteering clear is best :)
22:17:53Nz17Nah, just say, "Hey this is aftermarket, no warranty, as-is, but we will do our part to help out if reasonable."
22:18:06Tornethere are issues other than warranty/etc.
22:18:08TheSevenNz17: this is not possible in several countries
22:18:16Torneor at least, *potentially*
22:18:19bertrikyou'd probably have to pay taxes too
22:18:26Torneand we don't ahve the wherewithal to prove whether there are or not ;)
22:18:27Nz17Those countries can go back to the medieval ages.
22:18:34Bob_C_You can say that, but it doesn't override consumer laws
22:18:59Nz17Then there is no choice.
22:19:14Nz17"High enough" donations get various levels of "gift" players.
22:19:21Nz17Problem solved.
22:19:27Bob_C_An advantage of open software is not having to sell stuff
22:19:36TheSevenin Germany at least, and I think this is actually EU law, you *always* have to provide at least 1 year of warranty on *used* goods and 2 years on *new* ones...
22:20:23Nz17Can your warranty be to guarantee that it might fail?
22:21:22TheSevenif anything, we could offer some kind of rockboxing service for devices that users send in, but this is still pretty risky from several legal aspects
22:21:29 Join B4gder [241] (n=daniel@rockbox/developer/bagder)
22:21:57Nz17Yeah, but that loses the appeal.
22:22:05Nz17You'd still need to find compatible players.
22:22:39TheSeventhis could be solved easier, trust me.
22:22:48 Quit DerPapst ("Leaving.")
22:22:49TheSevenby just supporting players that are still being sold
22:23:08Tornewell people are *trying*
22:23:19Tornesansa aren't making it easy ;)
22:23:27TheSevennobody is making it easy
22:23:33gevaertsNz17: so why would we want to make it easier for other people to get players with rockbox?
22:23:35n1siirc the ipod video was the last target that got a usable port while still being sold new
22:24:08shotofaddsThe D2's dead easy to hack and are still made, but no other devs seem interested :/
22:24:16Nz17gevaerts: I suppose because FREEDOM FEELS GOOD. ;)
22:24:19 Quit phanboy4 ("Leaving")
22:24:33gevaertsNz17: but we are already free!
22:24:42*TheSeven thinks the first device that will both be sold and supported for a long time simultaneously, if it will ever work, is the lyre.
22:25:02Nz17Let us free the other proletariats!
22:25:29gevaertsNz17: my point is that if you want to do this sort of thing, go ahead, but it's not something the rockbox project is interested in
22:25:50*bertrik agrees with gevaerts
22:25:53Nz17That's not 100% true.
22:26:01Nz17If so, the code wouldn't be public.
22:26:11Nz17The group wants people to use the software.
22:26:31Nz17Why what?
22:26:57gevaertsIf people want to use it, they're free to do so, but why would we necessarily want *more* users?
22:27:10Nz17Would you want less?
22:27:20TheSevenNz17: sometimes yes
22:27:21B4gderwe want more robots!
22:27:24TheSevenat least for the player :-P
22:27:38*TheSeven would like to abolish charcell lcds :-P
22:27:39TorneNz17: Users often clutter up the nice tidy lists/forums with useless stuff :)
22:27:58Nz17But who would be so bold? ;)
22:28:01TorneNz17: Open source projects often *really do not care* if they have more users
22:28:11Nz17Users beget developers, developers beget code.
22:28:14Tornei work on rockbox because it makes *my mp3 player* awesome
22:28:27Torneand then a bit more because it's kinda fun
22:28:39Nz17And code begets new features.
22:28:55TorneUsers generally *don't* become developers, though
22:29:12 Quit HellDragon (Client Quit)
22:29:13Torneand even if they do, the chance of them working on anything I'm that bothered about is not huge ;)
22:29:20*Torne is being cynical somewhat, but you see the point :)
22:29:23TheSevenTorne: developers don't become code either :-D
22:29:29TorneYeah, i hope not
22:29:44Nz17Become? I wrote, "beget."
