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 2009-10-29

00:03:11 Quit cas ("enjoying heekscnc software for CAD-CAM on his new CNC machine ;-)")
00:05:05CIA-85New commit by mcuelenaere (r23388): Add support to for Lua scripts
00:05:14CIA-85New commit by mcuelenaere (r23389): Add Boomshine to the games category
00:05:29 Quit mikroflops (Read error: 110 (Connection timed out))
00:06:26 Quit bertrik ("De groeten")
00:11:34 Quit bmbl ("Bye!")
00:15:59robin0800any one seen this and know what they want?
00:17:00 Join tanti [0] (n=me@
00:17:49 Quit saLOUt (Remote closed the connection)
00:21:15kugelJdGordon|: and "the area not being used by the sbs. " is %Vi
00:22:07JdGordon|the area not being used is.... the area not being used
00:22:15JdGordon|its not being drawn into by the sbs
00:22:26JdGordon|but the sbs doesnt need to explicitly say "im not using this area"
00:23:00JdGordon|which really does make it complelty seperate from both the UI area idea, and themes in general
00:23:23kugelyea, I've come across that idea too, and to the conclusion that it's next to impossible (and complex) for the sbs to predict a useful unused area
00:23:49kugelin a single viewport
00:24:05JdGordon|hang on... wha?
00:24:19stripwaxfor list viewer with a background bmp, do we print 'filled' pixels in foreground and 'blank' pixels in background; or do we blt the whole bitmap and then only print 'filled' pixels in foreground?
00:24:36stripwax^ unrelated to above viewport discussion (sorry)
00:24:45JdGordon|stripwax: 1) shoosh!, 2) yes :)
00:25:29stripwaxLooks like we explicitly print the blank pixels using the background bmp pixels .. which surely is less efficient than a giant memcpy followed by printing just what we need.
00:25:33stripwaxJdGordon : - :)
00:25:35JdGordon|kugel: what predicting? the themer knows where the sbs doesnt draw in... in general its a rectangle in the center of the display... so that goes in the setting
00:25:36kugelthe backdrop is in a separate buffer. depending on the drawmode some loops decide which value to take (fg, bg or backdrop) or even mix
00:25:50 Quit tanti ("[TalkSoup] Get it today:")
00:26:16kugelJdGordon|: oh I thought you would suggest generating %Vi at runtime
00:26:37stripwaxkugel - right, and it seems that the listviewer chooses to print text using 'filled=fg, blank=backgrop' rather than 'filled=fg, blank=don't even do anything'
00:26:40JdGordon|it would be gone! kaput! dead! removed! extinct!
00:26:58scorchethis %Vi would be no more!
00:27:05scorcheceased to be?
00:27:26kugelit surely doesn't belong in the settings
00:27:44JdGordon|"ui viewport" does exactly that!
00:27:49 Join Strife89 [0] (
00:28:39kugelback at the beginning :/
00:28:49JdGordon|looks like it :<
00:28:51 Nick uflops is now known as mikroflops (
00:29:27kugelcan you explain how it makes sense to have a arbitary value that's only dependant on the skin in the user settings?
00:29:34kugelsbs even
00:29:51JdGordon|I have in the ml.. 15 or 16 times!
00:29:55JdGordon|but sure... once more
00:30:08JdGordon|the sbs knows where it wants to draw in...
00:30:18JdGordon|the lists will happily draw where they are told
00:30:30JdGordon|the themer knows where it wants the list, and where the sbs doesnt draw
00:30:47JdGordon|therefore, put the list viewport in the theme and then the sbs is compeltly detahced from it
00:31:21JdGordon|now if you come along and like my sbs, but want a *smaller* ui area, you only need to change the theme
00:31:40kugelso, tying sbs to a single theme + abusing a completely inrelated feature + disrespect a user setting
00:31:44CIA-85New commit by mcuelenaere (r23390): Add support for running Lua games/apps from the Plugins menu
00:31:47CIA-85New commit by mcuelenaere (r23391): Remove unused #define and correct comment.
00:32:12JdGordon|did you not read? the sbs is completly detached from any theme
00:32:15kugelJdGordon|: that's how it works currently. just edit the theme
00:32:30JdGordon|I disagree that sbs + ui area are complelty seperate features
00:32:40JdGordon|and disrespect the user setting? i dont follow
00:33:31kugelHow can I just load a nice statusbar onto cabbiev2 that I didn't write myself. I need to actually find out the area myself, and edit cabbiev2?
00:34:06JdGordon|you said yourself that the setting is prioritised over %Vi....
00:34:17kugelI need to edit ever single theme that didn't even use UI vp before?
00:34:32JdGordon|you would get a nice GUI to do it
00:35:07JdGordon|from your email... "- UI vp is the ultimate viewport for the UI excluding skins (plugins override use it currently, but I'm planning to fix that). It's the viewport that has the top priority for the UI. That's what it has been committed for."
00:35:15JdGordon|which means %Vi is redundant!
00:35:24kugelreplacing a super simple info viewport with complete annoying bloat sounds bad to me
00:35:36JdGordon|oy gevalt
00:36:04kugelthe point is, a theme doesn't need to be forced to use a UI vp
00:36:14 Join ShapeShifter499 [0] (
00:36:39JdGordon|back to square one :/
00:36:58JdGordon|forcing the user to specify a viewport isnt a big deal
00:37:06kugel%Vi isn't a big deal
00:37:15JdGordon|its redundant!
00:37:20JdGordon|and inconsistant
00:37:22*gevaerts must admit that he also doesn't see at all why there need to be two ways to say exactly the same thing
00:38:31kugel%Vi is also always used for the WPS as default viewport. Do you want to waste space there by using and requiring an unrelated possibly tiny UI vp?
00:38:47kugelgevaerts: we say the same thing?
00:38:59JdGordon|yes... %Vi should definetly not be used by the WPS.... it shuold use the same as the user setting...
00:39:00gevaertskugel: do we?
00:39:18JdGordon|I would have no problem with %Vi as a fallback if the WPS and lists used the same logic
00:39:51mcuelenaereShould do_menu() be able to handle menus with a STRINGTABLE of 150 strings?
00:40:03*mcuelenaere is having crashes in the sim with Lua and do_menu
00:40:16 Quit esperegu_ (Read error: 113 (No route to host))
00:40:26JdGordon|mcuelenaere: there may be a limit there.. honestly you'd have to check the menu.c code thoguh
00:40:34JdGordon|I havnt looked at the code in months
00:40:58 Quit TopyMobile__ (Read error: 113 (No route to host))
00:41:15kugelthe wps was always completely separated from the rest of the UI, w.r.t. to viewports and eveything
00:41:27 Quit mcuelenaere ()
00:41:30JdGordon|no its not
00:41:35kugelalways, not just since sbs
00:41:40JdGordon|and if it *is* now.. it shouldnt be
00:42:02JdGordon|no it wasnt... its default was always the same viewport that the lists are in
00:42:13JdGordon|if the bar is there it doesnt overlap it untill it sees %wd
00:42:22kugelback when the lists where plain fullscreen maybe
00:42:23JdGordon|s/bar/sbs/ and it starts making sense
00:42:29gevaertskugel: if %Vi is "the viewport that doesn't interfere with the status bar", for which subsystems exactly is it intended?
00:42:48JdGordon|kugel: in that case... you're ui viewport patch was incorrect
00:42:54JdGordon|wrong your
00:43:51 Join TopyMobile__ [0] (
00:44:02JdGordon|although.... *pre* sbs I can understand why it would work like that
00:44:32 Quit petur ("Zzzzzz")
00:45:03kugeland now not anymore?
00:45:12kugelthat puzzles me
00:45:47 Quit robin0800 (Remote closed the connection)
00:45:58kugelgevaerts: basically, everything. it's only overridden by specifying viewports (and deactivating the status bar of course)
00:46:06JdGordon|because before you were very limited with when you would want to use the ui viewport, it makes sense that you wouldnt use it in the wps... but now, with sbs, there is a much bigger reason to move the ui area
00:46:34JdGordon|in all honesty, we should force the WPS to default to using the fullscreen, and not enable sbs unless %we is there
00:46:39JdGordon|but that will break old themes
00:46:47JdGordon|that imo is the correct change
00:46:58gevaertskugel: if it's overridden, that means that the setting means the same thing, right?
00:47:35kugelyou mean even deactivate when neither %we nor %wd (which is supposed to follow user settings)?
00:48:01kugelgevaerts: yea, but only where the UI vp actually applies. and that's not the WPS for example
00:48:17kugelstrange ideas you get
00:48:22JdGordon|not sbs!!! USER SETTING!
00:48:58JdGordon|you cant pick and choose when to use what.... it gets confusing (oh think of the chil^H^H^H^Husers)
00:48:59kugel%we/%we only changes whether the status bar appears or not
00:49:15JdGordon|the statusbar doesnt mean anything anymore
00:49:18JdGordon|its DEAD!
00:49:30kugel%we/%wd also works for the sbs
00:49:31JdGordon|those tags now mean something else which we need to decide
00:49:48JdGordon|then they should use the same viewport the ui is put in... the user setting
00:50:11kugelwait a second, let me sum up
00:50:14JdGordon|(or fall back to the %Vi if its the same behind-the-scenes0logic as the ui area)
00:51:36kugelyou said you would accept %Vi if the wps would default to UI vp (that implies %Vi keeps alive). Then you said you understand that pre-sbs you would want the fullscreen for the WPS. Therefore, if %Vi stays, it's the best choice for the WPS (as fullscreen as possible, doesn't draw over sbs)
00:51:54gevaertskugel: the UI vp is meant to allow for things like margins that align with the backdrop, while %Vi is meant to indicate an area the sbs is guaranteed not to touch, right?
00:52:02*gevaerts is still exploring the area
00:52:13kugelgevaerts: exactly
00:52:30JdGordon|kugel: you are correct up to "Therefore"
00:52:57gevaertsDoesn't that mean that if you use an sbs on a theme designed for the UI vp, *both* %Vi and the UI vp are wrong? Shouldn't you use the intersection of them for the UI?
00:53:23kugelI considered that yes
00:53:39JdGordon|the intersection is defintly not what the user wants
00:53:53gevaertsJdGordon|: why not?
00:54:10 Quit flydutch ("/* empty */")
00:54:53JdGordon|because it will be either smaller or larger than the user wants?
00:55:34kugelonly %Vi guarantees the optimal maximized area for the WPS. the UI vp is either to small, or draws over sbs
00:55:35gevaertsI see three ways : (a) you use the intersection (b) you try to pick the "less wrong" out of thin air and are likely to get it wrong, or (c) you disallow non-matched theme+sbs
00:55:49JdGordon|kugel: I dont really care if %Vi stays or not (I tinhk its redundant, but fine).... I just care that the logic to determine the default viewport for the lists is different for the WPS... its confusing
00:55:54gevaertsJdGordon|: how can it be larger?
00:55:58kugelI would like to know why sbs changed your opinion about the best area of the WPS
00:57:00JdGordon|gevaerts: sbs specifies a viewport (x,y,w,h) (10,10,50,50), user setting wants (15,15,30,30)... whats the intersection?
00:57:17kugel(d) be deterministic and priotise the one over the other always
00:57:32JdGordon|thats what ive been arguing for
00:57:34gevaertskugel: (d) is (b)
00:57:37JdGordon|for criying out loud
00:57:39gevaertsIt might be smaller, but that's only because otherwise you either draw UI elements over the sbs which is Wrong, or you draw UI elements over bits of backdrop that clash with them
00:57:57kugelJdGordon|: for lists that is, for the WPS there is only one
00:57:58gevaertsJdGordon|: in that case the intersection is (15,15,30,30)
00:58:02kugelnothing to pick from
00:58:35JdGordon|kugel: pre sbs the only reason to use the ui viewport was to make the list look better with a background image.. that is entirely wrong now
00:58:44JdGordon|gevaerts: oh, intersection is always the smaller of the two?
00:58:49kugelit's not
00:58:58*gevaerts gives JdGordon| some coffee
00:59:02kugelthe UI is still only there for that very purpose
00:59:07kugelUI vp
00:59:09gevaertsJdGordon|: it's the overlap between the two
00:59:55kugelthe UI is only for areas which otherwise lack customization. That clearly excludes sbs and wps
01:00:05 Quit GeekShadow (Read error: 104 (Connection reset by peer))
01:00:26 Join GeekShadow [0] (
01:00:41 Join Thundercloud [0] (
01:01:41gevaertsAssume a theme that has a backdrop with a 15 pixel wide black border, otherwise white, with a black font, on a 100x100 screen. It specifies a UI viewport of (15,15,70,70). The sbs with a green font (always visible) sits at (0,70,100,30), so specifies a %Vi of (0,0,100,70). Now enable both. What's the optimal effective viewport for the UI?
01:02:10*gevaerts promises not to upload any theme that does actually look like that
01:02:16 Join roolku [0] (
01:02:29 Quit saratoga (Ping timeout: 180 seconds)
01:02:32kugelgevaerts: they don't overlap?
01:02:55kugeloh, they do
01:03:46gevaertsIf you take the intersection, you get (15,15,70,55), which is the white background that's not taken by the sbs
01:03:50kugelthe next question, what is the optimal viewport for the wps, which would have the borders removed?
01:04:24gevaertswps on its own or wps+sbs?
01:04:37kugelwell, both
01:05:28kugelafter JdGordon it would be the UI vp, I would rather use the %Vi/fullscreen for the no-sbs case
01:06:08JdGordon|ill be back in 10 min
01:06:11 Quit JdGordon| ("Miranda IM! Smaller, Faster, Easier.")
01:06:35kugelthe interesting thing is that wps w/o sbs (i.e. %wd) will get you what you had pre-sbs and what JdGordon actually understand why it's been done
01:07:20 Quit Thundercloud (Remote closed the connection)
01:07:24 Quit ShapeShifter499 ("ChatZilla 0.9.85 [Firefox 3.5.3/20091020102323]")
01:07:25gevaertsin the no-sbs case, I think the wps should use fullscreen. The wps itself is fully viewportified and usually specifies its own background, so I think it should take care of those things itself. Maybe for non-viewport+non-backdrop WPSes UI vp might make some sense
01:07:26kugelusing the %Vi also means fullscreen if there's no sbs displayed (I'm not sure if I have expressed that already somewhere)
01:08:12kugelit only matters for non-vp WPSes actually
01:08:15gevaertsFor the sbs+wps case I think there's no proper way to do it
01:09:42kugelthe UI vp is really only indetended for putting lists and other not-customizable screens into a rectangle. It never applied for screens that can do costumization itself
01:09:48kugelintended too
01:09:51 Join AndyIL [0] (n=pasha_in@
01:10:00gevaertsyou have to use the sbs %Vi, or you get drawing issues (assuming %Vi is specified correctly), and if the wps doesn't specify a backdrop you have to use the UI vp to avoid background/font clashes.
