This task is about the implementation of the MTP protocol for rockbox.

The code can be found in a github repository:
http://github.com/godeater/rockbox/tree/mtp The two main files are:
firmware/usbstack/usb_mtp.{c|h}
But there also are some other modification to be able to run it using the debug menu.

Adds “Binary Skip” to the skip length menu,

This allows navigation by means of binary search within a track.

If no search key is pressed for 5 seconds, the current binary search is considered as having “finished” and pressing a search key subsequently initiates a new binary search.

This is a very early patch to enable USB on the mini2440.

This patch let rockbox react on USB, but unfortunatly i dont get further then the first reset interrupt from USB. Something in the init has to be wrong, so i dont get ep0 interrupt which should follow the reset.

Because of this problems, no attempt has been made to implement the acutal data transfer functions.

If someone wants to step in and debug.. feel free !

The attached patch adds an indication of the current touchscreen mode to the "classic_statusbar" skin. This is useful since some screens automatically change mode, and there's currently no indication the user when this happens. The patch does the following:

- Adds a new WPS tag (%Tm) to return the current touch mode (Absolute|Grid)
- Updates classic_statusbar.sbs to conditionally show the correct icon

The reason I've made this a Flyspray task rather than just committing it is because the additions to classic_statusbar.sbs will also apply to non-touchscreen targets (though nothing will appear on screen). While the impact on these targets is minimal, I'm not sure about committing it since it will at least cause the new touchmode.bmp to be loaded unneccesarily.

Also, feel free to design some better icons!

This patch adds a secondary backlight timeout (similar to what the the iPod OF does).

It adds two extra settings: secondary backlight timeout and secondary brightness (I'm not sure wrt the naming of those, any suggestions?)

The code isn't very clean, but it works (except for CONFIG_BACKLIGHT_FADING == BACKLIGHT_FADING_PWM targets, for which I haven't implemented this).

Please test this.

I have started an discussion about the virtual keyboard in the mailing list.

In my opinion the text_editor’s control should be similar to the calculator:
- use the whole screen (no tiny font)
- support for touchscreen (raster map)
- separate characters into categories

I adjust the calculator plugin to show my concept.

if bookmark=ask is set and the mpeg is stopped, the user is asked for pressing select (for make bookark) or any key (for discard)
on apple devices: if the clickwheel is used for discarding, the bookmark is not set (as intended) but the clickwheel is not working anymore.
to get the clickwheel working again, the keyboard lock has to be set and unset and select has to be pressed.

This patch is a proof-of-concept for an adaptive runtime estimation.

Reason to implement such behaviour is that different user settings and user behaviour have large impact to the power consumption and – as a result – to the runtime. E.g. disabling LCD will save lots of power, but some users do not want to disable the LCD to keep readability in daylight.
This patch therefor adapts the estimation of runcurrent() based upon real measurements of the current consumption. The algorithm starts with a default (CURRENT_NORMAL which is defined in the config-file of each player) and uses a weighting scheme with a large time constant to adapt the estimation of the current consumption. The time constant is 1024, which should equal ~20 minutes as the runcurrent()-function is called once per second. With a smaller time constant the runtime estimation is “pumping” too much when the HDD is active.

This first version is just a hard coded experimental solution. Nevertheless something like an actual_current()-function can be implemented for each player in its own powermgmt-file. Players with capability to read the current will return this measured real current, all others still return the fixed default CURRENT_NORMAL.

This small patch adds a “jump to time” function to wps context menu. After clicking on jump 2 time item user, can specify time in mmmss format E.G:
- 130: jump to 1:30 in currently playing track
- 90: jump to 1:30 in currently playing track (shorter version of previous example)
- 1529: jump to 15:29
- 12345: jump to 123:45 (2 hours, 3 mins, 45 seconds)

I use it to easily transfer “last played position” time from computer to rockbox. Can be useful for audio books listeners.

during my research on FS#10883 i found out that it would be better do disable slide actions on ipods for yes no decisons. slides are too easy done and often cause an unwanted abort.

this patch solves this issue by only accepting pressed keys for abort actions. the solution for FS#10883 has been appended on the bug report seperately.