22:30:02TorneIn which case I hope users don't go around begetting developers too much :)
22:30:02Nz17But aren't most developers users first?
22:30:16*B4gder was a developer first
22:30:27TorneNz17: not necessarily
22:30:28*gevaerts was a user for more than a month
22:30:35TorneNz17: i got my ipodvideo to run rockbox
22:30:36Bob_C_Users beget support questions... so many questions...
22:30:40Tornei started building rockbox a couple of days later
22:30:50TorneI knew when I bought it that I was very likely to want to make at least minor changes
22:31:19Nz17And then this happened!
22:31:20Torneand iknew that i would probably be able to
22:31:26Torneso, i pretty much planned on *doing some development* for it.
22:31:40CIA-5New commit by bertrik (r23614): Meizu M3: configure and use SDRAM
22:31:42Tornei didn't actually plan on becoming a committer but they're nice guys and gevaerts bullied me
22:31:58Nz17Plus the benefits.
22:32:18Tornebenefits? :)
22:32:24 Join HellDragon [0] (
22:32:36Nz17Yeah, you know, like... such as... something kind... um, groupies?
22:33:02B4gderhe already mentioined gevaerts :-P
22:33:11*B4gder ducks
22:33:21*gevaerts throws eggs at B4gder
22:33:24 Quit Omlet ("( :: NoNameScript 4.22 :: )")
22:33:51TorneNz17: well it means i can gradually reduce the number of patches i need to apply to build for my ipod, by finishing and committing them ;)
22:34:01Tornethat is a benefit, tbh :)
22:34:06Tornethe stack was getting unmanageable
22:34:26Nz17Not to mention the buffer. *rimshot*
22:34:41 Join LambdaCalculus37 [0] (n=LambdaCa@rockbox/staff/LambdaCalculus37)
22:38:44shotofaddsNow this is ridiculous. I tried to get ROLO working again on the D2 to do some testing, of course this needed a build with IRAM enabled - that's easy since IRAM is only disabled ATM for speed reasons. But I'm getting some pretty weird side effects.
22:39:04shotofaddsLike the UI will freeze after the backlight's faded out due to inactivity. Music plays on, but I can't wake the UI. WTF?!
22:41:01Nz17OK guys, thanks for the advice and good conversation. Happy coding!
22:41:14 Part Nz17 ("to buy a player.")
22:42:57 Nick fxb__ is now known as fxb (
22:44:00 Join phanboy4 [0] (
22:45:31shotofaddsAs for ROLO, it still won't work due to some stupidity: the rolo_restart() is in IRAM, but it calls invalidate_idcache() which isn't (and has since been overwritten with the new ROLO'd image). D'oh...
22:46:04shotofaddsshouldn't those ARM cache functions be in .icode?
22:47:23TheSevensounds reasonable, as this is probably breaking rolo in even more places
22:47:43TheSevenit's probably the bug I'm observing on nn2g right now that rolo sometimes locks up (as soon as that function moved?)
22:49:23shotofaddssounds like it. presumably just changing those ".section .text" lines in mmu-arm.S to .icode should do it... if .icode even exists on all ARM :/
22:50:07*kugel wonders why other ARMs don't have that problem
22:51:09TheSevenkugel: because they forgot to flush the cache? :-P
22:51:12bertrikI guess rolo is probably not used that much so any problems are not reported
22:51:37kugelTheSeven: that's generic arm code in rolo.c as far as I know
22:51:53TheSevenbertrik: it's probably used by developers that always rolo very similar builds and then grumble from time to time if it fails 1 out of 10 times
22:52:37n1syeah, if the functions are "early" in the binary there's a good chance the addresses will match
22:52:45TheSevenshotofadds: .icode should exist on every arm target. if it doesn't, it should be added to the linker script in the .text output section
22:52:52kugelI assume it's used rather often (on targets that have usb), it kicks in if you update rockbox with rockbox usb
22:53:14TheSevenlet me check where that function is for me
22:53:25Bob_C_I don't think the Gigabeat F/X use IRAM do they?