01:10:26kugelwe do use the %Vi currently
01:10:29gevaertsso I think that in the reasonably specialised case of a non-VP wps without a specified backdrop the intersection is correct
01:10:36kugelthat's IMO the optimal choice for the WPS
01:11:19*kugel isn't sure if intersection should be really done
01:11:37gevaertskugel: for general UI or for wps?
01:11:43gevaerts(or both)
01:11:56kugelbut I wouldn't be against it
01:12:32gevaertsI think the example I gave shows that for some cases at least it's the only working solution. Obviously there are edge cases where the intersection would be empty, so there it definitely doesn't work
01:12:51gevaertsI'd consider those to be a case of "don't use this sbs with this theme
01:13:52kugelit's at least imaginable, although highly improbable
01:14:46 Join JdGordon1 [0] (
01:14:56JdGordon1back.. ok.. lemme read the backlog
01:15:25 Quit stripwax ("")
01:15:30gevaertsit requires half the screen to be declared off-limits by at least one of them. I'd only expect that to happen for people who have a broken LCD (in which case the sbs and UI would *really* clash), or for themes/sbses that are really not at all meant to mix and match
01:15:41kugelgevaerts: I can also imagine that intersection is what you would like to see and expect (from a user pov)
01:15:54gevaertskugel: I think so, yes.
01:17:18gevaertsYou have to handle the empty-intersection somehow of course. Either you refuse to load, or you fall back to either the UI vp or %Vi. I think you only need to make the system usable enough for the user to pick another theme or sbs though, not to actually be really usable
01:18:24kugelwe don't protect anywhere against empty viewports.
01:18:30JdGordon1ok... intersection is an unnessecary complication
01:18:44JdGordon1nothing says we have to look after people who put together badly designed themes
01:19:36kugeleach the UI vp with background and the sbs can be designed extra ordinarily well on their own but still cause redraw problems
01:19:44JdGordon1if the wps (for that matter... any skin) always is set up for the full display then there is no problem
01:20:18 Quit DerPapst ("Leaving.")
01:20:21JdGordon1which is why its silly to say that sbs/wps/etc are completly removed from themes
01:20:26kugelfull display? I thought UI viewport
01:20:34gevaertsJdGordon1: for the wps I tend to agree, although I'm not sure. For non-skinned UI, I don't see any other working solution
01:20:38JdGordon1it just doesnt make sense to put two together and hope they work
01:21:06JdGordon1kugel: full screen seems to make more sense to me... otherwise yes the ui viewport which is the same ordering as the lists
01:21:16JdGordon1so setting if exsits, otherwse %Vi
01:21:26kugela full-sized wps default vp (no matter of whether you actually draw stuff into or not) will just always conflict with a sbs
01:21:43 Quit Hillshum (Read error: 110 (Connection timed out))
01:21:47 Quit AndyI (Read error: 110 (Connection timed out))
01:22:15kugelthat's a problem
01:22:28JdGordon1so in that case you would have to add the %we tag to enable the sbs and then you do get the same ui area as the lists... if you dont like that then specify a different viewport
01:22:39JdGordon1seen as the defsault viwport doesnt get drawn in in viewported wps anyway
01:22:53kugelif there's no sbs in the wps then there's no problem
01:23:17JdGordon1if there is no sbs then there is no difference between ui viewport and fullscreen
01:23:28JdGordon1unless the setting is set
01:23:32kugelthe default viewport gets cleared though, several times. that's the main problem
01:23:38JdGordon1in which case.. if the wps doesnt like it, it creates a new one
01:23:46kugelas I said, whether you draw stuff into it or not
01:23:49JdGordon1so we should get rid of this idea of a default viewport
01:24:04JdGordon1or default to the same area as the lists
01:24:14JdGordon1those are your choices
01:24:50kugelbut the UI vp possibly also conflicts. %Vi is the only safe choice (or intersecting)
01:25:33gevaertsI'd go for using the intersection in the UI and in non-VP wpses
01:25:46kugeland the problem here is that you cannot simply fix the wps default viewport without impact on the whole ui
01:25:50gevaerts(if the sbs is shown of course)
01:26:19JdGordon1gevaerts: intersection doesnt make sense because the only reason you'd put the setting there is to overwrite the %vi... i think
01:27:02kugelor loading a sbs from a different theme
01:27:05JdGordon1kugel: easy fix... do the same ALWAYS_HIDDEN thing in all skins instead of just the sbs if any viewports are defined
01:27:33kugelthat will break a lot of themes
01:27:40JdGordon1intersection i think would lso confuse users... ("why is my list only half of what i specified?")
01:28:00gevaertsJdGordon1: no. The UI vp belongs to the backdrop
01:28:07kugelI think it's totally clear to the user when he spots that the sbs is where he had lists previously
01:28:11JdGordon1kugel: no it wont... any which have other viewports defined dont draw in the default one anyway... and big deal if some break
01:28:33JdGordon1gevaerts: it does?!
01:28:40gevaertsJdGordon1: conceptually, yes
01:29:06JdGordon1and wps' which dont have a backdrop? what do they belong to?
01:29:25kugelthe wps is likely to have a different one
01:29:37*gevaerts thinks that wps and non-wps cases definitely need to be looked at separately
01:29:56*JdGordon1 glares at gevaerts
01:30:07gevaertsJdGordon1: you mean wpses that inherit the UI backdrop, or no backdrop at all (i.e. plain colour)?
01:30:11kugelgevaerts: that's what I'm doing the whole time
01:30:17JdGordon1also.. remember that wps here also means radio screen, and other screens as thy get added
01:30:29JdGordon1gevaerts: both... which are i tinhk the same
01:30:36 Join Rand_Althor [0] (
01:31:25 Quit TopyMobile__ (Read error: 110 (Connection timed out))
01:31:28Rand_AlthorI'm thinking of switching from MP3 to OGG. Can anyone guess how much of a difference that would make in battery life? (e200v1)
01:31:30JdGordon1isnt the reason for the broken %pb because the default viewport isnt setup correctly now?
01:31:44gevaertsJdGordon1: the cases are different, because for the general list-based UI bits you can actually make the UI work with any arbitrary rectangle (within reason. I'm not asking for a 1x1 UI). For skinnable screens this isn't true
01:32:08kugelJdGordon1: it's set up to late, or rather the %pb is setup to early
01:32:45JdGordon1then skinable screens shold be given the whole display and at their discretion change to the ui area or something compeltly different
01:32:58kugelI'm not liking the fact that the default viewport size needs to be known at parsing, but I don't bother with that, the fix will probably partly revert some setup there
01:33:04JdGordon1set_defaults() should always return the same viewport... and fullscreen() should always return the full display
01:33:16gevaertsRand_Althor: the best way to find out is to test. Battery life is likely to go down a bit though, the rockbox mp3 decoder is more efficient than the ogg decoder on PortalPlayer (the CPU used in the e200v1)
01:33:19JdGordon1kugel: the default was always known... you changed it recently
01:33:47linuxstbRand_Althor: may have some clues (but lots of information there is out of date...)
01:34:19gevaertsHow many wpses are actually expected to work properly with a custom status bar? rockbox_default will work, but apart from that one?
01:34:45gevaertss/a custom status bar/an arbitrary custom status bar/
01:34:46kugelall that don't use viewports
01:34:50Rand_AlthorI probably will test, I'm just asking because if you told me I'd lose ~25% of my battery time, I wouldn't even bother. Could it be that high?
01:35:06kugelbecause of %Vi
01:35:12 Quit roolku ()
01:35:23JdGordon1gevaerts: thats not really the point though... most people want swap and change
01:36:22JdGordon1kugel: GOT IT! ok.. most wps which dont use viewports (i.e ones we expect to work here).. also dont set backdroips right? which means we would expect them to be in the same area as the list. therefore using Vi here instead of the setting is not what the user expects
01:36:59gevaertsJdGordon1: in that case, the user would expect the intersection between Vi and UI VP
01:37:21kugelJdGordon1: I can't follow that conclusion
01:37:30JdGordon1well maybe.. but intersection is an added complication
01:37:45gevaertswhy? It solves real problems...
01:38:41JdGordon1fine, if its always the intersection I'd be happy with that
01:38:50 Quit liar (Remote closed the connection)
01:39:05JdGordon1I've been arguing (for 3 days now) that the logic to get that area should not be different based on the screen
01:39:55kugelintersecting the lists would actually be harder to do than intersecting always
01:40:01kugellists only I mean
01:40:20JdGordon1I have faith that it could be done! :)
01:40:27JdGordon1so are we finally agreeing?
01:40:31gevaertskugel: don't forget that the status bar isn't always enabled on the wps
01:40:34JdGordon1do we have consensus?
01:40:47gevaertsso I suspect you need both sets of numbers anyway
01:41:00 Quit Rand_Althor ("ChatZilla 0.9.85 [Firefox 3.5.3/20090824101458]")
01:41:09kugeldon't worry, the framework for statusbar toggleging works so far :)
01:41:37JdGordon1gevaerts: no you wouldnt... you would initialise a viewprt with a different funciton which behind the scenes knows everything
01:41:39 Join MethoS- [0] (n=clemens@
01:41:43JdGordon1its either fullscreen or intersection
01:42:22*JdGordon1 is willing to bet that if this happens viewport.c would start being cleaner... and %pb will just work again
01:42:22kugelviewport_set_defaults doesn't initialize necessarily. it rather copies the UI vp
01:42:40gevaertsOf course you can also interpret no sbs as a zero size Vi, which means that the intersection would be the UI VP anyway
01:42:48kugel%pb can be fixed indepentendly
01:42:48JdGordon1thats where the intersect happens
01:43:11JdGordon1rename it to get_ui_Area()
01:43:11JdGordon1ditto if Vi is there and the setting isnt
01:43:30gevaertsyes, but that doesn't change all the time
01:43:45JdGordon1neither will change often
01:43:50*gevaerts keeps out of the where-do-we-code-this discussion
01:43:52JdGordon1and chaning isnt an issue
01:44:08kugelgevaerts: we can detect statusbar changes just fine, that's not a problem
01:44:19JdGordon1kugel: are we both happy with intersection or fullscreen always?
01:44:36kugelwhere/when do you want the fullscreen case?
01:44:46gevaerts"full" wps I guess
01:44:49JdGordon1graphic EQ
01:44:54JdGordon1fullscreen wps
01:45:07JdGordon1any screen which doesnt fit in the smaller one
01:45:18kugeland what is fullscreen then?
01:45:32JdGordon1sbs disabled
01:45:36JdGordon1it is really the entire display
01:45:46gevaertsbasically any screen that doesn't support a variable size at all
01:45:48kugelI rather fix the affected screens to be viewportified
01:46:01JdGordon1untill then!
01:46:06JdGordon1and even then.. they dont always fit
01:46:10kugelbut until we're there we can only do fullscreen anyway
01:46:17JdGordon1i.e quickscreen with a small ui area
01:46:28kugelquickscreen is pretty safe
01:46:38JdGordon1just an example
01:47:01gevaertsI suspect you'll want some plugins to remain fullscreen anyway, especially some of the games
01:47:10JdGordon1right, about to get off the bus... yes or no... are we in agreement?
01:47:15*gevaerts is
01:47:26*kugel didn't expect "what to do with non-resizable screens" to be a question
01:47:34kugelI think we are
01:47:49gevaertskugel: the answer is easy : we non-resize them :)
01:48:00kugelUI vp for the wps, %Vi as fallback, intersection
01:51:03 Quit phanboy4 (Read error: 131 (Connection reset by peer))
01:52:06 Join Thundercloud [0] (
01:53:48***Saving seen data "./dancer.seen"
01:58:19 Quit dmb ("Leaving")
01:58:41 Quit Thundercloud (Remote closed the connection)
01:59:06JdGordonkugel: UI setting always, %Vi as fallback.. if both are there is intersection
01:59:48JdGordonfreeking mazal tov! :p
02:00:03JdGordonwhat else can we argue about? :D
02:00:15*kugel offers JdGordon a handshake
02:09:08 Quit amiconn (Nick collision from services.)
02:09:10 Join amiconn_ [0] (i=quassel@rockbox/developer/amiconn)
02:09:21 Join pixelma_ [0] (i=quassel@rockbox/staff/pixelma)
02:09:22 Quit pixelma (Nick collision from services.)
02:09:30 Nick amiconn_ is now known as amiconn (i=quassel@rockbox/developer/amiconn)
02:09:41 Nick pixelma_ is now known as pixelma (i=quassel@rockbox/staff/pixelma)
02:14:00 Quit JdGordon1 (Read error: 110 (Connection timed out))
02:14:23 Quit dfkt ("-= SysReset 2.53=- Ph'nglui mglw'nafh Cthulhu R'lyeh wgah'nagl fhtagn.")
02:14:39JdGordonkugel: you know the script yeah?
02:15:54kugelit's perl so I would say I know it
02:16:13kugelI think I managed to prepare it for preinstalling .sbs though
02:17:01JdGordondoes it handle putting a theme together which doesnt have the screen dimensions in the actual .wps?
02:17:11JdGordon(thinking ahead to .sbs of course though)
02:17:35kugelI think so
02:17:49JdGordontthe file can be almost identicle for just about all targets so i dont want to have to make one for each differnt lcd size
02:18:13JdGordonexcept the RTC part is going to cause big problems though isnt it?
02:18:46kugelI don't think so
02:19:31kugelIIRC the rtc tags in the wps are working for non-rtc (returning −−:−− and the like), I think there's even a tag for testing the target's RTC capability
02:19:48JdGordonoh right.. ok that could work
02:20:00kugelyou might even drop the rtc then, which would even improve the classic statusbar
02:20:18JdGordonthat wouldnt improve it :)
02:20:23JdGordonthe clock is nice there
02:20:30kugel"−−:−−" also?
02:20:40kugelI mean for non-rtc of course
02:20:46JdGordononly if there is a rtc but its not set
02:20:58JdGordonnon rtc wouldnt show that.. if there is a tag for that
02:21:12kugelin the current status bar too?
02:21:18*kugel isn't sure
02:21:38kugelI think my samsung yh925 showed −−:−− before I enabled rtc in config-yh925.h
02:22:09JdGordonyeah, I dont see a way to know the difference between no rtc and unset rtc
02:22:41kugelJdGordon: don't forget early USB please (which connects to usb before any disk access), or you'll have a talk with amiconn
02:22:53JdGordonI know I know...
02:23:07JdGordonthere is a very boring hard coded one which is different to classic
02:23:48*JdGordon sees if he can change one of the rtc tags to return NULL instead of a string if there is no rtc
02:24:04*JdGordon is an idiot
02:24:11JdGordonscroll up a few lines.. that tag exists :p
02:24:34kugel" I think there's even a tag for testing the target's RTC capability"
02:24:56kugelit's not documented in CustomWPS :/
02:26:40*JdGordon goes to add it
02:26:58JdGordonhow badly is the manual and that wiki page out of sync?
