Building Rockbox fails if assert() (in firmware/include/assert.h) is used.

Then, I reworked assert.h.

changes:
When assert() failed, the correction that displayed the file name, the number of lines, and the condition of failing was done.
Moreover, when the log was enabled, the log was output to the file.

Please select the following to make assert() enabled when you make Rockbox.

1) Run tools/configure

2) Build (N)ormal, (A)dvanced, (S)imulator, (B)ootloader, …

    Select "A"

3) (D)EBUG, (L)ogf, (A)ssert, (S)imulator, (P)rofile, (V)oice

   Select "A"

4) make

I was reading http://forums.rockbox.org/index.php?topic=17079.0 and thought I’d take a swing at it. So far, I have added (potentially shaky?) support for the ‘#’ comments in the file posted, allowing both Windows and Linux line ending formats, and longer filename support.
Still having problems with the white space after each entry, and detecting or assuming the presence of the file extension.

This patch is a continuation of FS#9879 and includes it (but adds something more). It adds the possibility to switch the lamp on and off without leaving the plugin.

Key mappings were taken from FS#8934.

Here’s a patch to have a functional usb_serial driver allowing the use of “Logf over Usb” debug function.

Since FM stations list may require that the FM region is set to a certain value, it would be good if the FM region could be set along with the station list. The patch does just that.

The region is set by putting the line “region:N” into the FMR file (at any line there). N can be one of the values allowed for the fm_region setting.

The patch does not work on the simulators for e200 and h120 (the only two I tried) because, for some reason, the symbol HAVE_RADIO_REGION is not defined for them. Could anybody tell why?

Saving the region along with the station list is not implemented.

This plugin searches in the .rockbox/rocks/apps/drumkit directory for .raw sound files, loads them into memory, and assigns the keys to play them, creating a drumkit /sample trigger / instrument out of the player. Currently only designed for the Sansa e200.

Room for improvement:

- Add keymaps for all
Allow multiple samples to be played at the same time (by using
Allow for a variety of sound file types to be
Find a better name

This patch enables you to compile the Rockbox simulator with LLVM (you still need to alter the Makefile yourself though).

However, it fails at linking stage with 'file not recognized: File truncated' errors.

This bug currently exists as of rev 21444.

If settings are modified on the recording screen they aren't persistent after a reboot unless another setting is changed somewhere in the menu system.

The problem is that global_settings is being modified without a subsequent call to settings_save().

Attached is a fix for this problem. I've so far only tested it on the h1x0 simulator but I think it's a pretty benign change, unlikely to cause problems on other targets.

As of r21894, the pitch changing mode (in the pitch screen) can only be toggled (cycled through) in one direction. However it would be handy to be able to cycle in the other direction to change procentual vs. semitone while still staying in the time stretch mode. This patch adds such a possibility.

The key mappings is only defined for H1xx and H3xx (as a long press of Mode). Please add keys for other platfroms.

I tried to make the first shot on adding fm radio support for the Samsung YH-* devices.

–

The patch fix the multiple-language feature, maintaining the same behaviour in interactive and batch mode and with “make reconf”. Now it’s possible to build more than one voice language at once, specifying different engine and option set for each one.
So, running “make voice”, all the selected languages will be built.

To help the comprehension, here it is a batch mode sample:

../tools/configure –target=ipodvideo –ram=64 –type=av –language=italiano,espanol –tts=f,e –voice=lp_diphone –ttsopts=, –encopts=’-q 10 -c 10,-q 5 -c 5’

This is like to run:

../tools/configure –target=ipodvideo –ram=64 –type=av –language=italiano –tts=f –voice=lp_diphone –ttsopts=’ ' –encopts=’-q 10 -c 10’

../tools/configure –target=ipodvideo –ram=64 –type=av –language=espanol –tts=e –voice=lp_diphone –ttsopts=’ ' –encopts=’-q 5 -c 5’

In the first case, I set the language to italian, using festival with voice lp_diphone (female), no further festival options and some manual encoder options.
In the second case, I set the language to spanish, using espeak. Even if I set the voice, the value is ignored, since the selected TTS engine doesn’t use it.

I also added some cosmetics, like some very circumstantial indentation, blank check in rootdir too, simpler numbered menu generation, some terser command pipes.

The patch is applied to r22250.

I don’t know if it’s a really wanted feature, but the word in the script hurt me!