There should be another option on the list of prioritized options for where to place Album Art images and the filenames. There are currently 7:
http://www.rockbox.org/wiki/AlbumArt

My suggested addition would become the new option 2, pushing all others down.
The first option is currently the only option to allow for song-specific artwork to be displayed, but this requires naming the image files *exactly* like the corresponding sound file, which screws up directory sorting.

Therefore the option should be added that if no image file has an identical name to the sound file playing, it then looks for an image file which matches the current song's Title metadata, similarly to how the following option looks for a match to the Album metadata.

This could cause conflicts when two songs have the same name, or when a song shares the album's name, or the song name contains unusable characters, but such conflicts are rare and minimal, and if anyone cares all that much about such things there is already the built-in flexibility to get around such issues. The additional code required should be minimal (and already blueprinted within the other options), and have essentially zero impact on any other functionality.

I may try to write this patch myself (as i did with "shuffle + track skip"), but if anyone beats me to it, it would be appreciated :)

This patch adds to the simulator the ability to simulate usb on a Linux host.
To do this, you'll need to download, build and install a device driver. You can download it here:
http://sourceforge.net/projects/usb-vhci/files/ Just download the "linux kernel module". See install instruction in it but it's make && make install. Of course you'll need root rights. There also a wiki about it (http://sourceforge.net/apps/mediawiki/usb-vhci/index.php?title=Main_Page)

Now to have usb simulation works correctly you'll of course need to load the driver
modprobe vhci-hcd
And you'll probably need to edit the access rights of the device control device:
violent example: chmod 0666 /dev/vhci-ctrl
other example: chown root:<myself> /dev/vhci-ctrl
Or you can add a udev rule if you can make this work.

You can then run the simulator normally (note the the detection is at runtime so you don't need to have it work before launching the simulator).
Press 'u' to simulate usb insertion and 'u' once more the simulate usb removal.

Some miscellaneous remarks:
-the patch enables usb hid but for some reasons it doesn't work in main menu. However, the remote_control plugins works
-the patch enables usbserial which of no interest here but a way to test bulk transfers
-the patch DOESN'T enable usbstorage because storage simulation is a stub currently.

If it doesn't work, you can still uncomment this line in usb-drc.c:
/*#define USB_VHCI_DRIVER_DEBUG*/
And this line in usb-core.c:
/*#define LOGF_ENABLE*/
But beware ! Your console will end up full of obscure messages.

I'm particulary interested in compile feedback under Windows and OSX.
And of course I'm interested in working feedback ranging from device detection to working state (HID and usbserial) for example.

This patch enables changing the fractal's colors.

Currently, not all targets have proper keymaps defined. This is a work in progress.

I would appreciate any comments.

General:
The trainer works with (three) cardboxes. Known cards will be moved from left to right to the next box, unknown to the first. You have finished the test if all cards are in the last box. The actual filling status of the boxes is graphically shown.

How the user handles the trainer during the test:
- a word is displayed
- the user thinks about the translation
- after pressing a button the solution is displayed
- the user decides by pressing the forward or backward button if he was right or wrong
- according the selection the card is shifted to the next box or to the first

Options:
- select vocabulary file
- reset (statistic) of vacabulary file
- change direction of inquiry
- display comment on/off
- colour settings

Vocabulary files:

The vocabulary is stored in files in the following format:
word1;word2;comment (optional)

Example:
lernen;to learn;verb

File format:
Maximum length of words:50
Maximum number of lines:500
File format: UTF-8

Running “current build” r24477-100203 on Sansa e200 V1

Game Brickmania bug:
The ball just disappears (often on Level 2, but maybe on others). Ball will just usually go to the top corner and not “bounce” back, but the game acts as if a ball is still alive (doesn’t register a lost ball) though there is no further movement on screen. I just have to quit the game.