02:27:17JdGordonyou're apparently editing it
02:29:43JdGordon%?cc apparently doesnt work
02:31:58JdGordonwhich single display targets dont have a rtc?
02:36:52kugelno plan
02:38:18JdGordonhaha the ipodg1 sim is fucked :)
02:38:27JdGordon2 scroll events for each up/down press
02:38:50JdGordonthe custom statusbar option is impossibl to get to!
02:39:00kugelit affects all scrollwheel sims
02:39:04 Quit MethoS- (Remote closed the connection)
02:39:14JdGordoni havnt seen that in the e200
02:41:42 Quit kugel (Remote closed the connection)
02:43:43*linuxstb has noticed that in the Nano2G sim...
02:56:14JdGordonI think amiconn mentioned it a few days ago.. but i wasnt paying attention :p
03:06:15 Nick fxb is now known as fxb__ (n=felixbru@
03:08:10 Quit gevaerts (Nick collision from services.)
03:08:22 Join gevaerts [0] (n=fg@rockbox/developer/gevaerts)
03:16:21*linuxstb blames kkurbjun and r22977
03:17:56 Quit adiroiban (Read error: 110 (Connection timed out))
03:18:16*linuxstb also guesses it only applies to scrollwheels where HAVE_WHEEL_ACCELERATION isn't defined
03:21:49CIA-85New commit by kevin (r23392): Fixed the Boomshine plugin not to crash when there are no more levels to play
03:32:24 Quit Topy (Read error: 104 (Connection reset by peer))
03:34:59 Join TopyMobile [0] (
03:35:25 Quit TopyMobile (Client Quit)
03:35:28 Join Topy44 [0] (
03:36:09 Join TopyMobile [0] (
03:53:50***Saving seen data "./dancer.seen"
03:55:34 Quit Strife89 ("Bed.")
03:58:32*JdGordon stumbles on more fairly new #define mess :(
04:02:33 Quit TheSeven (Nick collision from services.)
04:02:53 Join The_Seven [0] (n=theseven@rockbox/developer/TheSeven)
04:03:06 Nick The_Seven is now known as TheSeven (n=theseven@rockbox/developer/TheSeven)
04:07:41JdGordonOH Bloody hell.... the %V line is different for grey/mono/colour targets :/
04:08:05 Quit Rondom (Nick collision from services.)
04:08:16 Join Rondom [0] (n=Rondom@
04:09:12JdGordonhmm... no?
04:09:44JdGordon"time format" setting obviously doesnt exist on non rtc builds
04:11:01 Join dmb [0] (n=Dmb@unaffiliated/dmb)
04:12:04 Join Hillshum [0] (
04:12:46 Quit panni_ ("( :: NoNameScript 3.81 :: )")
04:21:14 Join KBH [0] (
04:21:30 Quit KBH (Read error: 104 (Connection reset by peer))
04:27:16CIA-85New commit by jdgordon (r23393): * code police on a macro which has no reason for being ...
04:27:54 Join BHSPitMonkey [0] (n=stephen@unaffiliated/bhspitmonkey)
04:40:46 Join FOAD_ [0] (n=dok@
04:41:25 Quit JdGordon ("Leaving.")
04:42:01 Quit GeekShadow (Read error: 104 (Connection reset by peer))
04:56:49 Quit FOAD (Read error: 110 (Connection timed out))
04:56:50 Nick FOAD_ is now known as FOAD (n=dok@
04:57:52 Quit TopyMobile (Read error: 113 (No route to host))
05:00:28 Quit TheSeven (Nick collision from services.)
05:00:29 Join TopyMobile [0] (
05:00:31 Join JdGordon [0] (
05:00:48 Join The_Seven [0] (n=theseven@rockbox/developer/TheSeven)
05:01:00 Nick The_Seven is now known as TheSeven (n=theseven@rockbox/developer/TheSeven)
05:18:31CIA-85New commit by kevin (r23394): Boomshine plugin : Better messages at the end of the game (fixes the nblevel+1 being displayed)
05:40:36 Join phanboy4 [0] (
05:48:16 Quit TopyMobile (Read error: 60 (Operation timed out))
05:48:54 Quit Horscht ("Verlassend")
05:50:40 Join TopyMobile [0] (n=topy@
05:50:54 Quit TopyMobile (Remote closed the connection)
05:53:51***Saving seen data "./dancer.seen"
05:58:40 Join TopyMobile [0] (
06:06:15CIA-85New commit by tomers (r23395): Don't break line too early
06:28:33 Quit JdGordon (Read error: 104 (Connection reset by peer))
06:29:12 Join fejfighter [0] (
06:46:05 Join FangZor [0] (
06:48:54 Join JdGordon [0] (n=jonno@rockbox/developer/JdGordon)
06:55:22 Quit Sajber^ ("Leaving.")
07:00:16pixelmathe RTC exist tag works still on my Ondio at least with r22913 but I'll update and check, targets it is usefull for are the Ondios and the H100s an early Ipods I believe, the classic statusbar shows nothing on non-RTC targets
07:03:36pixelmathe tag was documented in the manual
07:03:46 Quit phanboy4 (Read error: 104 (Connection reset by peer))
07:04:07 Join phanboy4 [0] (
07:10:20JdGordonpixelma: yeah, my mistake.. the %?St tag doesnt work on non rtc targets because that time format setting doesnt exist
07:10:30JdGordonI'm using that to choose between 12 and 24 hour mode...
07:10:50 Part FangZor
07:15:21 Quit BHSPitMonkey (Remote closed the connection)
07:15:35*JdGordon just loves it when the WPS doesnt display at all, but it is in the screen!
07:15:54pixelmayou'll probably will also have a small problem with the virtual LED tag as it is only enabled on targets that have no real LED (or at least that's how it was)
07:17:53JdGordongrr :/
07:18:07JdGordoncan we do a features.txt thing for skins also :)
07:18:44JdGordonactually, something like the lang file so each theme has a default and a different file can be chosen per target
07:18:50pixelmaI never understood that *completely* (well yes it is called "virtual" LED and should be a replacement for the real LED) but if the classic statusbar showed disk activity there I would be for enabling that tag on all targets
07:20:30pixelmamake chose something like "cabbiev2.160x128x2.rtc.wps"? ... would be nice
07:22:53JdGordonsomething like e200*: cabbiev2.160x128x2.rtc.wps
07:23:13JdGordonsame idea as the lang file where the target chooses which string to use
07:23:33pixelmaok, that's slightly different
07:24:27JdGordonmore than slightly :)
07:24:33pixelmathen you have to go through all the targets and have to check if it'll work or not and get a long (or short) list
07:24:33JdGordonbut it gives us more control
07:26:05pixelmaI guess the control will only be interesting for certain features but display resolution is still a fact
07:28:27JdGordondisplay res doesnt have to be a fact at all...
07:28:53JdGordonthe classic_statusbar works fine on any display width above a certain size
07:29:09JdGordonyes its a crap one, but its one that does work no matter how wide the display is
07:30:45 Join BHSPitMonkey [0] (n=stephen@unaffiliated/bhspitmonkey)
07:32:19pixelmabut you won't make it work this variably if you keep it bitmapped with the current system (at least if the disk activity stays at the right corner)
07:36:59pixelmawell you would need to give the position for the virtual LED bitmap and this varies from display size to display size as it is right aligned so to speak
07:39:42JdGordonmy recent viewport changes fixes that... you can use negative numbers for the 4 pixel position values to mean "other edge minus BLAA"
07:39:54JdGordoni.e x=-50 is 50pixels from the right edge
07:41:04pixelmaah hmm
07:41:28JdGordonwhich is useful for more than just making a skin work on multiple displays :)
07:44:13pixelmaand in this example the viewport to the left of it will know about it if you don't give the width of it (I mean leave the width as - )
07:45:44pixelmasorry for weird way of putting this, I hope it's understood
07:47:23JdGordonthat would only work if the draw order is left to right (viewports)... but you could use -50 for the width of the viewport on the left so it guarentees to not overlap :)
07:50:43JdGordoneek... circular dependancies with the statusbar and wps and viewport initialising :(
07:53:53***Saving seen data "./dancer.seen"
07:57:32 Join Grahack [0] (
08:04:50 Quit bluebrother (Nick collision from services.)
08:04:52 Join bluebroth3r [0] (n=dom@rockbox/developer/bluebrother)
08:06:42pixelmaso, the themer has to take care that the width still is correct himself (or herself)...
08:07:06 Join Rob2223 [0] (
08:07:42JdGordonwell, the would anyway
08:07:57JdGordonwierd shit happens when there is no Vi and ui viewport setting :)
08:09:10 Join matsl [0] (n=matsl@
08:09:44 Join Zagor [242] (n=bjorn@rockbox/developer/Zagor)
08:21:07 Join ender` [0] (
08:24:57 Quit Rob2222 (Read error: 110 (Connection timed out))
08:27:44JdGordonwhats the %mv tag?
08:29:05amiconnJdGordon: That bug doesn't affect all scrollwheel sims. Out of the sims I am building, it only affects iPod G1/G2, but I didn't analyze the bug yet
08:30:19amiconnAnd the built-in status bar should be identical to the current one
08:31:29JdGordonit will be close... not identicle
08:31:41JdGordonthere is a .sbs file which will be identicle
08:33:05amiconnHmm, and could you explain what negative width/height actually adds?
08:33:12CIA-85New commit by jdgordon (r23396): minor changes... ...
08:33:28 Join esperegu [0] (n=quassel@
08:33:37pixelmaJdGordon: wasn't it you who implemented the %mv tag? IIRC it is for displaying something else while changing volume...
08:33:42amiconnIiuc you can always define the same viewport using positive width and height
08:33:44JdGordonbecause the bitmaps make it impossible, without adding extra tags to handle them correclty (whereas they could be completly removed :) )
08:34:02JdGordonpixelma: yeah, checked the wiki.. im half asleep here
08:34:35JdGordonamiconn: of course you can... but if you are lazy (which I am very muchly) you dont have to worry about lcd size.. ou can just give a negative value and it will do the magic for you
08:34:45JdGordonand it makes sharing skins between targets simpler
08:35:00amiconnFor x and y this is obvious, but not for width and height
08:35:12JdGordonits less obvious.. but it does work
08:35:30JdGordonnot when you use negative for both x and width though :p thats just silly
08:35:33amiconnI'm not asking whether it works, it's just unnecessary imo
08:37:10 Join LinusN [0] (n=linus@rockbox/developer/LinusN)
08:37:55amiconnNegative x and y allow you to define viewports which keep their positition relative to the right and bottom edge, which is perfectly understandable. But what does negative width and height actually add?
08:38:02 Quit fyrestorm ("lamers envy me like they envy bill g -- main boot xp, just the way it should be!")
08:38:29amiconnYou can always define the same viewport using positive width and/or height and negative x and y instead
08:38:41amiconnUnless I'm missing something, that is
08:39:58JdGordonyou are right.. its just another way to do it which some people might prefer... one use case I just thought of though is centering text inside this viewport, using a negative number moves the center depending on the lcd width
08:40:27JdGordonother than that, yeah there is no real need for it... It was a not thought out rushed commit
08:40:34JdGordonit seemed to make sense at the time
08:42:01pixelmaamiconn: it adds right or bottom aligned viewports independent of actual display size
08:42:34pixelmalike in the statusbar example
08:42:50JdGordonna, he is right :p only the negative x/y values are actually useful
08:43:00JdGordonwidth/height doesnt add anything
08:43:40pixelmaand how would you have the left viewport in the statusbar example have variable size then?
08:44:17amiconnpixelma: Negative x and y already adds that possibility
08:44:45amiconnI'm not questioning that, I'm only questioning negative width and height
08:45:25pixelmawell, he told me that you would need -50 for the width instead of just "skipping" the parameter with - for the left one
08:46:17amiconnHmm, then how does that actually work?
08:46:43pixelmaso I assumed it was needed to get the viewport size as - usually means "from here to the end of the display"
08:47:02amiconnMy (intuitive) assumption is that negative width just means the viewport extends to the left instead to the right, so that the x coordinate defines the right edge of the viewport
08:47:13JdGordonpixelma: well if you know you want -100 for x, and -50 for width, you could just use 50 for the width to get the same thing :)
08:47:40amiconnJdGordon: Then you'd need a different x though, correct?
08:47:46pixelmaJdGordon: that isn't variable then though... which was my whole point
08:48:10JdGordonI hate you guys :) you're confusing me
08:48:14JdGordonI agree with you both!
08:48:21JdGordonwhich means its way past my bed time!
08:48:23JdGordongood night
08:48:27CIA-85New commit by kevin (r23397): Boomshine plugin : port to Greyscale and B/W targets, and targets without touchscreen (currently using the virtual keyboard keymap)
08:48:43*amiconn thinks that this brings up a valid point though - the current system doesn't allow to define a viewport width which depends on the display width
08:51:25 Join flydutch [0] (
08:52:17pixelmayou could get the same effect if you make " - " clever enough to figure out itself that there is a viewport to the right that could get in the way
08:52:52pixelmaor below
08:55:46JdGordoncrap.. just had a idea for skins and dont want to forget it.... have a seperate settings file for skins so instead of having to comment out %Vd lines like that commit 10min ago allowed, you could just check the value of some tag which loads the one settings file shared between all skins, themers would be free to add as much configurability as they liked to skins then!
09:05:25amiconnHmm. Did anyone try lua on SH1 yet?
09:05:56amiconnIs that even built? If not, these lua plugins should be excluded for small plugin buffer targets
09:06:25pixelmaI think lua isn't build on lowmem (but am not sure at the moment)
09:06:42 Join petur [50] (n=petur@rockbox/developer/petur)
09:07:33 Quit Grahack (Client Quit)
09:10:24 Quit BHSPitMonkey ("Ex-Chat")
09:14:37pixelmaJdGordon (for later) - why would the setting if you want graphical or text battery and volume indicators be gone? I guess you could achieve the same effect with the %St tag if you keep the setting, no?
09:14:52pixelmasame effect as in the current status bar
09:15:55pixelmait lets people still chose the way they want to display it without an own .sbs file
09:16:33pixelmaor even with a custom sbs
09:27:09Unhelpfulamiconn: you may have seen, but kugel and i have been trying to build with an eabi toolchain again... we've got things to the point where rockbox builds, but we have three boot failures for three targets. i don't recall precisely where he said his were failing, but my e200 just hangs at the splashscreen... any ideas on how to proceed? look at asm for early boot, i would guess?
09:28:03 Join Thundercloud [0] (
09:32:19amiconnUnhelpful: Hmm. Checking the asm is a good idea. Perhaps you can also use an easy to control indicator (e.g the button lights on a c200 - they're gpio controlled), and move that forward through the boot sequence step by step
09:34:20amiconnThese long-call veneers - are they always put into the correct section?
09:34:46amiconnAlso, is eabi parameter passing and function return different from the old arm abi?