Added seeking by holdplus/holdminus buttons, pause/play by holdmenu button, hiding the OSD panel by briefly pressing power switch.

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.

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.

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 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.

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 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.

at least in the 16bit lcd drivers, all functions which use transparency (or clear pixels) get the pixel colour from the backdrop image, completly ignoring the bg_pattern.

This isnt a problem except in skins where it is perfectly reasonable to want a background image for most of the screen but a solid colour for one or two viewports.

There is a tiny problem though, bg_pattern is set to LCD_DEFAULT_BG which is a valid colour, so we probably need to add a flag to the viewport struct to say if it should ignore the background image or not.

This patch adds a frequency response graph below the equalizer's controls, making it an EQ that is actually "graphical".

I'm still not sure about the accuracy of the represented curve, in particular the Q factor on the peaking filters seems to have an effect opposite to the expected, though the equations are taken from the same "audio EQ cookbook" the equalizer is based on, any help is appreciated!

The user interface is still incomplete, I want to add proper tags on the axes and bars. I also thought about re-enabling the theme on this screen (why is it disabled?) and change the behavior of the "right" button, making it consistent with other menus.

Tested on a Sansa e200.

While the database initialisation is running, the tagcache queue can overflow due to system broadcast messages. This can be seen by running a simulator (it will crash with KERNEL_ASSERT after a while) or on a target by enabling KERNEL_OBJECT_CHECKS, then initialising the database and repeatedly connecting/disconnecting a charger.

The same problem may well happen with other queues (dircache, for example).

Bomberman like game with ai.
Not finished yet - lots of bugs and things to be done are known :)
Tested on Cowon D2 and also on some simulators with the same screen size (will work on other screen sizes too, but not now).
Somebody test it on iPod Video and on Zen Vision:M please.
Any suggestions greatly appreciated!
Hope the patch will work…

These experimental patches dynamically add new tracks to the database when they are played. This allows users who (for various reasons [1]) prefer to not fully initialize the database to take advantage of database-based features such as autoresume (remember a resume position for each file), runtime statistics, or track ratings.

The current patch set is only a prototype. It's still somewhat inefficient because it does a full database commit and reload for each new track that is played.

The actual database-extension patch (patch 0002) works by refactoring tagcache_build to create a new interface, tagcache_add_file, that adds a single file to the database. This new function runs in the database thread and takes a dircache_entry pointer, which is provided by way of a Q_UPDATE event. The audio thread generates this event in tagtree_buffer_event when the initial DB lookup fails, and later checks the database again in tagtree_track_finish_event. If the DB commit and reload have succeeded in the meantime, this second lookup will succeed and the runtime information the user is interested in (stats, resume info) can be stored in the DB.

Patch 0001 changes the DB commit's buffer-allocation preference from using the dircache buffer to using the RAM DB's buffer. (This used to be the preference until r10003 reversed it, supposedly because the dircache buffer typically is bigger. I don't know yet how valid this argument is today.) This reversal avoids needlessly flushing and refilling the dircache and makes the Q_UPDATE event-passing mechanism more robust (because in-flight update events contain dircache_entry pointers, which would have to be invalidated when the dircache is flushed – which is not done yet).

Current limitations/caveats:
* Requires dircache and Load-DB-to-RAM enabled (and only works on targets that include these features)
* tagcache_add_file does not respect any database.[un]ignore files in the directory hierarchy above the added new file.
* DB needs to be initialized to begin with. There's no UI function for a basic DB initialization (without scanning and adding all files) yet.
* Not very thoroughly tested (mostly in simulator, with small file sets)

Possible improvements:
* Delay committing DB changes if more track-add events are in the tagcache queue
* Incremental commit: Could be optimized for add-one-track case
* Incremental load_tagcache: Avoid flushing and reloading full database
* Add function to initialize an empty DB

Comments welcome!

Footnote:
[1] Reasons for not using the database, or not keeping it up to date, include: metadata parser instability; prefer not keep tags up to date; prefer file browser over database browser

This patch allows the line out on the Sansa Fuze V2 to be enable/disabled by means of the existing Lineout poweroff setting (previously used on some iPods).

By assigning this setting to the quickscreen I can switch between headphones and lineout after plugging into the dock. Still no luck in finding out how to autodetect the dock.

I'm not 100% sure that re-using the "Lineout Poweroff" setting is the right thing to do here as it's not powering on/off the line-out - just switching between headphones and line out

Hacking from Antoine Cellerier’s code for solitaire.c, I put together a rough freecell game.