The ball doesn’t bounce back. No further ball is issued. It just disappears and I can’t interact w/ the game anymore due to no active ball.

FIRST APPEARED: This began a few weeks ago (sorry I can’t be more specific), but has persisted in several “daily build” updates I’ve done and is in the current build I just installed.

REPRODUCIBLE: I reproduce the bug each time I just play through Brickmania. It generally happens in level 2 of the game though it may happen in subsequent levels– I don’t get through to them to see.

I’ve installed with rbuitil and manually, same thing happens. I have 2 different V1 e200s and it happens on both of them.

Thanks for all your work on the great Rockbox.

This patch tries to avoid the \opt / \nopt hell in manual's tex files, by concentrating all different action definitions together in the platforms keymap file.

Current patch contains changes only to the jpegviewer plugin, in order to demonstrate functionality. Please provide feedback before I spend a lot of time converting all other locations in the manual to this proposed form.

I think this way of organizing keymap assignments is better because it is neater, and each platform has its button assignment all in one place. Adding new platform is much easier this way.

The patch also use one keymap file to all ipod brands, because they are similar to each other (the ipod4g adds some keymappings, but it doesn't matter if they are defined in other ipod brands).

This is a first attempt at creating an entry for the new fft plugin. The button table was automatically generated with the attached python script.

The script takes a file (or - for stdin), looks for an #ifdef CONFIG_KEYPAD….#endif block and if it finds one (and it is relatively simple), it parses it and creates the corresponding LaTeX table.

The script is extremely simple but it does provide a good basis for further work.

This patch which is a small change in bidi.c contributes with the diacritics patch (http://www.rockbox.org/tracker/task/10720) to fix the diacritized Arabic text joining problem

Added new menu item under Settings.General.System, to select whether USB cable connection enters charging mode by default, or data transfer mode.

By default, the option is set to "No". Meaning, in order to have RB enter charging mode, a specific button should be pressed by the user.

If the option is set to "Yes", USB connection causes RB to enter charging mode, unless the specified button is pressed, causing RB to switch into data xfer mode.

This flyspray task will hold the development of a potential usb audio driver for rockbox.
It will also serve for the development of the isochronous transfer support support in usb code.

Screenshots in the manual are nested into a 'figure' environment, but the only purpose of this is to be able to have a caption. The floating feature of 'figure' by using the placement attribute 'H' (='Here!').

This patch makes screenshots non-floating while keeping captions below the images. This is done via the LaTeX package 'caption2'.

Placement attribute for figures is set back to the LaTeX default so that figures can really float should we need it.

This patch adds a per-thread runtime variable which indicates the ticks the particular thread has been running.

This can for example help identifying threads that occupy the CPU for a long period of time.

Emuriver is an emulator (not simulator) for iriver ifp-7xx series. It’s part of the http://www.rockbox.org/wiki/IriverIfpPort and has until now been hosted at the main author’s (Tomek’s) site http://www.mimuw.edu.pl/~tmal/rockbox/ It includes a tool to create flash images for use with the emulator, and a tool to mount such images and do basic file management on them.

This version is almost identical to the last version from Tomek’s site, but I had to fix a few bugs to make it usable on my system. Changes are as follows:

flashimgtool.c: Outputs usage information on errors. (I didn’t do this; I don’t know where this code came from but it’s useful.)

hardware.c: Bug: Arguments were missing in a function call which would sometimes cause segmentation faults, and (I don’t know why or how) slow down the emulator.

main.c: Changed a default file name from prog_flash.bin to prog_flash.img (I didn’t do this)

main.c: Bug: Added explicit return value to nonvoid function. The implicit return value was sometimes -1 which indicated an error so emuriver aborted.

make_flash_img.c: Optimizations: “if” replaced with “else if”. (I didn’t do this)

uicurses.c: Rename a variable into someting more unique. (I didn’t do this)

The reason I’m submitting this patch is that I think the emulator could be useful as it is for those trying to build the ifp port from SVN, that it could (maybe) be worked on to support other players, and that my bugfixes should be included for those building emuriver.

