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

It seems that operators are discarded when used in a tagnavi.config (or tagnavi_custom.config) format line condition.

For example,

%format “fmt_abtitle” “*%s - %02d:%02d” title Lm Ls ? playcount > “0” & lastoffset == “0” %format “fmt_abtitle” “%s - %02d:%02d” title Lm Ls Now if we used “fmt_abtitle” instead of “fmt_title” in the default tagnavi file, we would expect all tracks with a playcount greater than 0 and lastoffset equal to zero to be prefixed by three asterixes (*). Instead ALL the tracks with a playcount greater than zero are prefixed, leading me to think that the & lastoffset == “0” part is discarded. If I make the following change:

%format “fmt_abtitle” “***%s - %02d:%02d” title Lm Ls ? lastoffset == “0” & playcount > “0” %format “fmt_abtitle” “%s - %02d:%02d” title Lm Ls

Now all tracks with a lastoffset equal to zero are prefixed, which again proves that the and operator was clearly ignored and unparsed.

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

I primarily use my unit for audio books while driving, but the latest build (as I write this) won't play one of the files properly. It tries to start, stuck at 0:00. Trying to seek hangs the player altogether, requiring a hard restart.

I've recreated this problem on a flaky Clip+ and an AGPTek Rocker.

The 'file' tool identifies it as: ISO Media, MP4 v2 [ISO 14496-14]

Here's what mplayer has to say (while playing it back properly):

[lavf] stream 0: audio (aac), -aid 0, -alang eng

Right now if the voice file fails to load, there’s a splash message but blind users won’t be able to see it.

(usually this happens because they upgraded rockbox without upgrading the voice file too)

It would be useful if we could embed a failsafe voice file, containing only that one prompt, so the user will know what is going on.

Just English for now, though ideally one for every language we support.

(Oh, and a pony too. Please?)

On my X3, it triggers the MIPS equivalent of a segfault. (directory has 792 entries, tried to access it with only 500 allowed)

Rather than crashing outright, I think it would be better to detect this and somehow error out gracefully?

For a couple of months my Sansa Fuze v2 has been crashing randomly when playing music. Most of the time I just get a white screen with an error message. The last message I got was :

'undefined instruction at 0006DC8C'

(I think that last number varies at each crash)

This didn't happen with v3.13, but it's been happening with every daily build I've tested IIRC.

It doesn't happen very often so it's hard to reproduce. I listen to music in public transports about twice a day and this happens not more than once a week. It seems to happen more often when the battery is low (less than 10%). It generally happens early in a playback session, when the first or second track is played.

Sometimes I don't get a white screen, but the program freezes and some parts of the interface are surrounded with white frames. I don't know if it's the same bug or if it's unrelated.

Note that I'm always listening to music on a microSD card, and I haven't tried on the main memory.

PS: I chose "Sansa Fuze+" in the OS list because I couldn't find Sansa Fuse v2.

This patch changes some of the settings of the AS3515 to highest audio quality, similar to  FS#11304  but for AMSv1.

AUDIOSET2_IBR_DAC_50 → AUDIOSET2_IBR_DAC_0

0% reduction (instead of 50%) of DAC bias. This will probably increase power consumption.

Not set AUDIOSET3_IBR_HPH.

Increases HP amp bias. Recommended by the data sheet for 16Ohm loads (and will probably reduce THD).

Would be great if these settings could be changed during audio playback to listen for any actual changes in quality :)

The attached patch implements sinc based resampling in rockbox.
I have only tested it stand-alone so far (i.e. the code outside of rockbox – I am waiting for a replacement for my broken android phone to test it there).
There is (of course) quite some room for improvement.
Things which could be improved:
- buffer handling
- calculate the correct sinc spanning the whole table for downsampling. Then the downsampling and upsampling code would be the same (no need for sinc_increment, saving some additions). Also the result would be more accurate resampling.
- Save some memory by always calculating the correct sinc and putting it into resample_data. Would need to use fp_sincos() for that.
- Save some more memory by taking advantage of sinc's symmetry, only storing half of it in the table.

I know that I'm breaking the coding-style by using more than 80 chars in some lines. I will fix that when I have time. I wanted to get this code finally out to get some feedback after it sitting for two weeks (basically unchanged) on my HD.