09:34:55amiconnIf so, all asm functions need changing
09:36:18Unhelpfuli don't have a c200... is there a fairly easy indicator available on the e200? from what i could see, the *syscall* parameter passing convention changes, to align 64-bit parameters to even registers. i didn't see anything stating that that extends beyond syscalls...
09:40:10CIA-85New commit by kevin (r23398): Boomshine plugin : make speed and ball sizes adapt to the target screen size to get a constant gameplay
09:52:00Unhelpfulhm! well, apparently it does. a function with signature (int a, long long int b) expect a in r0 and b in (r2,r3)
09:52:30 Join liar [0] (n=liar@
09:53:56***Saving seen data "./dancer.seen"
09:57:27 Quit Thundercloud (Remote closed the connection)
09:58:35 Join cas [0] (
10:07:31 Join adiroiban [0] (
10:19:10 Join FOAD_ [0] (
10:19:45 Quit HellDragon (Read error: 60 (Operation timed out))
10:22:54 Quit phanboy4 (Read error: 104 (Connection reset by peer))
10:27:42 Quit FOAD (Read error: 145 (Connection timed out))
10:27:42 Nick FOAD_ is now known as FOAD (
10:28:49 Quit TheSeven (Nick collision from services.)
10:29:11 Join The_Seven [0] (n=theseven@rockbox/developer/TheSeven)
10:29:20 Nick The_Seven is now known as TheSeven (n=theseven@rockbox/developer/TheSeven)
10:43:04 Join HellDragon [0] (
10:48:06 Join kugel [0] (n=kugel@rockbox/developer/kugel)
10:50:02kugelJdGordon keeps adding unrelated changes to his patch :/
10:51:35Unhelpfulkugel: hey! the even-register-aligned doubleword parameter for eabi *do* apply to regular function calls... any idea how best to find any asm functions that take a doubleword?
10:52:34kugelI knew that, but I don't think we have asm functions taking long long
10:52:53kugelmaybe in codecs, but that shouldn't prevent booting
10:53:32kugelbut other than looking at each *.S for for target, no
10:53:48Unhelpfulare you entirely certain? isn't our size_t long long?
10:54:22kugelbtw, the sansa fuze failure is somewhere near mounting filesystems (the sd driver panics), I assume the e200 one is before that
10:55:13Unhelpfuli only see the bootloader splash screen on e200.
10:56:39kugelsame here
10:57:18Unhelpfulthere could also be struct packing issues, if any of the early stuff relies on struct portability between c and asm...
10:57:52kugelyou could try adding a lcd_clear_display() after lcd_init, for before you could use buttonlight as indicator
11:09:26jasioNew lua script boomshine causes a PANIC on vx777. most likely a problem on the player than with lua
11:15:52 Join b0hoon [0] (
11:20:50 Quit b0hoon ("CGI:IRC (Ping timeout)")
11:29:07 Join DerPapst [0] (
11:31:14 Join mr45 [0] (
11:36:20 Part mr45
11:38:59 Quit kugel (Read error: 110 (Connection timed out))
11:54:00***Saving seen data "./dancer.seen"
12:04:54 Quit GodEater (Read error: 113 (No route to host))
12:15:15CIA-85New commit by roolku (r23399): Fix FS #10734 - scoring in boomshine.lua is broken (r23391)
12:16:36 Join roolku [0] (
12:23:12 Join kugel [0] (n=kugel@rockbox/developer/kugel)
12:32:05 Join rasher_ [0] (
12:36:13 Quit rasher_ (Client Quit)
12:37:42 Join mt [0] (n=mtee@rockbox/developer/mt)
12:39:35 Quit roolku ()
12:39:57 Join gingermig [0] (
12:42:15 Quit liar (Read error: 148 (No route to host))
12:43:12gingermigHi devs and others! I have googled the following idea, but found nothing special. Here goes:
12:43:26gingermigThe X5, among others, has USB OTG capability.
12:43:35gingermigWould it, at some point in the future,
12:43:51 Join mcuelenaere [0] (
12:43:58gingermigbe possible to use a USB microphone connected to the OTG port,
12:44:23mcuelenaerejasio: what panic are you getting?
12:44:32gingermigand thereby use the media player and Rockbox to act as a high-quality mobile recorder?
12:44:45 Join MethoS- [0] (n=clemens@
12:45:07jasiomcuelenaere: some characters are printed out of the screen but I get this:
12:45:13jasioException occurred: Reserved Instructio [0x080100A1] at 0x80F37890 (stack at 0
12:45:21jasioException occurred: Reserved Instructio [0x080100A1] at 0x80F37890 (stack at 0
12:45:51mcuelenaereodd, what SVN revision are you using and how do you launch the game?
12:46:30mcuelenaerewow the balls are much smaller now with those recent commits
12:47:14jasior23391 I think
12:47:37jasior23395 actually
12:47:48mcuelenaerehmm I used r23399 and it works perfectly (VX747)
12:48:15mcuelenaerejasio: can you run any other Lua scripts?
12:48:49jasioi don't have any other one. any one you want me to try?
12:48:50mcuelenaerejasio: also, do you experience a noticeable delay when connecting your DAP to your computer between physically connecting it and the OS mounting the disk? (~10-20 secs)
12:48:58*mcuelenaere looks
12:49:48mcuelenaerejasio: try this one
12:50:48jasiomounting on WinXP at work is quick.
12:51:20 Quit kugel (Read error: 60 (Operation timed out))
12:52:10mcuelenaerehmm ok, I get bus resets when mounting it here on Linux (which didn't appear in earlier SVN revisions)
12:52:50 Join GeekShadow [0] (n=Antoine@reactos/tester/GeekShadow)
12:53:34jasiothat lua script works.
12:53:58mcuelenaerejasio: what happens if you go to /.rockbox/rocks/games/boomshine.lua and run the script from there?
12:54:05mcuelenaere(in the file browser)
12:54:46jasioi get the "level 1" text and a second or two later the PANIC
12:55:21mcuelenaereah so it's dependant on some actions runned by the script
12:58:03mcuelenaerehuh? VX777 simulator doesn't build
12:58:28 Nick fxb__ is now known as fxb (n=felixbru@
12:59:54jasioi have the sim running but not updated it for while
12:59:54mcuelenaerelooks like it needs a rockbox-ondavx777.bmp
13:00:36jasioI might have copied that pic from vx747
13:00:40mcuelenaerebah, grid mode is still the default?
13:00:42*mcuelenaere did that too
13:00:53mcuelenaereit needs a picture of the VX777 though
13:02:28jasioit doesn't look significantly different from 747 so it was enough for my first try
13:02:28mcuelenaeregreat, no crash in the sim
13:02:28*mcuelenaere adds a :/ to that
13:03:56jasiogot to go to meeting. brb
13:04:39 Join KBH [0] (
13:06:10 Join fyrestorm [0] (
13:06:43CIA-85New commit by mcuelenaere (r23400): Remove svn:executable from simulator UI
13:07:15*mcuelenaere got to go too
13:07:17 Quit mcuelenaere ()
13:08:43 Quit HBK (Read error: 60 (Operation timed out))
13:13:06 Join GeekShado_ [0] (
13:14:14 Quit antil33t (Read error: 54 (Connection reset by peer))
13:14:20 Join antil33t [0] (n=Mudkips@
13:15:23 Quit GeekShadow (Read error: 104 (Connection reset by peer))
13:16:46 Join kyle6513 [0] (n=kyle6513@
13:20:04 Quit freqmod_qu (Remote closed the connection)
13:22:44 Quit DerPapst ("Leaving.")
13:23:28 Quit gingermig ()
13:24:02 Quit mt (Read error: 110 (Connection timed out))
13:33:35CIA-85New commit by mc2739 (r23401): Remove svn:executable from wps
13:41:37 Quit kyle6513 ("Leaving")
13:47:59 Join funman [0] (n=fun@rockbox/developer/funman)
13:49:40 Join dfkt [0] (i=dfkt@unaffiliated/dfkt)
13:54:02***Saving seen data "./dancer.seen"
13:54:23 Join kugel [0] (n=kugel@rockbox/developer/kugel)
13:58:58*gevaerts grumbles
14:03:41 Join Horscht [0] (n=Horscht2@xbmc/user/horscht)
14:11:21 Join teru [0] (
14:13:11 Quit kkurbjun (Read error: 110 (Connection timed out))
14:14:09 Quit fyrestorm ("lamers envy me like they envy bill g -- main boot xp, just the way it should be!")
14:46:53 Quit funman ("free(random());")
15:00:46 Join evilnick_B [0] (i=0c140464@rockbox/staff/evilnick)
15:12:43 Quit GeekShado_ (Connection timed out)
15:20:37 Quit MethoS- (Remote closed the connection)
15:31:28 Join toffe82 [0] (n=chatzill@
15:34:15 Join n1s [0] (n=n1s@rockbox/developer/n1s)
15:37:20 Quit linuxguy3 (Read error: 110 (Connection timed out))
15:41:48 Quit fejfighter ("Leaving")
15:44:30 Join Sajber^ [0] (
15:47:00 Join liar [0] (n=liar@
15:47:11 Quit liar (Client Quit)
15:47:35 Join liar [0] (n=liar@
15:52:42 Quit liar (Client Quit)
15:53:09 Join liar [0] (n=liar@
15:54:05***Saving seen data "./dancer.seen"
15:56:32 Join sirdaniel [0] (
15:57:54 Quit sirdaniel (Client Quit)
15:59:59 Join funman [0] (n=fun@rockbox/developer/funman)
16:00:42 Part LinusN
16:03:24 Quit teru ("Quit")
16:04:22 Join Strife89 [0] (n=michael@
16:09:11CIA-85New commit by nls (r23402): Const correctness
16:10:30 Join Grahack [0] (
16:11:49 Join jhulst [0] (
16:13:28 Quit liar (Client Quit)
16:14:20 Nick YPSY is now known as Ypsy (n=ypsy@
16:22:50 Quit Grahack ("Leaving.")
16:24:09 Join freqmod [0] (
16:24:47 Join phanboy4 [0] (
16:26:04 Quit freqmod (Read error: 54 (Connection reset by peer))
16:29:42gevaertsany theme site admin around? PHI-infoScreen for 320x240 contains non-theme settings
16:36:51 Join freqmod_qu [0] (n=quassel@
16:38:55 Join bmbl [0] (n=Miranda@unaffiliated/bmbl)
16:47:36 Quit rasher ("moving day")
16:49:56mc2739Thingie200 for 176x220 also has non-theme settings
16:50:54 Quit cas ("ChatZilla 0.9.85 [Firefox 3.0.14/2009090216]")
16:52:08gevaertsmc2739: which?
16:55:24pixelmaand there is still the problem of some themes showing for wrong resolutions
16:56:15 Join webguest65 [0] (
16:56:57 Join liar|netbook [0] (n=liar@
16:56:58 Quit webguest65 (Client Quit)
16:57:26mc2739gevaerts: which settings?
16:57:50gevaertsmc2739: yes. I've looked at it, and I don't see anything special.
16:58:11mc2739it has fmpresets - is taht theme related?
16:58:59gevaertsoh indeed. I'd missed that :\ I only looked at the cfg
16:59:01 Join rasher [0] (
16:59:31HillshumIt would be nice if rbutil could download and install fmpresets for users...
16:59:40 Quit matsl (Read error: 145 (Connection timed out))
16:59:59mc2739there is also a .kbd file and a .colours file that colud be considered theme related, but are not in the .cfg
17:00:26gevaertsI don't think a .kbd belongs in a theme
17:00:30 Quit bubsy (Remote closed the connection)
17:02:22 Quit funman ("free(random());")
17:03:51kugelHillshum: that would be fine indeed
17:04:14kugelgevaerts: kbd is an edge case IMO
17:04:27HillshumAlong with other regional settings (DST?)
17:04:53gevaertskugel: why? How is that related to theming?
17:05:15kugelit changes the look of a part in rockbox
17:05:51gevaertswith that argument themes can ship cover.jpg files
17:06:33kugela fallback one in .rockbox/albumart, indeed. nice idea
17:07:38 Quit kugel (Remote closed the connection)
17:07:59 Join bubsy [0] (n=bubsy@
17:13:38 Quit Strife89 ("Lunch.")
17:17:41 Join kkurbjun [0] (
17:23:16 Join funman [0] (n=fun@rockbox/developer/funman)
17:30:41 Nick liar|netbook is now known as liar (n=liar@
17:31:35 Quit Zagor ("Don't panic")
17:36:59 Quit antil33t (Read error: 104 (Connection reset by peer))
17:37:08 Join antil33t [0] (n=Mudkips@
17:40:54rashergevaerts: you certainly dug your own grave there
17:41:54gevaertsrasher: maybe, but I still don't see why a file that configures which characters you can type belongs in a theme :)
17:42:35rasherI agree with that.
17:48:28 Join GeekShadow [0] (n=Antoine@reactos/tester/GeekShadow)
17:49:55 Join DerPapst [0] (
17:52:10CIA-85New commit by mcuelenaere (r23403): Also expose functions having (s)size_t or off_t as argument/return type to Lua.
17:52:13CIA-85New commit by mcuelenaere (r23404): Expose pcmbuf_beep() to plugins (based on FS #10608 by Christophe Gragnic)
17:54:09***Saving seen data "./dancer.seen"
17:56:24 Join linuxguy3 [0] (n=timj@
18:01:23 Quit J-23 ("ZNC -")
18:04:20 Quit petur ("work->home")
18:13:37CIA-85New commit by mcuelenaere (r23405): Lua: return nil when function failed instead of returning nothing
18:15:47 Quit freqmod_qu (Remote closed the connection)
18:16:32CIA-85New commit by mcuelenaere (r23406): Only expose pcmbuf_beep on SWCODEC targets (aka fix red).
18:21:42 Join bertrik [0] (
18:24:50 Join domonoky [0] (n=Domonoky@rockbox/developer/domonoky)
18:31:22 Join kugel [0] (n=kugel@rockbox/developer/kugel)
18:35:15JdGordonpixelma: well, those settings would be unused, so they shold be removed... which is why I blurted out that idea a few min before your message
18:36:26JdGordonkugel: everything is releated to everything :D
18:42:58pixelmawhy would they be unused?
18:43:34 Quit DerPapst ("Leaving.")
18:44:44JdGordonthey arnt used elsewhere...
18:45:00pixelmabut you can still use it in an sbs
18:45:19JdGordonwhich is why I suggestesd my suggestion
18:45:31JdGordonbut leaving them in the core is a waste, its just clutter
18:45:49pixelmaI mean code your sbs in a way the user has a choice
18:46:06pixelmawithout having to make his own sbs
18:46:21Hillshuma = class() should make a an instance of class class correct?