22:53:50shotofaddsooh look, ROLO works now ;-)
22:54:15shotofaddsBob_C_: but .icode is still defined, I did check that one at least
22:54:16 Quit kugel (Nick collision from services.)
22:54:24 Join kugel [0] (
22:54:49n1sBob_C_: they do use it for rolo
22:54:49n1sfor this exact reason (code can't overwrite itself)
22:54:49n1sthe iram on those chips is small though, 4kB iirc
22:54:52TheSeven0x08063a08 invalidate_idcache
22:54:57TheSevenbroken for me, too
22:55:22*shotofadds wonders how many reds this might get
22:55:37Bob_C_I tried ROLO a couple of times without success, but that was on some dodgy builds
22:55:48 Join kugel_ [0] (
22:55:54 Quit kugel (Nick collision from services.)
22:56:06TheSevenand it's pretty much at the end of the image, so it should fail rather often
22:56:47 Join kugel__ [0] (
22:56:50 Nick kugel__ is now known as kugel (
22:56:53 Quit kugel (Read error: 104 (Connection reset by peer))
22:57:13shotofaddspresumably I just need to change invalidate_idcache and invalidate_dcache, as those are the only ones called from ROLO. Any benefit in moving the others?
22:57:35 Quit adiroiban (Read error: 60 (Operation timed out))
22:57:56 Join kugel [0] (n=kugel@rockbox/developer/kugel)
22:58:28 Quit kugel_ (Read error: 60 (Operation timed out))
22:58:36 Join adiroiban [0] (
22:59:11 Join JdGordon|| [0] (n=Miranda@nat/microsoft/x-vudzhukktxaksaum)
22:59:34TheSevenshotofadds: for consistency at best, but you'll need to watch for the gigabeat iram size
23:00:01*shotofadds does a couple of test builds
23:00:04 Join kugel_ [0] (
23:00:08 Quit kugel (Nick collision from services.)
23:00:12 Nick kugel_ is now known as kugel (
23:00:24kugelthe question why it works on other targets is still open (?)
23:00:30kugelmaybe we should have a section that is protected against rolo and move it there instead of relying on magic/iram
23:00:59shotofaddsI don't think relying on IRAM is magic
23:01:24TheSevenkugel: it's definitely broken for me in a way that should fail almost always (judging from it's address) but it still works 90% of the times
23:01:38kugelit is since many targets can remap it whereever they want
23:02:02shotofaddsbut why's that a problem?
23:02:24n1skugel: the point is that iram is not overwritten untill rockbox starts
23:03:11kugelanyway, some targets might not even have iram/or too little
23:03:34TheSevenkugel: then these have either some extra precautions or rolo just doesn't work for them
23:03:44TheSevenat least for the targets that don't have any iram
23:03:53 Quit JdGordon| (Read error: 60 (Operation timed out))
23:04:11shotofaddsthis won't make any difference on F/X anyway since their .icode is inside .text
23:04:16kugelso it's luck that it works on all other targets for years?
23:04:43TheSevenkugel: seems so
23:04:50shotofaddskugel: this error is in mmu-arm.S so only relevant to ARMv4/v%
23:05:18 Quit Lynx_ (Read error: 54 (Connection reset by peer))
23:05:21kugelthat basically only excludes PP
23:05:38TheSevenshotofadds: the v6 one was split off only very recently and probably has the same issues (if it isn't even using mmu-arm.S for the common things)
23:06:08 Join Lynx_ [0] (
23:06:39kugelshotofadds: these days we have almost more targets that actually have a caches than those that dont't I believe
23:07:30shotofaddsyes, but this is a sane fix regardless
23:08:09kugelgo for it, I'm just curious that it never caused problems
23:08:22shotofaddswhich targets are ARMv6? just Gigabeat S?
23:09:11shotofaddshmm. it won't make any difference there either since IRAM isn't enabled on the S either.