Trying to cross-compile GCC for the ipod with GCC 9.2 I keep running into errors:

In file included from ../../gcc-4.4.4/gcc/diagnostic.h:25,

               from ../../gcc-4.4.4/gcc/tree-mudflap.c:43:

../../gcc-4.4.4/gcc/pretty-print.h: At top level:
../../gcc-4.4.4/gcc/pretty-print.h:310:6: error: ‘cgraph_node’ is not defined as a type

310 |      ATTRIBUTE_GCC_PPDIAG(2,3);
    |      ^~~~~~~~~~~~~~~~~~~~

This would seem to be the same issue as the one described on https://gcc.gnu.org/bugzilla/show_bug.cgi?id=90677

The attached patch against GCC seems to fix the problem for me.

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.

Rockbox version 3.13
Wenn I delete a file being played, it continues to play (which is OK for me). But when I rewind in this file while playing, Rockbox hangs and does not respond to any key presses. The screen light stays on. I have to press the hard reset or wait until the battery is empty.

This also sometimes happens when I press fastforward, but it always happens when pressing rewind.

Desired behavior:
- Preferred: permit rewinding/fastforwarding (and really delete the file only after stopping the replay)
- ignore rewind/fast forward; alternatively
- skip to the next file immediately after deleting.

Thanks for reading, and please contact me if you cannot reproduce this! I have a Iriver H140 and I must say that all my recordings are in wavpack (recordings made using SPDIF-IN), so I do not know whether this also happens with other file formats and with other players.

-Stefan

Hello there,

When I reboot my Rockboxed iPod nano second generation, it changes the time +5 hours.

Does anyone know why this can be? Thanks

The remaining time estimate shown under System>Rockbox Info reports just “0h -1m”.

The battery indicator on the status line works OK… maybe it’s just a display bug?

I have written a basic flashcard program for LUA and am attaching source code and an example data file. I am using a Sansa Clip+ so YMMV. It uses ANKI format text files which are tab delimited txt. You may use UTF8 for the flashcard if you have your device set as such. To use, save the two included text files in the same directory and run. The card data file, Russian.txt, is a typically UTF8 encoded text file exported straight from ANKI with no modifications. I am using a size 8-fixed font in the screen shots.

Things I would like to see done:

1. Making the code more generic to run on all rockbox platforms.

2. If someone would tell me how to play vorbis files from LUA on a clip+, I might be able to implement playing the vocabulary words as they are displayed on the device. I am already parsing the audio file names from the database, but I need a function to play them.

3. If LUA/Rockbox supported SQLite, It might be possible to read and write native ANKI database with scheduling information. This would facilitate programming of advanced flashcard scheduling for efficient learning.

4. A menu to select different databases.

Looking forward to feedback!

Hi, thanks a lot for all of your coding efforts, support, etc. I noticed that the WPS layout changed slightly after installing the latest dev build.

===

Problem: An unexpected indent appears in the WPS after installing the latest dev build on a Sansa Clip+. Please see attachment for an example.

Steps to Reproduce:
1. Install 3.13 with Blind theme
2. Play some tracks to see WPS
3. Upgrade to latest dev build
4. Play some tracks to see WPS
5. Problem occurs here: there’s an extra indentation after the “{current track number} / {number of tracks in playlist}” text, at the bottom-left of the screen.

Player: Sansa Clip+
Rockbox version: 1879de7-140525
Binary: rockbox.sansa
Size: 582384

Additional: Possibly related to #12953

Recipe:
1) Play an album. Wait until buffering completes.
2) Skip from track 1 to track 2. No buffering occurs.
3) Skip from track 2 to track 1. No buffering occurs.
4) In the playlist viewer, select track 2. Rebuffering occurs, even though the track is buffered.

The same problem occurs in the opposite direction (e.g. after step 2, it should be possible to switch to track 1 using the playlist viewer without rebuffering).

The attached patch slightly modifies the keymap of the H300 non-LCD remote to allow access to the browser from the WPS (by use of LONG PLAY).

I know that it works ok from an H300 with a non-LCD remote, but I don’t know if I’ve broken the LCD remotes completely.

Please could users of both the H100 and H300 LCD remotes try this patch and tell me that the WPS context keymap is unaffected. For reference, you can see the intended mapping here: http://www.rockbox.org/twiki/bin/view/Main/KeymapIriverHSeries

This patch adds an option to disable the speed limiting that slows down seeking as you approach the end or beginning of a track.

It defaults to limiting enabled, to maintain compatibility. Personally, I prefer the fast seeking without the limiting.