It currently works for Sansa E200 targets. Others should work once button mappings are done.

This is not heavily tested, so feel free to report or fix any problems you encounter.

The patch adds two new WPS tags:

1. %Ln yields the number of the list item currently being drawn

2. The operator "/?" (which can be used in the %if tag) yields the string "div" if the first operant is divisible by the second and NULL otherwise.

With these two additions it should be possible to create themes where list items are coloured alternately (as an example).

The whole work has been done by gevaerts on 2011-09-12 (and all the credits also go to him), I only placed the patch to the patch tracker, changed the operator from "/" to "/?", and changed the function name from "skinlist_get_item" to "skinlist_get_item_number".

He posted the patch on the binpaste and it would be pity if it got lost.

Update to my tools/talkclips.py script. Minor updates which include:
- Better handling of logfile eg. we don't leave it open as was the case
before
- Report a few more details in the log eg. start/end time, duration, and
number of talk clips created. This is useful to dettect how many files
had problems…
- Summarise output for user at end, of processing, number of talkclips
created and how long this took (in seconds).
- Report to user how many talk clips will need to be created to give
them an indication of potential waiting time etc. This is calculated by
first determining how many files are on the device excluding .talk files, and how many .talk
files there are, and calculating the difference.
- This information allows a percentage indication to be drawn throughout
the script, so now the user can track the percentage as processing takes
place.

Patch is attached.

Thanks.

This patch adds a %ch tag that returns the number of channels in the current file.

Unfortunately, the large majority of codecs don't fill in this field in struct mp3entry. In the codecs covered by the rockbox testfiles collection, only atrac, cook and tta actually work.

I don't think it makes much sense to commit this with so few supported codecs, so help is appreciated on the codec side of things

Rockbox WebOS app.

Built on r31467, 32/64-bit Ubuntu 11.10, running on a 32GB Touchpad

Directions (only tested on the above):

* Install WebOS PDK: https://developer.palm.com/content/resources/develop/sdk_pdk_download.html

* Install platform-native libsdl-dev if you haven't already. Only needed to pass configuration check as the PDK doesn't properly install staging sdl-config for some reason.

* Apply patch (patch -p0 < filename.patch)

* cd webos

* ../tools/configure, select WebOS target (206)

* make && make install && sh ./buildpkg.sh

Notes:

* This is one big hack – I fudged just about everything in /firmware (how the target system works is still not altogether clear to me) and piggybacked a lot of this off of the SDL target. Mostly in the interest of getting it to run without adding ifdefs all over the place. Obviously we'll want to be more methodical and thorough about exactly which changes need to go where moving forward.

* The Touchpad's SDL video implementation appears to adjust orientation and scale output to fit within 1024x768 as best as possible. This means that a 800x480 build will appear in landscape mode, letterboxed, while a 320x480 will appear in portrait mode, pillarboxed. The device itself automatically orients to landscape – meaning, if you're in a portrait resolution and then you exit, everything will initially be sideways – but that doesn't affect input. Scaling routines are noticeably slow.

* I tried building this as a hybrid Enyo/PDK (window mgr friendly) app but had some problems with the native part failing to execute at SDL_Init. So for now no backgrounding or notification bar.

* Make scripts use the ARM compilers installed with the PDK. I'm wondering though if this is superfluous given that the Rockbox toolchain also supports ARM7. I did try exposing some of the ARM-specific assembly but the compilation failed with the PDK compiler.

800x480 build: http://www.jasonyu.net/webos/org.rockbox.raap_0.0.1_all.ipk

I noticed that the playlist seems to become one item shorter when I shuffle it, using the latest Git master branch (commit 70d57080) on a Sansa Clip+.

Steps to reproduce:
1. Clear the playlist by pressing Stop (power button)
2. Add a small number of tracks to the playlist
3. Shuffle the playlist (playlist view, menu, Shuffle)
4. Observe that playlist is one item shorter than before

I tracked the reason for this down, and attach a patch. This isn't a "proper" fix, and is a bit hacky, but it does illustrate what causes the problem.

I would like to run my iPod with its keypad locked all the time, thus I wrote a little patch to an option to disable the settings reset during startup regardless of the hold button's position.

You will find the patch attached to this message. I'm not entirely convinced by the name of the new option (or my wording in general) so if you have a better idea, feel free to rename things.