There are a few issues: The flashimgtool subdir in emuriver duplicates a few (three years old) files from RB, disk.c, fat.c, and some more. You might not like the directory structure. It uses an ancient version of autotools. I haven’t been able to create flash images that actually work with flashimgtool. I haven’t built the RB port for ifp yet, so I don’t know if it will run in emuriver or not.

Yes I have discussed the patch in IRC.

This patch implement alternative hold mode, mentioned in FS#10698.
To enter this mode you must:
1) switch hold button to on position(this results in normal hold mode).
2) press menu button.
3) switch hold button to off position
4) release menu button
To leave alternative model:
1) switch hold button to off position
2) press menu button
3)switch hold button to on position
4)release menu button(now you in normal hold mode)

Additionally I implement simple keymap which working only in wps contex:
if hold button off:

 Plus button increase volume

New build target, for iriver E10.

A lot of target-specific functions are missing, so it doesn't compile when it reaches the linker. But I think it's about time to insert the build target into SVN, and #rockbox seems to agree.

Also, building a simulator build works quite well. The simulator BMP and the keymaps needs some work, but it works pretty well, including music playback.

I've done some testing with building for other targets and nothing seems to be broken by this patch. I couldn't figure out how to include the BMP in the patch so i attach it separately.

Initial support for showing input in the fft plugin. Now, on targets supporting recording, instead of quiting when no track is playing, fft starts monitoring the audio input (using the settings under Recording Settings).

The first patch is just the functional changes, the second patch fixes the indentation of the entire file.

I’m not sure if it works at all since I don’t have a device that can record.

Hi all,

Currently for all models, the Morse Code input method can only type lowercased letters. Here I provide a patch which enables people to switch between lower cased and upper cased letters, I implemented this for Fuze and Fuze v2, and further fixes will be provided soon (or someone else want to add some key mapping options into your keymaps?)

The patch (also attached):

Index: apps/action.h

   ACTION_KBD_DOWN,
   ACTION_KBD_MORSE_INPUT,
   ACTION_KBD_MORSE_SELECT,

   /* the following are helper actions for touchscreen targets,

   "abcdefghijklmnopqrstuvwxyz1234567890,.?-@ ";

                       state.morse_code |= 0x01;
               }
               break;

           case ACTION_KBD_SELECT:

                   /* turn off hangul input */
                   state.hangul = false;

               }

               break;

           int morse_code;

           outline[0] = morse_alphabets[i];

           sc->putsxy(x, y, outline);

           morse_code = morse_codes[i];

   { ACTION_KBD_DONE,         BUTTON_UP,                        BUTTON_NONE },
   { ACTION_KBD_ABORT,        BUTTON_HOME|BUTTON_REPEAT,        BUTTON_NONE },
   { ACTION_KBD_MORSE_SELECT, BUTTON_SELECT|BUTTON_REL,         BUTTON_NONE },

   LAST_ITEM_IN_LIST

This patch modifies hardware registers for pclk & fclk in one instruction.

I’m currently testing it on Clipv2 to see if it gets rid of crashes

This is a patch meant for Clip+ that adds a “Reflection” setting to
allows choosing if to show a reflection or not. When no reflection is shown the images are moved to bottom of screen allowing for title at top and keeping them in the blue
will work for other devices. I tested it on Fuze v2 but others may have unknown problems related to screen size differences. For screens > 100 high it doesn’t move the albums
against r26185 in UISim
you want this as a patch against my pictureflow-wps integration patch let me know as I have one

This very small mod changes the tracklist so it doesn't fade lines when on a mono display (bit depth 1). Tested against r26241 in Sim only (because I don't have a mono display device).
Please report back if problems.