18:46:37HillshumWait a sec. Wrong channel
18:47:44pixelmaJdGordon: I don't see why those settings need to be removed, they could serve the same purpose as they do now
18:48:38kugelJdGordon: it is totally unrelated
18:49:00kugelread the task's title and explain how modifying a option_screen() call is related
18:49:31JdGordonthey take up useless space in the menus, and binary... my suggestion was to have a seperate file in with the wps's which has a listing of all the different config options different skins might have... so a theme could add any customisability they wanted. The suer would just need to change the skins config file (one file shared by all)
18:49:56JdGordonkugel: ok, then I'll go and remove that line in svn then
18:50:08kugelkeen on introducing bugs?
18:50:09JdGordonalthough I thought you meant the rest of the viewport.c changes
18:50:35JdGordonas much as I'd love to continue this.. i have to get to work
18:51:31kugelchange some option in pictureflow in a build with that change if you want to know what happens
18:51:56 Join Rand_Althor [0] (
18:51:59mc2739JdGordon: there seems to be a problem with the classic_statusbar bmp files - they load in gimp, but not Rockbox or windows
18:52:38Rand_AlthorAny chance that RB will one day compile in Visual Studio?
18:52:49kugelI'm sure you changed much more unrelated code, I haven't looked it in detail because it's mixed with #if 0'd which is confusing
18:54:25AlexPRand_Althor: Does visual studio let you compile for other archs? e.g. SRM, Coldfire, sh, mips?
18:54:57 Nick Ypsy is now known as YPSY (n=ypsy@
18:55:06Rand_AlthorI don't know. I'm thinking about buying it.
18:55:19AlexPI very much doubt it does
18:55:53AlexPPersonally I would use the free open source tools that exist and are know to do the job
18:56:17JdGordonRand_Althor: wait for VS2010 :)
18:56:25JdGordonbut no, rockbox likely wont ever work with it
18:56:38 Join freqmod_qu [0] (
18:56:47JdGordonmc2739: I may have saved them in the wrong format.. although they work in the sim
18:57:05JdGordonor oyu got a corrupted file?>
18:57:06 Join Grahack [0] (n=chri@
18:57:37pixelmaJdGordon: having options without them being settable through the menu isn't nice and I thought we agreed that this isn't wante
18:58:27JdGordonthis is different
18:58:33JdGordonback in 30
18:58:44pixelmaI don't think so
18:59:32 Join andrewRB [0] (
18:59:46 Quit Rand_Althor ("ChatZilla 0.9.85 [Firefox 3.5.3/20090824101458]")
19:00:52 Nick Hillshum is now known as myself (
19:01:15 Nick myself is now known as Hillshum (
19:02:39mc2739JdGordon: sorry, maybe I did have a bad copy, they seem to be ok now
19:13:37Bob_CRand_Author:if by "compile from" i.e. edit in VS then launch build tools then yes. I worked with a setup like that once
19:21:42AlexPedit source seems quite a way from compile to me
19:29:12 Join _zic [0] (
19:31:16 Quit funman ("free(random());")
19:31:18 Quit phanboy4 ("Leaving")
19:33:42 Join JdGordon| [0] (n=Miranda@nat/microsoft/x-qrhtgnynuoepaplm)
19:37:35 Quit Grahack (Client Quit)
19:37:35JdGordon|pixelma: yes it was agreed that settings shouldnt be available if they arnt in the menus, but what I'm thinking would allow any skin to create their own settings, named whatever they want, meaning anything they want.. which is imbossible to put in the menus
19:38:30HillshumThey could make their own?
19:39:08linuxstbYou mean a skin author can create a setting as part of the skin?
19:39:45 Quit _zic ("Ухожу")
19:40:23JdGordon|I dont know how useful it actually is though
19:42:23 Quit shai ("Leaving")
19:43:26evilnick_BIs this (from the D2 Wiki) still accurate? "NOTE: Newer D2+ hardware revisions use an updated power management chip, and some functionality is not yet implemented on these players (eg. touchscreen support)."
19:44:17Hillshumevilnick_B: The first part is accurate
19:45:04 Join J-23 [0] (
19:45:14evilnick_BSo the touchscreen will work on ALL D2+s?
19:45:15 Nick J-23 is now known as Guest32413 (
19:45:41 Nick Guest32413 is now known as J-23 (
19:45:58pixelmaJdGordon: I can't even imagine how that would work - code wise as well as for the user
19:46:21pixelmaevilnick_B: I think the forum has some more details about it
19:46:35JdGordon|code wise you mean in the skin?
19:47:08JdGordon|%?Ss|my setting name|<option 1|option 2|option 3|....>
19:47:22pixelmain the skin as well as "behind the scenes"
19:47:22JdGordon|in the settings file it would be "my setting name: 2"
19:47:35evilnick_Bpixelma: Thanking you
19:47:45JdGordon|dont you worry about blank... let me worry about blank :D
19:48:14pixelmaI can't imagine currently that it will not confuse users
19:48:19JdGordon|I agree its not nice, and the only case I can tihnk of that would use it is the icon/text stuff for the classic bar
19:49:39pixelmaso remove an easy setting and replace it with something possibly more complicated and confusing? (putting it a little harsher but)
19:50:41JdGordon|I dont really mind leaving them in.. but the people that scream blue bloody murder when a new patch adds a couple hundred bytes bin size, but then want this to stay....
19:51:54JdGordon|also, at least half the targets wont use these ever because the default cabbie theme doesnt display the bar
19:54:13***Saving seen data "./dancer.seen"
19:55:02JdGordon|which part of that was the ? directed at?
19:55:55pixelmawhy you think the default cabbiev2 doesn't display "the bar" (which bar)
19:56:31*JdGordon| is going to get in trouble for rmemebering wrong isnt me?
19:57:08 Join funman [0] (n=fun@rockbox/developer/funman)
19:57:16JdGordon|the bar being the classic statusbar...
19:57:29pixelmaif I remember correctly it was agreed on that cabbiev2 won't touch the user setting there, and the default is on I believe
19:57:52JdGordon|oh? bah
19:58:11 Quit Hillshum (Read error: 104 (Connection reset by peer))
20:00:28pixelmaalso, as I said - you can still offer an sbs that shows either a graphic or the number
20:00:29kugelUnhelpful: it freezes in settings_list.c the first time
20:01:47pixelmaJdGordon: which reminds me... I still don't understand why you stopped implementing %mv into cabbiev2 after converting the Clip's wps... ;)
20:02:20JdGordon|because I'm a lazy bugger and only did it on the clip because thats the one I happened to be using at the time and wanted it
20:03:14kugelwhy do you want to remove volume/battery display?
20:03:32JdGordon|oh for crying out loud! I DONT! I'M NOT!
20:03:53pixelmahe only wants to remove the setting to chose between graphical and numerical display
20:04:00kugelI mean that
20:05:08 Join domonoky1 [0] (
20:05:27 Join einhirn [0] (
20:06:44pixelmaI don't understand that either, at least it's unrelated to removing the hard-coded statusbar
20:06:50n1sthat setting is a bit out of place when you have proper skins imo
20:07:56pixelmaI don't think that it has to do with "proper skins" - you can still offer the user both ways of displaying in your "proper" skin, designed how you want it
20:10:00pixelmaI find most - if not all - pure graphical battery displays I've seen so far too coarse (classical statusbar is an edge case with something like 15 steps or so and I still use numerical there)
20:10:09kugeln1s: I don't think so. In contrary, proper skins can use %St to react nicely on user settings
20:10:11pixelmajust as an example
20:10:12n1spixelma: yes, but why should that one detail have a setting while other things don't
20:10:24 Quit flydutch ("/* empty */")
20:10:39pixelman1s: can you give an example?
20:10:52n1skugel: my point is not that it's useless or won't work it's that it's arbitrary and out of place
20:10:53kugela sbs version of the current statusbar can implement the current way of showing volume and battery (depending on the setting) using %St
20:12:03kugelmy point is that it's removing an existing "feature". I mean I don't really care about this two settings anyway
20:12:13kugelI think I never even changed those
20:13:44n1spixelma: well theoretically most things in a skin could be modifiable with settings
20:14:28pixelmacan't imagine one currently which doesn't have a setting already
20:14:46n1sI don't care that much about it and i don't agree with the argument to keep existing features, if we can add stuff we should be able to remove stuff
20:16:50n1swe could add enable/disable for stuff like peakmeter and aa
20:19:55*pixelma remembers hearing a story about the removal and readding of the pointer selector... ;)
20:20:38*Bagder likes a good story ;-)
20:21:38 Quit domonoky (Read error: 110 (Connection timed out))
20:23:09JdGordon|It doesnt make sense that we bloat the menus with settings that most people wont use
20:23:39kugelUnhelpful: in lcd_init() too
20:24:18pixelmaJdGordon: where do you get that knowledge from how many people use it?
20:25:22JdGordon| :)
20:25:41JdGordon|rasher: can you run a script to see how many themes on the site disable the statusbar?
20:26:06rasherNot.. readily
20:26:16JdGordon|and even those that done disbale it, I bet you only a very tiny percentage would change those two settings
20:26:40pixelmaI don't think it tells much as it is currently a setting that is easy to change by the user without "destroying" the theme
20:27:53pixelmaand a setting I would not put in a theme so that the user can keep his settings
20:28:50kugelJdGordon|: you aren't serious are you?
20:28:53pixelmamine would set the battery display to numerical then
20:29:16kugelYou are editing the task title just to make random unrelated changes?
20:29:16JdGordon|kugel: right this very moment? or some time in the future? or past?
20:29:41 Quit goffa (Remote closed the connection)
20:31:10*kugel wonders when we established an exception in the rule of 1 change per patch.
20:31:44*JdGordon| wonders where kugel got the impression that that task was supposed to be commit ready?
20:32:25kugelyou're putting words in my mouth
20:32:30 Join GodEater [0] (
20:32:31JdGordon|no im not
20:32:51kugelI've not said/implied that I think it's commit ready or any of that sort
20:32:57JdGordon|the only reason you would be complaining is because you think it could be ready for commit!
20:33:09JdGordon|or you think I tinhk its ready
20:33:21JdGordon|or I tihnk you think i think they think its ready
20:33:45kugelI'm complaining because you make it difficult for me to look at/review the patch
20:33:48JdGordon|or you tihnk i tihnk you tihnk they think i think you think donkeys think
20:35:15 Join goffa [0] (n=goffa@
20:35:35kugeland because you're sneaking in random changes unrelated to the actual purpose of the patch, but I assume you just magically avoid that by changing the task title
20:36:04JdGordon|I changed the name because of what you said in irc 90min ago
20:36:20JdGordon|the only reason its there is because I'm not a git
20:36:30JdGordon|user and tend to lose patches
20:37:15GodEaterwhat an awesome place to accidentally hit enter ;)
20:37:58pixelmaJdGordon: you could make two patches out of it
20:38:27kugelif that means that you're going to extend the task in order to upload several, nicely separated patches, that might be fine
20:40:58 Join froggyman [0] (
20:41:27kugelJdGordon|: what other benefits does knowing the default vp dimensions at parsing time have?
20:42:47kugelUnhelpful: it also freezes in create_thread() :S
20:43:50CIA-85New commit by bluebrother (r23407): Improve a few debug logger messages.
20:44:13 Join robin0800 [0] (
20:45:00kugeldoes anyone know where our size_t is defined or are we using the host system's definion (I tried greppin already)
20:47:13gevaertskugel: include/sys/types.h
20:47:23*gevaerts also used grep :)
20:47:39gevaertsfirmware/... of course
20:47:55kugelI didn't know that file exists! Thanks
20:48:18kugelgevaerts: yea, I just said I tried it, not that I did it correctly :)
20:51:10 Quit funman ("free(random());")
20:57:56 Join FOAD_ [0] (
21:02:20JdGordon|kugel: I dont know... right now pb might be the only thing affected, but its not a far stretch to assume other will have that problem... there is no hard doing it
21:02:27JdGordon|the load order isnt a big deal
21:03:39kugelwell, an event is send when the UI vp changed, the WPS could easily have an handler and refresh its default vp at any time
21:04:34kugelwithout introducing a bogus load dependency
21:04:35*bertrik made himself a splash with rounded edges :)
21:05:54JdGordon|when the ui vp is changed after loading a setting? or if something forces it to change?
21:06:44JdGordon|ok, the first is acceptable.. I dont tinhk the second is
21:06:50gevaertsbertrik: commit!
21:07:06 Join petur [50] (n=petur@rockbox/developer/petur)
21:07:18kugelit would guarantee that the WPS always has the viewport which the previous list was drawn into (although there's no list from where you can go to the wps and which overrides the UI vp)
21:07:39kugeli.e. consistency!
21:07:53kugeldidn't you want that, have lists and WPS really use the same UI vp?
21:08:04JdGordon|well hang on... when would a screen ever change the ui area on the fly?
21:09:51JdGordon|the only time I can think where this could happen is say a popup (like the splash or context menu), and then the WPS should not know about it
21:10:06JdGordon|of course, changing the ui area for them is the wrong way to do it
21:10:24kugelit's not used for pop ups
21:10:30JdGordon|the ui area shold only change through a sbs loading, or from a cfg load
21:11:09kugelthe vp itself isn't changing, but there are some screens which set the ui vp pointer to a different one temporarily
21:11:25JdGordon|thats the wrong way to do it
21:11:33JdGordon|give me an example...
21:11:36kugelplugins mostly
21:12:04kugelthe loader sets a different viewport
21:12:14JdGordon|plugins are a seperate beast, plugins shuold use the ui area directly, i.e copy the setting into their local viewport vairable
21:12:27JdGordon|the loader *should* set the fullscreen viewport.. thats correct
21:12:30CIA-85New commit by rob (r23408): Implement dbg_save_roms for TCC780x.
21:12:46kugelwell, the plugin loader is the only place where it's done
21:14:15 Quit FOAD (Read error: 110 (Connection timed out))
21:14:15 Nick FOAD_ is now known as FOAD (
21:14:49 Part froggyman
21:14:59kugeland as I mentioned, I want to fix most plugins to use vps better so that won't be needed anymore
21:15:09 Join einhirn_ [0] (
21:16:03JdGordon|fine, but thats still the wrong way to do it
21:16:21JdGordon|plugins should be using the same viewport_get_defaultS() function that the core uses
21:16:38kugelthat's what I'm planning to do
21:16:47JdGordon|the *only* thing that should actually know about the ui area is viewport.c
21:16:58kugelso that game menus for example show the statusbar and draw into the UI vp
21:17:36JdGordon|using get_defaults() or passing NULL to do_menu(), and enabling the bar
21:17:44 Join merbanan [0] (
21:18:02JdGordon|which goes back to me asking, why you think any screen shuold ever change the ui area?
21:18:24kugelI don't think any screen should ever do that
21:18:48JdGordon|great, so then why would the WPS need to be told the area chaged, except for when a new sbs or cfg ios loaded?
21:19:54kugeldo you remember that you were largely against reparsing all skins on loading a single oen?