Holding the Pause button (to move/paint down) will shut down rockbox.

This shutdown combination should be disabled for Rockpaint!

While I’m at it (and you can put this in a different Flyspray if you want), I suggest that we make use of the iPod scrollwheel somehow in Rockpaint, maybe to change the size or shape of the brush or something and/or in the menus.

I would also like to be able to set the colour of the background to start with.

When utilizing the database on the iPod Photo (as well as other 4G iPods I have used), on occasion when entering a directory or going to the parent directory, any button on the iPod will need to be pressed before the others will have effect. I.E. before I can scroll to select a song, I must press Select or one of the other wheel buttons. It is severely annoying, in part because it is too sporadic to even be assumed to be a “feature”.

The (software-)equalizer is adding high noises to the playback.

I made a recording, the first part with eq enabled the second one with disabled equalizer.

I could also make a test on my older H110, on which I never used the sw-eq though.

also category:
Some containers, like Matroska MKA or like Ogg can contain numerous audio tracks, for example the whole AudioCD contained within single Ogg file.

For example VideoLAN.org plays such files ok, while fails to show separate tracks/tags in playlist.

However, RockBox justs freezes and until reboot can no longer play music (each file is detected as having zero duration)

example can be found: [link deleted - it was a full copyrighted album]

While examining disk performance with  FS#9708 , I found one album where the disk stayed spinning with no activity for 5 s longer when using PIO. This is repeatable with that particular album. With the downloaded r19739-090109 build, this did not happen with the default settings but it started happening as soon as I enabled dircache. Prolonging the disk timeout prolongs the extra sleeping time.

Experimentation with shows that the storage_sleep() part of fill_buffer() (in apps/buffering.c) is executed early in the buffering and it is not executed at the end. With another album where spindown happens properly, ata_sleep() is called twice. Note that ata_sleep() just sets last_disk_activity, and sleep happens over half a second later if there is no disk activity during that period and last_user_activity has timed out.

I think this is a race condition based on how dircache and transfer rates change things. I wonder if it is related to  FS#9541 .

The data mentioned is located at http://spreadsheets.google.com/ccc?key=pkeRcfM0sg949P3a5EYXtew and the delayed spindown is seen in the EP7 PIO test. In that data, also note how sectors written are 5, 10 or 15, with differences between different tests with the same album.

I have a 30GB 5th generation iPod, but I expect this issue affects all SW-codec players.

Sometimes when I shut my unit off and then turn it on again and resume playback, RB resumes from a different place than where it left off, and even though when I resume shuffle is on (as indicated by its icon), it isn’t. It’ll play the current track, then the next track in the playlist (which in my case is in order by track number, so it’ll play the next track on the album), and so on.

If I turn shuffle off and then on again, shuffle works correctly again.

I’m using a build from January 19 (19800-090119) on an Archos Recorder. I
noticed this bug occurring in the 3.1 build as well.

It seems to happen randomly – I can’t figure out anything specifically that I can do to always make it happen.

The H10 5G MTP needs a different bootloader than the UMS version (see http://www.rockbox.org/twiki/bin/view/Main/IriverH10PortDevInfo#Build_a_Rockbox_Bootloader).
I’m not sure about the correct way to do this in configure, so I’ll leave it to someone who is.

After releasing hold after several minutes of playback with hold enabled, the backlight flickers on and off, if the following settings are enabled:

Backlight on hold: off
Backlight fade in: on
Backlight fade out: on
Caption backlight: on

My theory is that fade-in/out events from the caption backlight somehow get queued up, and happen once hold is released.

Happened as recently as today, with r21438. Reports in IRC said it appeared somewhere between 3.0 and 3.1.

The uisimulator cannot be build by using the source of the latest version on MinGW.

I send this bug’s patch.

It is necessary to apply the patch of FS#10021 to build the firmware on MinGW.