This patch adds a couple more lines to Track Info for Play Count and Rating.
Currently these don’t seem to show up anywhere else.

Tested against r26427 with Fuze v2 and USim.

preamble first :) I am new at the rockbox-project, so this is my first commit. I hope that I am right here with that. If not, please notify, and say, where somewhat would be better.
Strictly speaking this is my first contribition to an opensource-project ever. I am a friend of OSS since years and now I want to give something back ;) I am thinking, rockbox is a nice project to get in the open source scene.

To get in the rockbox-project I tried to improve the jewels-game. First I changed the controls for the Sansa Fuse in the Simulator. The old controls are good for the real Fuse, but very unconfortable at simulator.

Then I wanted jewels to get more difficult with higher levels. To do so I changed it to start with 4 types of jewels only and get one more each three levels till they are 7 types.

Greetings,
Michael

This adds the translatable string LANG_DISC for display in WPS themes. E.g. a theme could use the following to show “Disc 2” for a multidisc album:

%?ik<%Sx|Disc| %ik|>

Disc is a 4 or 5 letter word in most languages, so it won’t add much clutter in themes which would like to make use of this. A comparable example is the “Next:” (LANG_NEXT) which is used when displaying the next track in a playlist.

This is a beginning of support for line out and dock connection detection for the e200v2 (and possibly fuze).

The first patch enables line out in as3514.c. Line out is always enabled at this point, so battery life may suffer. Line out volume is set to 0dB and does not change unless the volume is set below -40dB at which point the DAC volume begins to decrease. This patch does not support as3525v2 at this time.

The second patch detects a connection to a dock on e200v2 (and possibly Fuze) and prevents switching to USB mode and rebooting to the OF. For this to work properly, you need to have Rockbox already booted. There is no need to press any button on connection. If a normal USB cable is attached, the OF will be booted.

TODO:
1. Only enable line out when connected to a dock and disable line out when the dock is disconnected.
2. Provide support for additional as3525 and as3525v2 devices. Information on IO port changes when connecting a dock is needed for this.

Steps to reproduce:

1. Create a directory: /AUDIO;BOOKS
2. Place a file in it: /AUDIO;BOOKS/file.mp3
3. Start playing the file, and then create a bookmark
4. Do “List bookmarks” for the file
5. The bookmark menu lists something like this:

BOOKS/;file
0:02, 2, Shuffle

The problem seems to be that .bmark files use semicolons to separate the fields of information. If there’s a semicolon in the path, it thinks that it indicates the end of the path and the beginning of the filename.

This patch adds a sector-based read/write cache between the storage and fat drivers.
It has a static pool of 64 sector buffers which can be locked by the file handling code to replace the sector buffers in the file and dir structs, and serve as cache as long as they aren’t used. It can also be used to enforce exclusive access to a sector.

If requested using the STORAGE_NEEDS_ALIGN define, all storage accesses will be aligned to cacheline boundaries.

This patch wil allow the use of the power key to respond yes in the yes&no screen, cause if in absolute touch screen mode would be impossible to respond yes.
It also will enable the hotkey option.
Italian and english translation for the yes&no.

So, when you compile and run the bootloader on VX767/VX757 or any other buttons-only model, it can only boot to the Rockbox or the OF. It cannot use USB mode or reset RB configuration.

Here is a patch.

Some collections (such as ftp://sunsite.unc.edu/pub/Linux/games/amusements/life/dblifelib-2.tgz) use an extension of “.l” instead of “.cells” for what appears to be the same file format (same comment character and same characters for dead and live cell).

This patch just adds an entry to viewers.config so that Rockbox will treat .l files as supported and open them with Rocklife.

This patch adds an additional menu option to the list of WPS hotkey choices, “Announce Information”. This voices basic information about the player, currently playing file and current tracklist.