21:20:27kugelif we ever happen to have it said way, the sbs is going to touch the UI vp, while the WPS wouldn't notice because it's not reparsed
21:20:51kugelwhat you do is simply introducing a load dependency which can be avoided entirely
21:21:35JdGordon|kugel: you havnt noticed by now that I have no problem changing horses mid race if I'm backing the wrong one?!
21:22:13JdGordon|it makes sense that the ui area and sbs are loaded first if they control other skins! I cont see why you tihnk load dependancies are an issue
21:22:16kugelnot to mention that parsing the UI vp setting in a function which is meant to only load skins is a bad hack anyway
21:23:25kugeland load dependencies are bad, they are confusing, error prone and inconsistent
21:24:10kugelwhile letting the skins that really depend on UI vp know when the UI vp changed is perfectly clean
21:24:16JdGordon|bad - no, confusing - maybe without documentation, error prone - not if you follow said doocs, inconsistant - not at all
21:24:32JdGordon|then they should *fully* reload when it changes
21:25:10kugelthe event handler could do that, or just do what's really needed
21:25:42JdGordon|you dont know whats really needed...
21:26:19kugelI surely do
21:26:21JdGordon|there is already a big dependancy chain on themes... 1) init skin buffer, 2) init ui vp 3) init sbs 4) other skins... thats not a problem
21:26:21 Quit einhirn (Read error: 110 (Connection timed out))
21:26:42JdGordon|that deterministic... using events is random
21:30:52 Quit n1s (Read error: 60 (Operation timed out))
21:32:14kugelfine, defend your band aid, I don't think it's the correct way to do this (btw, didn't you say the event is accaptable some lines up?)
21:32:46 Quit parafin (Read error: 104 (Connection reset by peer))
21:33:14 Join parafin [0] (
21:33:41kugelthe event is much more elegant to notify any parts of the code on changes on what they really depend on
21:33:43 Quit parafin (Remote closed the connection)
21:33:59JdGordon|I have no idea what you are takling about anymore
21:34:02JdGordon|what band aid?
21:34:18JdGordon|events are good WHEN THERE IS NO DEPENDANCY
21:34:20JdGordon|there is here
21:35:04 Join parafin [0] (
21:35:04 Quit parafin (Client Quit)
21:35:21JdGordon|good... now neither of us know what we are arguing...
21:35:36kugelI don't understand that bit: "events are good WHEN THERE IS NO DEPENDANCY"
21:36:23 Join parafin [0] (
21:37:10kugelall your changes to settings_apply_skins() are band aid IMO
21:38:00JdGordon|oh you're talking about the patch again...
21:38:29kugelthe whole time
21:39:14kugelI'm also unsure why you're messing with viewportmanager_set_statusbar() in there
21:39:34kugelviewportmanager_theme_changed(THEME_STATUSBAR) does that for you (have you looked at that function?)
21:40:38JdGordon|I cant look at code now...
21:40:48JdGordon|but iirc the if for that is empty with my change
21:41:14JdGordon|viewportmanager is doing too much imo anyway.... sbs isnt really part of the viewportmanager
21:42:55kugelah, and I thought you went back to the current api as you said
21:46:15 Quit robin0800 (Remote closed the connection)
21:46:39 Join phanboy4 [0] (
21:48:47JdGordon|:< that said, for simplicity I dont have major objections to the current linking
21:52:02 Join robin0800 [0] (
21:54:15***Saving seen data "./dancer.seen"
21:54:44kugelJdGordon|: I'm not sure why you're rushing the api changes that way. I'd welcome if we can just focus on getting rid of the classic statusbar now, and discuss and concentrate on changing the api when it's finally dead
21:56:12JdGordon|i didnt change the api
21:56:19JdGordon|unless i uploaded the wrong diff?
21:56:50 Join jgarvey [0] (
21:58:14JdGordon|if you can remove the old bar without doing any other changes then go for it.. imo its easier to do it all in one hit
21:58:49kugelwell, viewportmanager_[get|set]_statusbar are back compared to the previous patches, but there's a lot of (to me) unrelated - lines in viewport.c
21:59:54JdGordon|ok... the removal of the statusbar is actually a side effect of other viewport changes
22:00:44 Quit bubsy (Remote closed the connection)
22:00:48 Quit robin0800 (Remote closed the connection)
22:01:02JdGordon|but once again.. the damn patch isnt ready for commital.. so looking at it and doing a critical review is a waste of time
22:02:39kugelto me the alternative looks like you continue that way I and I complain even more at the end, which also is a big time waste
22:03:40*kugel wants to know why eabi breaks settings_reset()
22:12:21amiconnJdGordon|: I don't think your settings-in-wps idea will work well
22:12:57amiconnWhat happens if you change the theme and then save settings? The settings you have defined for your previous theme will be gone.
22:13:18amiconnThere is no way for the settings code to know which settings a theme has without that theme being loaded
22:17:16JdGordon|amiconn: no, the setting would be saved into a completly sepearte text file under .rockbox/wps
22:19:35 Join TopyMobile_ [0] (
22:21:26 Quit jgarvey ("Leaving")
22:22:25*amiconn doesn't understand what kugel means when talking about "knowing the dimensions of the default vp"
22:22:44amiconnThe default vp is fullscreen, and those dimensions are always known...
22:23:24kugelthe default vp of the wps
22:23:38amiconnThere is only one default vp per display
22:24:22kugelthe wps has an implied non-configurable vp
22:25:29*amiconn would probably call that ui vp or wps root vp
22:25:48 Join tomers [0] (
22:25:53CIA-85New commit by rob (r23409): Make the TCC78x HW debug screen yield by including a button wait timeout.
22:25:55JdGordon|amiconn: the default viewport is the area not being used by anything else.. yes its possibly badly named, but it is not nesecarily the full sceen
22:26:28tomersAny comment reharding "FS #10728 - Cowon D2: Add support for D2 in rbutil" ? I think I would like to commit it soon....
22:26:31 Join bubsy [0] (n=bubsy@
22:26:38*amiconn thinks we should try to improve naming so that it's not ambiguous
22:28:26 Join Matsy [0] (
22:28:34 Quit TopyMobile (Read error: 113 (No route to host))
22:31:16 Join Hillshum [0] (
22:31:51CIA-85New commit by tomers (r23410): FS #10728 - Cowon D2: Add support for D2 in rbutil
22:32:01kugelwasn't there a way to use e200tool to run code directly?
22:32:43JdGordon|amiconn: well, the name "default vp" is only used internally, I think in one #define, otherwise if its only me and kugel arguing we know what it refers to
22:32:59JdGordon|but yes, viewport is full of badly named functions :)
22:34:09MatsyWow, this project is amazing.
22:34:17 Quit esperegu (Read error: 113 (No route to host))
22:34:23andrewRB^^ insightful +5
22:34:30JdGordon|Troll -1
22:36:47 Join shotofadds [0] (n=rob@rockbox/developer/shotofadds)
22:36:52 Join bertrik_ [0] (
22:36:54 Quit bertrik ("De groeten")
22:37:14Bagdertomers: I don't think bootloader/cowon/d2.bin is a good path for the d2 file
22:37:22Bagderand I haven't seen any discussions on that
22:37:26Bagderso I rather not move it there
22:37:43tomersbagder: What do you think is the best location?
22:38:16shotofaddsI suggested /bootloader/cowon/d2/bootloader.bin or somesuch instead
22:39:07tomersshotofadds: Yes, you did that in
22:39:12 Join esperegu [0] (n=quassel@
22:39:40tomersbagder: Do you accept shotofadds' suggestion?
22:39:45BagderI think we could have the cowon and iaudio stuff in the same root dir
22:39:55 Join AleMaxx [0] (
22:40:10shotofaddsBagder: under cowon/ hopefully ;-)
22:40:11tomersAll cowons seems to be labeled as iriver, except the d2
22:40:25BagderI can accept any suggestion, I just want to have it discussed properly first
22:40:32tomersAFAIK, cowon is the company, iriver is the product line
22:40:35Bagderiaudio you mean
22:40:47andrewRBtomers: you mean iaudio?
22:40:48tomerss/iriver/iaudio/ :-)
22:40:49andrewRBoh. ^^
22:40:58shotofaddstomers: yes, and the d2 is not part of that product line so doesn't have iAudio in the name
22:41:02Bagderbut yes, most are named iaudio but I believe there are a few others (non-rockbox) models under the cowon brand too
22:41:27Bagderso yes, the root dir would then most suitably be called cowon actually
22:41:32tomersSo /bootloader/cowon/d2/bootloader.bin is ok?
22:41:44linuxstbBagder: I agree it makes sense to put cowon and iaudio stuff together. Maybe move things into cowon and setup a "iaudio -> cowon" symlink for compatibility?
22:42:02Bagderthat's not a bad idea
22:42:08shotofaddsor /bootloader/cowon/bootloader-d2.bin to follow another inconsistent convention :p
22:42:09 Join HBK- [0] (
22:42:29BagderI prefer /bootloader/cowon/d2/[something]
22:42:32tomersthe symlink should be removed when the next version is released
22:43:14*shotofadds still thinks there should be some kind of warning in the installer if it's not aware of the "SD card caveat"
22:43:17AleMaxxin what way are jz47xx and cc1600 similar? Im only asking since I read some things on the internet. Im aware that the first uses a MIPS code while the latter is ARM powered.
22:43:19linuxstbshotofadds: I think we should try and avoid using short names like "d2" - so bootloader-cowond2.bin
22:43:33tomersMaybe we should rename all files according to a convention we decide, and set code accordingly, once and for all. Symlink can be used to provide backward compitabilityu
22:44:02AlexPOn a vaguely related note (i.e. not really), is there are reason why the beast isn't unstable now? It has a released linux and windows beastpatcher and bootloader, and install instructions in the manual
22:44:23linuxstbAlexP: I was going to ask you that - were you the person who added instructions to the manual?
22:44:30*rasher points once again to">
22:44:31tomersshotofadds: I agree about the warning. Needs some work though
22:44:39AlexPlinuxstb: yes
22:44:55 Join Llorean [0] (
22:45:17linuxstbshotofadds: So what does a bootloader build currently call the D2 bootloader?
22:45:37shotofaddsbootloader.bin, so I reckon the path should be bootlaoder/cowon/d2/bootloader.bin
22:45:37 Join saratoga [0] (i=9803c6dd@gateway/web/freenode/x-sdrqtzrlmvamobib)
22:45:50shotofaddswithout the typo, obviously :)
22:46:16BagderI put it there now
22:46:20JdGordon|keep the typo!
22:46:38MatsyOh, I am amazed. I joined this channel because of some discussion in January 2006, and most of the users of the chat are still here.
22:46:56MatsySome devotion ;p
22:47:05JdGordon|yep, no matter how har we try... we just cant get rid of them!
22:47:07BagderMatsy: or just lack of real lives ;-)
22:47:18AlexPSad but true
22:47:23linuxstbshotofadds: I get a rockbox.d2...
22:47:42shotofaddslinuxstb: that's just to throw you off the scent :p
22:47:44linuxstbshotofadds: "bootloader.bin" is the binary, which is then processed by "boottool"
22:47:56shotofaddsit really shouldn't be processed by anything
22:47:59tomersbagder: Are you going to fix rbutilqt.ini, or should I do it?
22:48:00linuxstbIf that's not the real bootloader people should use, then change boottool to "cp"
22:48:08 Join funman [0] (n=fun@rockbox/developer/funman)
22:48:14linuxstbAnd "bootoutput" (IIRC) to "bootloader-cowond2.bin"
22:48:17MatsyBagder: Haha, I certainly don't hope that's the reason ;o
22:48:18kugelhm, /me thouht he could use e200tool recover
22:48:30shotofaddslinuxstb: I think boottool *is* cp, it's just using the wrong filename ATM
22:48:52linuxstbshotofadds: No, it's scramble
22:48:52MatsyIt would be safe to ask a question here, then. ;p
22:49:16Bagdertomers: please go ahead, I'm actually stuck somewhere completely different...
22:49:29tomersbagder: OK
22:49:30BagderMatsy: fire away and we'll find out
22:49:33MatsyI am trying to debug the firmware of a device that runs on a ARM720t processor.
22:49:34shotofaddslinuxstb: really? I'm trying to think what possible use that would have
22:49:43linuxstbshotofadds: Uploading it directly with tcctool?
22:50:02*linuxstb guesses this was copied from the logik dax...
22:50:15tomersbagder: Have you move the file? I can't see it
22:50:18MatsyYet, I've tried IDA to do it for me. But, it doesn't seem to want to disassemble it.
22:50:25shotofaddsno, you need to build a different version if you want to do that (something to do with the dual-boot, but it's so long ago I forget)
22:50:34Bagdertomers: and you check the download master server?
22:51:01BagderMatsy: then you probably do it wrong, as IDA can indeed disassemble ARM code, both normal and thumb
22:51:03linuxstbBagder: What did you change it to? I thought the discussion was still ongoing...
22:51:08MatsyYeah, exactly what I thought.
22:51:13tomersbagder: Checked
22:51:13 Quit HBK- ()
22:51:13MatsyIt normally goes perfectly fine.
22:51:18Bagderyeah, I'm ready to move it again ;-)
22:51:22shotofaddslinuxstb: the D2 is happy with any raw binary through tcctool anyway - the scramble stuff is only necessary if you want to use the OF to flash it.
22:51:37MatsyHowever, I don't see ARM720T in that list. Does IDA even support ARM7?
22:51:50Bagdertomers: that's one of the mirror servers, they sync once per hour...
22:51:56saratogaARM720T is ARMv4
22:52:02linuxstbshotofadds: Ah yes, I think the DAX was the same. But anyway, I would vote for "bootloader-cowond2.bin", so it's fully self-contained.
22:52:12saratogawhich everythign supports
22:52:45BagderMatsy: I know (many) people have used it on ARM720* chips
22:52:49AleMaxxMatsy, its the first one in the listbox
22:53:03MatsyMy firmware file might be corrupted then.
22:53:16Bagderor it isn't code that you look at ;-)
22:53:18tomersbagder: Is there a direct link to the master download server?
22:53:22MatsyYeah, or that.
22:53:23LloreanJdGordon|: Instead of that "custom options" why not just allow users to define a variable or few at the beginning of the WPS. Since the custom options would require users to edit a file, why not just have that file be the .wps? One or two quick edits to the first two lines and they get their customized layout.
22:53:41AlexPlinuxstb: So can you think of anything holding it (the beast) back?
22:54:08Bagdertomers: I try not to expose that, to prevent "people" from going there directly ;-) but yeah (haxx [dot] rockbox [dot] org)
22:54:24shotofaddslinuxstb: yes, I agree with the naming. I wonder if tomers,Bagder are listening :)
22:54:34*Bagder will adapt
22:54:34shotofaddsbootloader-cowond2.bin, that is.