The manual is not included, I can't compile them still…

Where possible (i.e. the buttons aren’t needed for somethinig else), the PLAY button should go straight to the WPS (resuming as required) or radio.

These are almost the case, but on the H300 there are the following exceptions:

- Rockbox Info (PLAY performs the disk free space scan, same as SELECT)
- Playlist Viewer when selected from the Main > Playlists menu (PLAY returns to the menu above, same as LEFT)
- Sleep Timer (PLAY does nothing)
- Debug Manu (PLAY enters the selected item, same as SELECT)
- Browse x in Settings, where x = Themes, Fonts or .cfg/.wps/.rwps Files (PLAY returns to the menu above, same as LEFT). (Note that the plugin browser works fine!)

Similarly, the STOP button should stop playback, which is largely ok, except:

- In the Playlist Viewer (STOP does nothing)

I don’t have any other targets handy, but feel free to add any other suspect screens here.

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).

If I set my Gigabeat F20 touchpad sensitivity to High, the touchpad and all buttons stop working - except for the long POWER action which shuts down in the normal manner.

I then need to reset settings on boot (or connect to a PC and edit them) to restore usability.

If other owners could try this out, I'd be interested - it may be that I have a stuck button under the touchpad, of course.

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 occurs on Stable 3.8.1 and all builds currently up to Tuesday, May 10, 2011 (r29850-110510).

When my iPod 3G is powered on with no song in the queue to be resumed, and the "resume Playback" item is selected, the screen flickers, the quick glimpse of the selected Now Playing theme can be seen (for me, DancePuffDuo, but also confirmed for every other one I've tried) and it returns me back to the Main Menu without any warning… just an abrupt return.

I've reinstalled stable as well as various builds leading up to the date above and all have the same effect.

The way to reproduce this bug is to Play a single song from the database that is by itself (in the artist/album category of your choosing, or, the last song in an artist/album grouping) and then, after it is done playing and returns you to the database or main menu, shut down and restart Rockbox. Select the Resume Playback item on the Main Menu. The screen should flicker and abruptly return you to the Main Menu without any warning that there is nothing to resume playing.

This bug does not reproduce with the iPod 3G simulator with 05/10/2011's build (r29850-110510).

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.

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.

Player: Sansa Clip+
FW: Rockbox 3.12, official stable build, but it happens on older versions too

While playing some FLAC file, RB skips to the next track automatically. I know there is a lot of similar bugs, but this is unique:)
Firstly, the problem is repeatable on my Clip. Skipping is always at one time, e.g. 00:19 for one file and 01:40 for other one. I think that these files are stored on corrupted part of SD card. OK, if it's corrupted then I have to buy a new one, but when I switch to original Sansa Clip+ firmware it play normally! At problematic places (00:19 or 01:40) there is a little 'click', but it continues to play. My conclusion is: if original FW handles such a files on bad memory, Rockbox should has workaround also.
Files are OK, when I copy them to internal memory, Rockbox plays continuously.
I didn't notice skipping of MP3 files on the same SD card, only with FLACs.
Computer players (e.g. AIMP) plays FLACs from this SD without any problems.

I've attached movies that presents the bug on 2 different tracks. Hope it helps.
http://youtu.be/0e4Xf8rKTh0 http://youtu.be/zl4LX4wEwlI

Rockbox Version:r16780-080324

Description:
If the “Bookmark on Stop” option is set to “Yes” and there is no free space left on the storage, rockbox freezes/takes a REALLY long time if pressing the Poweoff(=Stop) button, because it cannot save the bookmark(or has to push out the less recent?).

I wondered why rockbox crashed if doing this til I discovered this bug.

Expected behavoiur: Instantls notify user there’s no space left and don’t save the bookmark

How to reproduce:
1. Power on the player, launch rockbox. Make sure there’s enough disk space.
2. Go Settings→General Settings→Bookmarking→Bookmark on Stop→Set to “Yes” 3. Reboot, connect to PC
4. Create a text file which is big enough to consume the rest of your free disk space
5. Restart Sansa into rockbox
6. Play a track
7. Press poweroff (short)

I think there is a problem with a .ape codec ,when i play .ape files it will have a sound like radiosound.The time happen randomly with each .ape files and when i playback again it still happen at that time.Exp:ABC.ape will have have a noisy sound at 0:10 and when i play again it still happen at 0:10

I would like to change voice strings to make them sound better with a specific SAPI voice.