This feature reads a simple definition with a single line of no more than 20 characters from the file “announce.cfg” in the .rockbox directory.

This definition consists of a string of single character tokens, each of which announces a particular canned piece of information.

The tokens are:

A - current time
B - elapsed time and total length of the current track
C - current track number and total tracks in playlist
D - battery level as a percentage
E - battery remaining in hours and minutes
F - sleep timer remaining in minutes

If more than one token is specified, more than one piece of information is announced, so:

AB

would announce the time followed by elapsed time and track length

Multiple announcements may be bound to the hotkey by separating one or more groups of tokens with a “:”

A:DF

In this example, the first time the hotkey is pressed, the time is announced. If it is pressed again within 10 seconds, the battery level and sleep timer are announced and so on.

A space between tokens adds a short pause to the announcement.

If no announce.cfg file is present, the time is announced by default.

This patch enables display vsync on the iPod Nano 2G, currently only for the Type 1 (LDS176) display. It removes tearing and can make animation smoother. However, right now it uses an extremely CPU-unfriendly method of waiting for synchronization, which makes the display quite slow in some circumstances. I hope to implement DMA to copy the framebuffer, which should help somewhat. However, I am probably going to need some help to do this.

TheSeven gets all the credit for this patch, since he found where the Tearing Effect (vblank) signal was connected to the s5l8701 and wrote the vsync code. All I did was drop it into the display driver in the right spot.

This patch contains several improvements to the Chopper game designed to make it run with a high framerate instead of a low, jerky one. Previously, the game depended on sleeping for a number cycles before drawing another frame, causing very jerky motion. Now the game should run at a high framerate all the time. The gameplay should be about the same, at least on targets with a display of >64px. I don’t have any of these smaller targets, so I wasn’t able to test. Please tell me what you think of my changes.

This patch reimplements FS#11541 using the skin engine.

new skin tag %Vx(params) will talk whatever is in the params (unlimited….)..
Params can be any string or other skin tag (no conditionals just yet).
if the first option is - it will shutup anything currently talking. otherwise - is a voiced pause.
otherwise the only special string is the word “text” which says to spell the next tag (used when the value of the tag is usable in a conditional, i.e %cP need talk or you will get “1” or “2” :/ ).

quick examples:
%?mh<%Vx(-, %cl,-, %cM, text, %cP)> ← Talk the current hour, minute, am/pm when hold is toggled.

%?pE(15)<%?It<%Tx(-, %Ia, -, %It)» ← untested but should spell the next tracks title and artist when there is 15s left of the current track

This patch adds an option to allow the Fastforward/Rewind buttons to be used to skip tracks entirely.

This is an itch of mine that I scratched and am offering up to others if you are interested.

I’ve been using audio lessons to learn Japanese and ran into issues navigating with the normal FF/RW buttons. The acceleration based on track length left me with no sense of time since I don’t keep track of how far I am into my 30 minute long lessons. Simply backing up 5 seconds to catch something I missed was a problem because I’d have to look at what I was doing to see how far I’d ff/rw. Kinda hard when one tick moves 30 seconds too. This is problematic when driving! The existing skip-length feature is really good for this. However, if I use that to skip 5 second increments then I abandon any way of skipping entire tracks. So I re-purposed the FF/RW buttons for skipping tracks and use skip-length for navigating within a track.

I’ve been using this in r27841 for about a week on my Gigabeat, but it seems like it should work for r27942 as well.

This patchs modifies the simulator so that it can emulate the storage layer. More preciserly, it enables all the storage code and simply redirect all storage_{read,write, …} to a file. This way, all fs functions used are the one of rockbox, which allows to debug the FAT code for example.

To run this, apply the patch and modify the makefile to add this: -DROCKBOX_SIMULATE_STORAGE_LAYER
to the line:
export GCCOPTS=….