22:54:38AleMaxxMatsy, check it in a hexeditor with column length multiple of eight, there should be rows with a lot of 0xEx, if these are not on dword boundarys you need to align that
22:55:03AleMaxxmultiple of 4 is of course fine too
22:55:06 Join T44 [0] (
22:55:15 Quit bmbl ("Bye!")
22:55:18JdGordon|Llorean: thats like what I've done in but the problem with that is (especially for shippd themes), is they get overwritten.. also if all themes used the same settings then you would set them once and they would instantly work for every skin
22:55:24tomersshotofadds: Listening. bootloader/cowon/d2/bootloader-cowond2.bin ?
22:55:40shotofaddsyes, check the server...
22:55:44shotofaddsit's there already
22:55:52JdGordon|the bigger problem is the skin engine cant handle these variable thingies in the parser... thats maybe something to add though
22:56:15tomersIn rbuti.ini, I set bootloadername=cowon/d2/bootloader-cowond2.bin - is that OK?
22:56:18AleMaxxand I meant row length
22:57:14shotofaddstomers: what does platform id mean in rbutil.ini? Is it supposed to be related to the id in tools/confiure?
22:57:21linuxstbshotofadds: We should change tools/configure (and the build system so it checks for bootloader-cowond2.bin as the output file)...
22:57:29*shotofadds was about to do that
22:57:33MatsyI can't find any rows with a lot of 0xEx entries.
22:57:45AleMaxxI meant columns
22:58:03AleMaxxbasically there should be a lot of dwords beginning with 0xE
22:58:07MatsyOh, right
22:58:10MatsyI found a piece
22:58:18MatsyEvery 4 bytes ends with 0xEx
22:58:29tomersshotofadds: Platform id where?
22:58:42AleMaxxMatsy, then its probably good arm code
22:59:02MatsyThat only applies for a certain piece of the binary file I am looking at.
22:59:04AleMaxxMatsy, when youre in IDA you need to hit C, you know that right?
22:59:05 Quit KBH (Connection timed out)
22:59:07MatsyThe rest doesn't seem to follow that rule.
22:59:10MatsyYeah, I do.
22:59:14 Join Omlet [0] (
22:59:16 Join Omlet05 [0] (
22:59:17 Quit Omlet (Broken pipe)
22:59:17 Quit Omlet05 (Broken pipe)
23:00:13tomersshotofadds: This is what I assumed
23:00:48*shotofadds doesn't know, I just wondered if the number was significant anywhere
23:01:18AleMaxxMatsy, so these 0xEs should be on offsets 3,7,B,F
23:01:53AleMaxxthats the case?
23:02:02MatsyOn some parts of the binary file, yeah.
23:02:16AlexPWhat player is this for?
23:02:20AleMaxxthen go to these offsets in the firmware and press 'c'
23:02:49MatsyOh, I'll try that.
23:03:01shotofaddslinuxstb: how is the DAX coming on, btw ;-)
23:03:06linuxstbshotofadds: I think you need to ask domonoky1 or bluebroth3r
23:03:10 Join froggyman [0] (
23:03:16linuxstb(re: the platform ID)
23:03:26tomersshotofadds: I think it is actually like the mapping you can see at ./tools/release/ +136
23:03:42linuxstbshotofadds: Waiting for some talented person to finish it...
23:04:37*domonoky1 wakes up..
23:05:39 Quit froggyman (
23:05:39 Quit Llorean (
23:05:39 Quit shotofadds (
23:05:39 Quit phanboy4 (
23:05:39 Quit merbanan (
23:05:39 Quit einhirn_ (
23:05:39 Quit petur (
23:05:39 Quit FOAD (
23:05:39 Quit AleMaxx (
23:05:39 Quit J-23 (
23:05:39 Quit JdGordon| (
23:05:39 Quit andrewRB (
23:05:39 Quit Sajber^ (
23:05:39 Quit adiroiban (
23:05:39 Quit ender` (
23:05:39 Quit amiconn (
23:05:39 Quit AndyIL (
23:05:39 Quit elcan (
23:05:39 Quit AB3JU (
23:05:39 Quit Bob_C (
23:05:39 Quit Zarggg (
23:05:39 Quit ej0rge (
23:05:39 Quit togetic (
23:05:39 Quit Overand (
23:05:39 Quit Bagder (
23:05:39 Quit markun (
23:05:39 Quit Tuplanolla (
23:05:39 Quit BlakeJohnson86 (
23:05:39 Quit gitster (
23:05:39 Quit saratoga (
23:05:39 Quit evilnick_B (
23:05:39 Quit antil33t (
23:05:39 Quit rasher (
23:05:39 Quit jhulst (
23:05:39 Quit Lss (
23:05:39 Quit AlexP (
23:05:39 Quit topik (
23:05:39 Quit yosafbridge (
23:05:39 Quit ThomasAH (
23:05:39 Quit jds2001 (
23:05:39 Quit pixelma (
23:05:39 Quit jasio (
23:05:39 Quit blithe (
23:05:39 Quit jordan` (
23:05:39 Quit n17ikh (
23:05:39 Quit fish_ (
23:05:39 Quit tarbo (
23:05:39 Quit GeekShadow (
23:05:39 Quit liar (
23:05:39 Quit Horscht (
23:05:39 Quit TheSeven (
23:05:39 Quit Rondom (
23:05:39 Quit gevaerts (
23:05:39 Quit Hadaka (
23:05:39 Quit jon-kha (
23:05:39 Quit Slasheri (
23:05:39 Quit shodanX (
23:05:39 Quit krazykit (
23:05:39 Quit beta2k (
23:05:44 Quit Dhraakellian (
23:05:44 Quit fxb (
23:05:44 Quit YPSY (
23:05:44 Quit HellDragon (
23:05:44 Quit gtkspert_ (
23:05:44 Quit Unhelpful (
23:05:44 Quit Kohlrabi (
23:05:44 Quit pjm0616 (
23:05:44 Quit rphillips (
23:05:44 Quit Tristan (
23:05:44 Quit chaos (
23:05:44 Quit ChanServ (
23:05:44 Quit esperegu (
23:05:44 Quit Hillshum (
23:05:44 Quit bubsy (
23:05:44 Quit tomers (
23:05:44 Quit TopyMobile_ (
23:05:44 Quit parafin (
23:05:44 Quit dfkt (
23:05:44 Quit Xerion (
23:05:44 Quit xavieran_ (
23:05:44 Quit lostlogic (
23:05:44 Quit dionoea (
23:05:44 Quit zu_ (
23:05:44 Quit r00s (
23:05:44 Quit ps-auxw (
23:05:44 Quit bzed (
23:05:44 Quit goffa (
23:05:44 Quit freqmod_qu (
23:05:44 Quit kugel (
23:05:44 Quit toffe82 (
23:05:44 Quit Topy44 (
23:05:44 Quit Galois (
23:05:44 Quit jfc^3 (
23:05:44 Quit FlynDice (
23:05:44 Quit niekie (
23:05:44 Quit Shaid (
23:05:44 Quit maraz_ (
23:05:44 Quit linuxstb (
23:05:44 Quit tmzt (
23:05:44 Quit tha (
23:05:44 Quit avacore^ (
23:05:44 Quit daurn| (
23:05:44MatsyHahahaha, AleMaxx
23:05:44MatsyThanks ;p
23:07:50NJoinChanServ [0] (ChanServ@services.)
23:07:50NJoinmarkun [50] (n=markun@rockbox/developer/markun)
23:07:50NJoinej0rge [0] (
23:07:50NJoinBagder [241] (n=daniel@rockbox/developer/bagder)
23:07:50NJoingitster [0] (
23:07:50NJointogetic [0] (n=togetic@unaffiliated/ibuffy)
23:07:50 Join Tuplanolla [0] (n=jani@unaffiliated/tuplanolla)
23:07:50NJoinBlakeJohnson86 [0] (
23:07:50NJoinZarggg [0] (
23:07:50NJoinOverand [0] (
23:07:50NJoinBob_C [0] (
23:07:50NJoinAB3JU [0] (
23:07:50NJoinelcan [0] (
23:07:50NJoinAndyIL [0] (n=pasha_in@
23:07:50NJoinamiconn [0] (i=quassel@rockbox/developer/amiconn)
23:07:50NJoinender` [0] (
23:07:50NJoinadiroiban [0] (
23:07:50NJoinSajber^ [0] (
23:07:50NJoinandrewRB [0] (
23:07:50NJoinJdGordon| [0] (n=Miranda@nat/microsoft/x-qrhtgnynuoepaplm)
23:07:50NJoinJ-23 [0] (
23:07:50NJoinFOAD [0] (
23:07:50NJoinpetur [50] (n=petur@rockbox/developer/petur)
23:07:50NJoineinhirn_ [0] (
23:07:50NJoinmerbanan [0] (
23:07:50NJoinphanboy4 [0] (
23:07:50NJoinshotofadds [0] (n=rob@rockbox/developer/shotofadds)
23:07:50NJoinAleMaxx [0] (
23:07:50 Join Llorean [0] (n=DarkkOne@rockbox/user/Llorean)
23:07:50NJoinfroggyman [0] (
23:07:50 Join MethoS- [0] (n=clemens@
23:07:50NJoinsaratoga [0] (i=9803c6dd@gateway/web/freenode/x-sdrqtzrlmvamobib)
23:07:50NJoinesperegu [0] (n=quassel@
23:07:50NJoinHillshum [0] (
23:07:50NJoinbubsy [0] (n=bubsy@
23:07:50NJointomers [0] (
23:07:50NJoinTopyMobile_ [0] (
23:07:50NJoinparafin [0] (
23:07:50NJoingoffa [0] (n=goffa@
23:07:50NJoinfreqmod_qu [0] (
23:07:50NJoinkugel [0] (n=kugel@rockbox/developer/kugel)
23:07:50NJoinGeekShadow [0] (n=Antoine@reactos/tester/GeekShadow)
23:07:50NJoinantil33t [0] (n=Mudkips@
23:07:50 Join rasher [50] (n=rasher@rockbox/developer/rasher)
23:07:50NJoinliar [0] (n=liar@
23:07:50NJoinjhulst [0] (
23:07:50NJointoffe82 [0] (n=chatzill@
23:07:50Mode"#rockbox +o ChanServ " by
23:07:50NJoinevilnick_B [0] (i=0c140464@rockbox/staff/evilnick)
23:07:50NJoinHorscht [0] (n=Horscht2@xbmc/user/horscht)
23:07:50NJoindfkt [0] (i=dfkt@unaffiliated/dfkt)
23:07:50 Join HellDragon [0] (i=jd@Wikipedia/HellDragon)
23:07:50NJoinTheSeven [0] (n=theseven@rockbox/developer/TheSeven)
23:07:50NJoinRondom [0] (n=Rondom@
23:07:50NJoinTopy44 [0] (
23:07:50NJoingevaerts [0] (n=fg@rockbox/developer/gevaerts)
23:07:50NJoinpixelma [0] (i=quassel@rockbox/staff/pixelma)
23:07:50NJoinLss [0] (
23:07:50NJoinGalois [0] (
23:07:50NJoinjfc^3 [0] (
23:07:50NJoinXerion [0] (
23:07:50NJoinHadaka [0] (n=naked@
23:07:50NJoinjasio [0] (
23:07:50NJoinjon-kha [0] (i=jon-kha@
23:07:50NJoinblithe [0] (
23:07:50NJoinAlexP [0] (n=alex@rockbox/staff/AlexP)
23:07:50NJoinjordan` [0] (n=jordan@
23:07:50NJoinlinuxstb [0] (n=linuxstb@rockbox/developer/linuxstb)
23:07:50NJoinxavieran_ [0] (
23:07:50NJoinniekie [0] (
23:07:50NJoinFlynDice [0] (
23:07:50NJoingtkspert_ [0] (
23:07:50 Join Unhelpful [0] (n=quassel@rockbox/developer/Unhelpful)
23:07:50NJoinn17ikh [0] (
23:07:50NJoinTristan [0] (i=tristan@
23:07:50NJoinThomasAH [0] (
23:07:50NJoinavacore^ [0] (
23:07:50NJoinkrazykit [0] (n=kkit@
23:07:50NJoinyosafbridge [0] (
23:07:50NJoinjds2001 [0] (n=jds2001@fedora/jds2001)
23:07:50NJoinbeta2k [0] (n=beta@
23:07:50NJoindaurn| [0] (n=daurnima@freenode/staff/daurnimator)
23:07:50NJoinps-auxw [0] (
23:07:50NJoindionoea [0] (n=dionoea@videolan/developer/dionoea)
23:07:50NJoinpjm0616 [0] (n=user@
23:07:50NJointopik [0] (
23:07:50NJoinrphillips [0] (
23:07:50NJoinr00s [0] (
23:07:50NJointmzt [0] (
23:07:50NJoinfish_ [0] (
23:07:50NJoinlostlogic [50] (n=lostlogi@rockbox/developer/lostlogic)
23:07:50NJointha [0] (
23:07:50NJoinchaos [0] (n=ch4os@gentoo/user/ch4os)
23:07:50NJoinzu_ [0] (
23:07:50NJoinfxb [0] (n=felixbru@
23:07:50NJoinKohlrabi [0] (
23:07:50NJoinshodanX [0] (
23:07:50NJoinYPSY [0] (n=ypsy@
23:07:50NJoinDhraakellian [0] (
23:07:50NJoinShaid [0] (
23:07:50NJoinbzed [0] (
23:07:50NJoinSlasheri [0] (i=miipekk@rockbox/developer/Slasheri)
23:07:50NJoinmaraz_ [0] (
23:08:02CIA-85New commit by rob (r23411): Update the Cowon D2 bootloader binary name
23:08:14tomerslinuxstb: This FS stood there for two days... I got many comments and fixed them all, then waited for couple of days without any further remarks from anybody
23:08:19MatsyEverything works like a blast, now ;)
23:08:39tomerss/two days/one day/
23:08:47linuxstbtomers: Sure, but there's no point asking for final comments, then only waiting a couple of minutes...
23:08:49CIA-85New commit by rob (r23412): Change the D2 bootloader output filename to bootloader-cowond2.bin
23:09:16*shotofadds didn't see the final patch til after it was committed... but that's my fault
23:09:39 Join HBK [0] (
23:10:36tomerslinuxstb: That was just for making sure nobody had 'something in his stomach'. If there was some serious comment, I would expect it to be written in the FS item. Do you think I should have asked everybody in the chat personally?
23:10:44linuxstbtomers: But yes, I did have chance to comment, but hadn't got round to it. I didn't like a lot of your changes to mktccboot, such as using the "OF" abbreviation in lots of places instead of the existing "original firmware".
23:11:34AlexPtomers: His point isn't that you should ask everybody, but that if you ask for comments it is pointless if you then only wait two minutes.
23:11:41 Quit TheSeven (Nick collision from services.)
23:11:44AlexPPeople don't stare constantly at their screens
23:11:53AleMaxxMatsy, congratulations :-)
23:11:57 Join The_Seven [0] (n=theseven@rockbox/developer/TheSeven)
23:12:05domonoky1how this rbutil d2 patch handle the sd/bootloader on nand thing ?