(Replacing  FS#10527 )
Peakmeter on ipod video 5g uses too much CPU, such that audible glitches occur during even flac/wav decoding .

to reproduce:

1. start a big (20mb or more) track. delete it immidiately after starting it.
2. let the song play until it stops. i guess it stops at the end of the buffer. as an alternative you can ff to the end of the track (which shouldn't be buffered)
3. now you're not able to play any track. if you select a new one from the file browser or database, the wps will show up, but nothing is played and the progress bar won't progress. ff and rw also don 't work. only rebooting helps. shutting down needs really long (~15s).

if you delete a fully buffered song, everything works as expected (well, the song continues till the end, and the next song starts. some devs call that "expected" :D ).

tested with r24091 and an e200, although i guess that the bug is target independent and occurs in the last few thousand revisions :)

This is actually release 3.5.1

I have quite a few low bitrate .wma audio files that play fine in the original firmware, play fine in VLC, but skip, stutter, and hiss in Rockbox.

The file is 20kbps, 22 kHz, mono, Windows Media Audio V2.

I have attached a sample cut from a larger file. Like the original files, it plays fine in the original firmware and in VLC.

I've been working on PictureFlow - WPS Integration. I repeatedly had cases where showing a tracklist or playing for certain albums would cause a segfault. I tracked this down to a bug in the track list indexing. It only occurs under certain memory allocation circumstances that doesn't occur as readily with the std build. I have attached here a patch to fix it.

Explanation: the create_track_index requests memory for building a track index. The variable "tracks" references an array descending from the buffer top end. When a reallocation occurs it currently only moves the "tracks" position when a track count > 0. However, in some cases a reallocation occurs before the first track is stored and in that case the "tracks" position gets left in the middle of the track_names data causing corruption and then segfault. I've moved the (track_count > 0) decision so that it only controls moving tracks struct data. The "tracks" offset must happen even when track_count is 0.

This patch fixes the segfault on my Sansa Fuze v2 and in UISim. Tested against r26114.
I would expect the same result in other platforms.

I've also included this bug fix in my PictureFlow - WPS Integration patch until this bug fix gets accepted.

When the option “Auto-change directory” is enabled, my Ondio crashes at the end of the current folder’s playlist when it would have to create a new one.

The error it shows is “I09: CPUAdrEr at 09010FF8”. The address varies a bit with the actual revision, this example is with r26246 and an OndioFM backlight build (but the problem is present in plain SVN builds too). Looking at the map file shows that the crash happens in the cue_find_current_track function.

Bisecting build turned up that the error was introduced with the cuesheet rework in r21978/9 and the fix in 21982 fixed usual folder playback but not auto-change directory. It is reproducible on the Archos Recorder so I assume a general hwcodec problem.

Hello.

I found that bookmarks which was automaticaly created (by the function Bookmark on Stop with set to Ask) from tagcache don’t work. Problem is in missing path to file in bmark file. I made more test and now I attach bmark where first line/bookmark is from tagcache and second line/bookmark is from files browser.
Both are on approximately same time.

I testted on Sansa c250 and Sansa Fuse2 with Rockbox r27481-100718.

Can anybody from developers focused on it, please.

Much more thanks.
Salaba

In my customized rockbox theme, there is a short (half second?) but irritating delay before the progressbar is drawn. This delay is apparent when pressing a button to turn on the backlight, so it doesn’t show up in the simulator. Here is a video of the issue happening: http://www.youtube.com/watch?v=1LAfW8-BjjE

I’ve tried moving the progressbar to the very top of the wps. I also tried changing the wps around to have each icon in a seperate viewport which didn’t help.

I’m using rockbox version 27667.

This bug already stands for a while and was reported in IRC but since nothing happened so far, I’m going to report it here.

On targets with early USB (”usual” looking USB screen when plugging from off-state and probably only hardware bridge USB) the statusbar that should be there in the screen is missing and when you disconnect the player so that the “rest” of Rockbox would start, it crashes.

The address where the crash happens with the build currently flashed to my Ondio is in “lib/libskin_parser.a(skin_parser.o)” - more exactly “9035954: 8b fb bf 903594e <_skin_parse+0×152>”. I pastebinned the complete disassembly of the _skin_parse function here: http://pastebin.com/LZMNW96a .

I will also try to find the exact revision that introduced the bug but that’ll take a while. If I remember correctly it came with some statusbar skinning commit which is already a while ago now and crept up twice in a row (first was fixed, then it came back). Plus it is not the easiest and no quick task testing flashed builds on the Ondio….

The Last.fm scrobbler lists in the scrobble log songs that ive never listened to

In my sansa Fuze v2 build 28208 when I insert some album to the playlist and after that I insert another album to the playlist I got this error massege:

Data abort
at 3001AE90
FSR 0x8
(domain 0, fault 8)
address 0xB0166244

thanks!