23:09:28shotofaddsROLO has no hope on the Gigabeats at the moment, by the looks of it
23:10:01*shotofadds wonders whether kugel's "safe area" might be a better bet ;)
23:10:02TheSevenshotofadds: they could just use some fake iram at the end of dram for rolo's purposes
23:10:02kugela separate section makes it work on all targets
23:10:41shotofaddsbah, that's a lot of .lds's to edit :/
23:11:29 Join killan [0] (
23:12:29 Quit killan_ (Read error: 104 (Connection reset by peer))
23:13:27***Saving seen data "./dancer.seen"
23:15:54kugelcouldn't it be fixed by just invalidating caches before loading the "huge" binary?
23:17:07 Quit killan ("( :: NoNameScript 4.22 :: )")
23:17:33 Nick JdGordon|| is now known as Jdgordon| (n=Miranda@nat/microsoft/x-vudzhukktxaksaum)
23:17:37 Nick Jdgordon| is now known as JdGordon| (n=Miranda@nat/microsoft/x-vudzhukktxaksaum)
23:17:45shotofaddskugel: no, there's a final copy stage after loading the binary (from audiobuf to DRAM start addr)
23:17:51 Join killan [0] (
23:18:25kugelinvalidating before that final copy?
23:19:10kugelor even turning the caches off entirely
23:20:06*kugel doesn't think it's a good idea to call functions after they're overwritten generally ;)
23:21:35shotofaddshaving cache functions available at all times is generally a Good Thing, no?
23:22:17Bob_C_turning off dcache also means no MMU?
23:22:26kugelall Good Things have an exception. Are the caches critical for rolo to work?
23:22:44 Quit jgarvey ("Leaving")
23:23:57kugelBob_C_: AFAIK, no
23:24:23Bob_C_oh, afaik it does :)
23:24:42Bob_C_perhaps depends on arch
23:27:24 Quit Strife89 ("Going home.")
23:27:29kugelBob_C_: mmu, dcache and icache all have separate enable/disable registers
23:27:46 Quit evilnick_B ("Page closed")
23:30:00Bob_C_ok, in that case it should be ok
23:34:24 Join shot0fadds [0] (
23:34:56 Quit phanboy4 (Read error: 104 (Connection reset by peer))
23:35:17 Join stripwax [0] (
23:35:17 Quit n1s ("Lämnar")
23:36:24JdGordon|can I convince anyone to work on please??
23:38:34 Nick Ypsy is now known as YPSY (
23:39:33*shot0fadds doesn't think adding a .safe section to ALL targets is a very clever idea. take a look at imx31/rolo_restart.S for a rather easier solution...
23:40:45*shot0fadds goes for a cuppa
23:42:33kugelJdGordon|: I thought I made it work with sbs?
23:43:20JdGordon|I need it fixed so it can handle a single .sbs for all targets unless there is one for -rechwcodec and -recswcodec or something similar
23:44:02kugelwhy exclude those?
23:44:13JdGordon|they would be from a different file
23:44:13kugelI thought you added a number of redundant tags to make that workj
23:45:51kugelwhat where those recording tags for then?
23:45:57JdGordon|it could be done using feature tags, but only if one is added for swcodec and hwcodec
23:46:05JdGordon|they have different tokens
23:46:43JdGordon|and anyway... it was ONE "redundant" tag...
23:46:44kugelwhich tags exactly?
23:48:31 Quit shotofadds (Read error: 110 (Connection timed out))
23:49:25 Quit bertrik ("De groeten")
23:49:26JdGordon|which are you talking about?
23:49:30*JdGordon| is a bit confused
23:50:06CIA-5New commit by mc2739 (r23615): Add Diego Herranz to docs/CREDITS for r23613 / FS #10773
23:51:42kugelJdGordon|: which ones are incompatible?
23:52:12JdGordon|off the top of my head I couldnt tell you...
23:52:36JdGordon|I tihnk actually only one of the tags makes sense in both
23:53:02JdGordon|and, unlike rtc, swcodec and hwcodec are mutually exclusive so there is no need to compile in the tags for both of them
23:53:51 Quit kugel (Remote closed the connection)
23:55:22 Quit domonoky1 (Read error: 104 (Connection reset by peer))

Previous day | Next day