23:12:09 Nick The_Seven is now known as TheSeven (n=theseven@rockbox/developer/TheSeven)
23:12:10AlexPPeople might not have seen the task, and gone to look at it when you mentioned it for instance
23:12:11tomersI agree with you. I should have waited longer
23:12:26 Quit Topy44 (Read error: 110 (Connection timed out))
23:12:57tomerslinuxstb: What variable names do you think I should fix?
23:13:21 Quit merbanan (Success)
23:13:24 Quit Hillshum (Operation timed out)
23:13:26*tomers don't know what to set in bootloaderfile (rbutil.ini) any help?
23:13:29linuxstbtomers: I don't really care enough for you to go back and change things. It's mainly the principle of not changing things without a good reason.
23:13:40 Nick bertrik_ is now known as bertrik (
23:13:48domonoky1does the user have to do anything after the d2 bootloader install ? if so, a case to BootloaderInstallBase::PostInstallHints should be added.
23:14:23 Join DerPapst [0] (
23:14:42tomerslinuxstb: I wanted it to be the same abbreviation as the helper function I wrote there. Nothing more.
23:14:44shotofaddsdomonoky1: yes, they need to reboot into the OF to flash it, like the iRivers
23:14:54domonoky1tomers: bootloadername is the bootloader on the server. bootloaderfile is the bootloader filename on the target (if needed by the install method)
23:15:25linuxstbtomers: My other comment would have been about the Makefile - it seems very hard to read with the "PLAT" variable and the name MKPLATBOOT. Why did you use that?
23:15:30 Quit HBK ()
23:15:39 Join HBK [0] (
23:15:46shotofaddsthat is the other problem: depending on the firmware version the on-target filename can be any of D2.bin, D2_N.bin, D2P.bin, D2P_N.bin...
23:16:02tomersdomonoky1: Does the cp install method uses this value? If so, it should stay d2n.bin.
23:16:28shotofaddssorry, 50% right: D2N.bin and D2N_P.bin
23:16:40pixelmashotofadds: really reboot into the OF and chose a menuitem there or more like the other Iaudios - insert charger and if there is a file with a certain name in a certain place it'll be picked up by the Cowon loader for flashing the new bootloader?
23:17:33tomerslinuxstb: I agree with that. I tried to work on making the mkamsboot's Makefile similar to the mktccboot's Makefile, so they could somehow be united... I created a diff and worked my way with constants to achieve similarities. I think now it's bloated with constants
23:17:34shotofaddspixelma: the latter, actually
23:17:34linuxstbtomers: Also, shouldn't mktccboot.c #include your new mktccboot.h ?
23:17:35domonoky1tomers: asfar as i know, bootloaderinstalltcc.cpp should do all coping itself.
23:18:11bluebroth3rtomers: I'm still missing handling of the two drives issue :(
23:18:43 Nick bluebroth3r is now known as bluebrother (n=dom@rockbox/developer/bluebrother)
23:18:48 Join Topy [0] (
23:18:48*shotofadds is surprised not to have got any red
23:18:57domonoky1tomers, shotofadds: you could use the filename of the file given by the user, as done in bootloaderinstallams.cpp
23:19:01linuxstbtomers: I think there's a danger of trying to unify things too much - copy/paste is bad, but sometimes unifying things just makes it too hard to read, and hence less maintainable than two copies of essentially the same thing.
23:19:12shotofaddsdomonoky1: yes, that would be the best thing
23:19:40tomerslinuxstb: I'll fix that tomorrow.
23:19:56*bluebrother wonders why that patch already got in
23:20:05shotofaddsit would also be nice to check if the device the user has plugged in can accept the filename the user has chosen
23:20:12shotofaddsyou can detect that using the USB ids
23:20:50domonoky1bluebrother: i see no problem with commting a not completly working bootloader/target for rbutil, we can easily disable it if it still doesnt work completly when we make a new release.
23:21:20shotofaddscommit early commit often, no?
23:21:31CIA-85New commit by tomers (r23413): rbutil: Update bootloader name for cowon d2
23:21:50linuxstbshotofadds: Depends... That could also be "post patches early, post patches often..."
23:22:17bluebrothershotofadds: no. It's "release early, release often".
23:22:18shotofaddstomers: thanks for looking into this. Do you have any other areas in mind that you want to work on>
23:22:27linuxstbBut it's also true that most people ignore patches until they're committed...
23:22:34bluebrotherand for work in progress I go with linuxstb
23:22:35tomersshotofadds: No
23:22:45shotofaddshehe ;)
23:23:41bluebrotherthe big problem with patches is that irc-used guys expect them to be handled fast. But that's not the case (and is unlikely to be anyway, who can monitor the tracker all day long?)
23:23:49CtcpVersion from freenode-connect!freenode@freenode/bot/connect
23:23:53tomerslinuxstb: The patch went through about four iterations, and in the last day nobody comment on it, after I fixed all what was said. I thought its good enough to be committed, knowing development should continue by further patches/commits - like the two drives issue.
23:24:27 Quit T44 (Read error: 60 (Operation timed out))
23:24:56linuxstbtomers: That's true, but a day isn't that long. I think that when you ask in IRC if a patch is OK to commit, you need at the very least need one person to say "yes, I agree", rather than just assuming that silence means agreement.
23:24:59bluebrothertomers: "in the last day". Great argument. May I suggest that not everybody can monitor / work on patches each day? I've had several days in a row without any time for monitoring development in the past, is this that uncommon?
23:25:07AlexPtomers: I think people are saying that you should have brought it up here then waited a good long time (a day isn't long) for more comments
23:25:17*bluebrother thinks patches should be in the tracker *at least* a week
23:25:20*domonoky1 thinks someone how knows the d2 bootloader install process should commit a postInstallHint() for d2 :-)
23:26:15*tomers I totally agree with you all. I should have had more patience. SORRY. Let's end this discussion please
23:26:33domonoky1looks like only m:robe500 is missing from the unstable targets in rbutil ?
23:26:46AlexPdomonoky1: And shortly the beast :)
23:26:59pixelmadomonoky1: if I understand shotofadds' answer correctly it works like the other Iaudios
23:27:15*tomers was a bit over-motivated with the D2 install issue :-)
23:27:24linuxstbWhat about supporting everything except the bootloader install for the Beast? Do we have any existing targets like that? e.g. e200r?
23:28:16tomersshotofadds: I didn't get what you were suggesting (testing filename) can you explain again?
23:28:22shotofaddspixelma, domonoky1: the post install step is just disconnect the player and turn it on
23:28:24domonoky1e200r is just unsupported in rbutil, it doesnt know a e200r target (although of course everthing exept bootloader install works)
23:28:35shotofaddstomers: the firmware filename differs depending on the OF version
23:29:20domonoky1shotofadds: ah, then no postInstallHint() needs to used. but the filenaming should perhaps be done similar to ams.
23:29:26shotofaddsD2 v1.x/3.x expect D2.bin, D2 v2.x/4.x expect D2N.bin. D2+ is the same, except *_P.bin
23:29:35shotofaddsit's a mess
23:29:42tomersshotofadds: So some players accept *only* d2n.bin, while other accept only ___.bin (some other name) ?
23:29:56pixelmashotofadds: then slightly different as on the other Iaudios you have to insert the charger to push off the update
23:30:20shotofaddstomers: yes, you need the correct filename for the OF version that's currently on the player
23:30:36shotofadds(although you can flash any version by simply renaming the file)
23:30:44shotofaddsit's all rather confusing
23:30:46pixelmas/as on/to
23:31:13linuxstbdomonoky1: What I'm suggesting is that "we" add the capability to rbutil to support a target, but with bootloader installation just being a link to the manual/wiki, rather than actually doing the install.
23:31:16shotofaddspixelma: inserting the charger would have the same effect on the D2 as well, it's just not strictly necssary
23:31:28domonoky1so we could also just add the file 4 times with all names ? :-)
23:31:35shotofaddsor that :-)
23:32:17tomersshotofadds: Is there any other player install method who dynamically decides the bootloader filename according to USB ID?
23:32:23domonoky1linuxstb: sounds good. could perhaps be done with a bootloaderInstallNone class :-)
23:32:33shotofaddstomers: actually we can't auto detect the filename based on USB ids, since all D2s use the same ID (only D2+s have different ids for different firmware versions)
23:32:55shotofaddsaargh, just use the filename supplied by the user
23:33:15*shotofadds will be afk for a bit
23:33:20linuxstbSave 4 copies? ;)
23:33:40bluebrotherlinuxstb: that was what I intended to do for the beast for now :)
23:33:46tomerslinuxstb: If linking to manual, you should also provide the patching tool as WIn32 binary for Windows users
23:33:59linuxstb"I" should? ;)
23:34:23linuxstbWhich patching tool are you talking about?
23:34:34pixelmaI believe a similar idea was discussed for the X5 and X5Vs (the ones without radio which also need a slightly different name for the firmware files and won't accept the other)
23:34:38tomersmk(whatever)boot and such...
23:34:50pixelmanot sure how it is done at the moment
23:34:55AlexPtomers: If you mean the beast, then it is provided, and is linked in the manual...
23:34:57linuxstbtomers: Yes, we do.
23:35:35linuxstbtomers: Targets don't (or at least shouldn't...) be "unstable" without binaries on the download servers.
23:37:21*linuxstb is reminded that he doesn't understand what is meant by "compiled bootloader" on
23:37:33linuxstb(as compared to a released bootloader)
23:38:19domonoky1linuxstb: perhaps it means it can be a RC (or beta) bootloader
23:38:40kugela EABI build apparently doesn't boot because the scrolling thead doesn't get a slot! :/
23:38:48*linuxstb pings saratoga (who wrote that page...)
23:39:34linuxstbdomonoky1: Maybe, although I'm not sure we should be making those available to end-users, especially on the targets where it's hard/impossible to recover from a bad bootloader.
23:40:03*linuxstb would prefer that to be changed to say "released bootloader"
23:40:35linuxstbdomonoky1: And if we distribute them, we still need to go through most of the release process - i.e. tag it in svn so the source is easily available for that binary.
23:40:36domonoky1linuxstb: for example the ams sansa long had (or still have a rc bootloader), it was pretty safe, but not declared 1.0
23:40:40*tomers Got to go. Will read your posts tomorrow...
23:40:41 Quit ender` (" We are surprised at our own versatility in being able to fail in so many different ways. -- Samuel McChord Crothers")
23:40:58 Quit tomers ("ChatZilla 0.9.85 [Firefox 3.0.14/2009090216]")
23:44:02linuxstbdomonoky1: So you're happy with the term "compiled bootloader" ?
23:44:44domonoky1linuxstb: jup, barriers for unstable should be low.
23:45:36 Join webguest51 [0] (
23:46:05webguest51can anybody help me? i have a question
23:46:12linuxstbdomonoky1: But releasing a bootloader is just a matter of some administrative work, rather than the state of the code.
23:46:13domonoky1webguest51: just ask.
23:46:23webguest51does this work for 2nd gen nano?
23:46:31 Quit Matsy ()
23:46:58webguest51when i run the utility it only shows the first gen and nothing for the 2nd
23:47:08linuxstbwebguest51: See for the list of supported devices, then read the manual for the installation instructions.
23:47:28webguest51okay will it also tell me what it will let me do, like play movies and such?
23:47:31linuxstbRockbox Utility doesn't support the 2nd gen Nano. That's stated in the manual.
23:47:37 Join linuxguy4 [0] (
23:47:37domonoky1linuxstb: the administrative work was probably the same for the RC bootloaders (tag and put binarys to the server) it just wasnt deemed ready enough to be called "release worthy" :-)
23:47:43 Join T44 [0] (
23:47:50linuxstbwebguest51: What it will let you do is also described in the manual... ;)
23:48:24 Quit webguest51 (Client Quit)
23:48:38 Quit einhirn_ (Read error: 104 (Connection reset by peer))
23:48:59linuxstbdomonoky1: If the bootloader isn't deemed release-worthy, then I wouldn't call the device unstable (and hence supported).
23:49:34linuxstbA 1.0 bootloader doesn't have to be perfect, just safe from bricking the device, and functional.
23:49:40domonoky1linuxstb: that depends on how you define "release-worthy".
23:49:43 Quit TopyMobile_ (Operation timed out)
23:49:51linuxstb"safe from bricking the device, and functional"
23:50:15domonoky1linuxstb: not-bricking bootloader is of course neccesary for unstable.
23:50:36linuxstbdomonoky1: But I think that's all I'm trying to say - if the bootloader is good enough for the target to be called unstable, then that bootloader should be released.
23:50:49kugelsomething goes REALLY wrong there
23:51:37domonoky1its just that the word "released" implys much more for some people.. but who cares about them, so change it :-)
23:53:47*petur is looking for a manual guru
23:53:51*domonoky1 goes sleeping..
23:53:58 Quit domonoky1 (Read error: 54 (Connection reset by peer))
23:54:04kugelto illustrate the problem
23:54:15petur"LaTeX Error: File `xcolor.sty' not found." <−− what's wrong?
23:54:18***Saving seen data "./dancer.seen"
23:54:25 Join saratoga_lab [0] (i=9803c264@gateway/web/freenode/x-ombmoqbdaemgdfus)
23:54:30saratoga_lab"Note: Rockbox Utility does not currently support the Gigabeat S and you will therefore need to follow the manual install instructions below."
23:54:45saratoga_labcouldn't the automatic install section simply be omitted on the beast?
23:54:51gevaertspetur: try installing the latex-xcolor package
23:54:52funmankugel: aren't (some?) thread functions written in asm ?
23:55:11funmanpetur: you need to install the package providing xcolor
23:55:15kugelthe two functions from the pastie are entirely C
23:55:16linuxstbsaratoga_lab: Doesn't the manual also say that that section still contains useful information?
23:55:28linuxstb(the Nano2G has the same Note:)
23:55:49funmanpetur: according to apt, latex-xcolor on ubuntu
23:55:53linuxstbsaratoga_lab: But yes, I agree that could probably be written better.
23:55:58saratoga_labah so it does
23:56:13saratoga_labregardless, I think the S should be supported now
23:56:27linuxstbsaratoga_lab: Did you see my question about your term "compiled bootloader" on the TargetClassifications page? (about 30 minutes ago in this channel)
23:56:52saratoga_lablinuxstb: no i haven't been following the logs today due to being in and out, i'll look now
23:57:36saratoga_lablinuxstb: I meant a bootloader binary that can be used without setting up the rockbox dev tools
23:58:30 Quit linuxguy3 (Read error: 113 (No route to host))
23:58:41saratoga_labi guess that would be the same as released, although for unstable I wanted a lower bar if possible for targets where the bootloader install is completely safe (AMS, etc)
23:58:59saratoga_labbut i don't really care about that now

Previous day | Next day