When running the simulator, it expects to read from file simdisk.raw. You can change this by using the –root option:
./rockboxui –root <file>

The code doesn’t work for me, FAT code seems to choke on something for now, I’m not sure why.

Overview:

This adds the hotkey [short home], and allows you to stop/start play more or less anywhere [long home] which I find quite useful. Hopefully, it’s a little more consistent too, though it still needs a bit of polish.

Problems:

I’d quite like [short power] to stop play in the menus by calling ACTION_TREE_STOP, but unfortunately that only works for local files, not for the radio, which feels inconsistent. I’ve put in a feature request; http://forums.rockbox.org/index.php?topic=25657.0

The recording screen isn’t very consistent with the rest of the keymap - you can press [left] or [home] to get back to the menu, and [short power] and [right] both open the recording menu. This seems to be a peculiarity of the way the recording screen is designed, though I’ll try to improve it.

There is a ‘glitch’ in that when you press [long home] deep in the menus, it’ll sometimes jump through another menu before dropping back to the WPS / Radio screen.

Some users prefer to have a free button with which to turn the backlight back on - they are seemingly used to using [short down], which is bound to playlist here. However, if they don’t set the WPS hotkey, they’ll still have a ‘functionless’ button, so no functionality is lost.

I haven’t touched the HID part at all, because I have no clue what it does…

Details of major changes:
WPS:
[Long up] Pitch screen
[Short down] Playlist screen
[Long down] Quick screen
[Short home] Hotkey
[Long home] Stop

Pitch screen:
[Short select] Cycle mode
[Long select] Reset

Recording screen: (troublesome)
[Short select] Start / pause
[Short left] / [Short home] New file / exit
[Short right] / [Short power] Recording menu
[Volume] Adjust settings

Radio screen:
[Short select] Scan / search
[Long select] Presets
[Short down] Radio menu

Built against r2799. All comments and suggestions welcome, and thanks for reading.

This patch converts LCD_{WIDTH,HEIGHT} into variables, allowing Rockbox to change screen size at runtime.

I’ve tried to make the changes as less intrusive as possible, but still, there will be some binsize increases.

Most of the #if’s were converted to C if()’s which should be functionally equivalent (and if the compiler is smart enough, no increase in binsize should occur).

In order to use this, just define DYNAMIC_LCD_SIZE in your target config and compile without plugins (if the preprocessor spits errors, try clearing ENABLEDPLUGINS in your Makefile).

What still needs to be done:
* figure a way out how to compile several rockbox logos into the binary and choose the correct one at runtime
* do something about the hacks in apps/bitmaps/SOURCES and apps/settings_list.c
* get the Android port working

This patch should save pitch and speed settings to nvram, and restore them after reboot if needed.

If I read the code correctly, there should be enough room in nvram (and anyway, the way this is done doesn't apply to any targets that actually use real nvram)

This code compiles, but apart from that it's completely untested.

Of course, since this patch increments the nvram data version, you'll lose resume info the first time you reboot.

The Rockbox utility generates its talk files and directory files using the filename and directory names.

If the filenames do not contain spaces, but are CamelCased, they are currently synthesized badly, for example: “I Do I Do I Do” gets pronounced eeedoyeeedoyeeedoy.

The patch to the rbutil/base/talkfile.cpp processes CamelCased filenames and directory names just before synthesis and produces much better sounding talk clips.

This patch introduces 2 new activities, one which lists available themes for a specified target and another which lists more details and the option to install the specific theme. Currently this is more a proof-of-concept, as the communication with the theme server is handled by the same API RBUtil uses, which isn't perfect for the intended usage (e.g. offers no possibility to filter by screen size, only uses INI as output format, ..).

(this also includes some minor unrelated changes to the main Rockbox extraction functionality which won't be in the final patch)

TODO:
- make translatable
- better UI
- uninstall/update functionality
- show only themes fit for the displays' resolution
- use different protocol to communicate with theme server to get rid of iniparser dependency (XML, JSON, .. ?)