I don’t intend to change string in *.lang directly.
It should be done in voice.pl dinamically while building voice, shouldn’t it?

I've tried a dozen of builds around, and it seems that the new touchscreen on D2+ is not working properly - there're wild distortions (sometimes it's possible to press a button on lower bottom of the panel, but usually it's to no effect at all) at the edges of panel - i have to press almost on the edge of panel in order to tap the buttons, for instance in this (http://themes.rockbox.org/index.php?themeid=1146&target=cowond2) theme.

Device: Sansa Fuze v2 (stable port)
Rockbox Release: 3.12

Recently I have been noticing a problem with themes that allow radio station logos to be displayed on the FMS screen. This is most noticeable with the analog and Fuzemod themes. While the logo displays properly, other images on the screen do not. In the case of the analog theme, the VU meter appears corrupted and with the Fuzemod theme the battery animation, tuning bar slider and playmode indicator icons fail to appear as expected. This problem was not present when either of these themes was used with the releases before version 3.12. In fact this bug appears to have been introduced by one of the experimental daily development builds either in May or June. I have tried different sets of radio station logos and preset lists with these themes but I always get the same result when using version 3.12; everything is fine with any previous release. Since this problem occurs with two different themes along with any combination of radio station logos and preset lists, I can only conclude that it originates with the firmware.

I have attached to this report three generic radio station logos in bitmap format along with a generic preset list. You can put these into /.rockbox/fmpresets and try your own tests. Since the frequencies were selected at random, don’t expect them to tune in anything. Besides, there is nothing wrong with the ability of the Rockbox firmware to tune in radio stations or play their music. The problem starts as soon as the user starts flipping through the station presets, so don’t be shy about changing the station frequently during your tests. I also find that in order to accurately reproduce this problem the Sansa Fuze v2 should be powered off then on again whenever a new theme or preset list is loaded.

this is an attempt to implement the idea suggested in the forum (http://forums.rockbox.org/index.php?topic=20120.0).
I take values from “skip length” setting for variable steps.

This attempts to implement the idea in http://forums.rockbox.org/index.php?topic=25169.0 this only works for the Settings menu in the main menu.

With this patch, if you reenter the Settings menu, it shows the menu shown when you leave the menu just like when you return to the browser or the database.

On my newly purchased (but used) Archos Internet Tablet 70 250GB with Android Version 2.2.? (Froyo) I installed the daily build of rasher. Everything works so far, but playback of MP3 (mostly 160 Kbit CBR and ~200 Kbit VBR) pauses briefly every 2 to 3 minutes. I just reformatted the drive, which now has ext3 and that seems to end the pause which I got on FAT32 after ~4 Seconds of every new song, but every 3 minutes or so the playbacks stops for about a second - very annoying. Don't tried other codecs as I don't own any files other than MP3.
I was in the developers menu during one of this pauses, looking at the list of the contributors. The old names fly away to the right, but this was frozen also during this one second timespan - as if the rockbox-process was paused entirely or so.

Hello,

empty menu items crash rockbox due to NULL pointer access.
Easily reproducable with the "LCD settings" in RaaA / SDL app build.

The resulting valgrind trace looks like this:
==22619== Invalid read of size 4
==22619== at 0x407FDA: do_menu (menu.c:555)
==22619== by 0x41B198: miscscrn (root_menu.c:323)
==22619== by 0x41B377: load_screen (root_menu.c:520)
==22619== by 0x41B72A: root_menu (root_menu.c:688)
==22619== by 0x406DFA: main (main.c:196)
==22619== Address 0x0 is not stack'd, malloc'd or (recently) free'd

Question:
Should we protect against "menu→submenus" being NULL or
should the menu system hide empty menus without any items at all?

Cheers,
Thomas

Hi,

the random_folder_advance_config plugin fails to build the
folders on N900 / RaaA. I guess it suffers from the same
as the database update did in the past: It probably follows
starts over and over again in the /proc or /sys filesystem.

This task is to develop a stand alone command line interface for use with SAPI. This will be designed to run under Windows or WINE.

This will replace the platform dependent VB script currently in place adding new features and try to improve performance.

I have attached an initial release, run with -h option for usage info.

This is not completed had very litttle error handling and due to performance issues under wine I am considering a new server/client model for a future release.

I'm using Rockbox 3.10 with the 01.02.16 Clip+ Firmware on a Sansa Clip+ 8GB, installed via the installer.

There are a lot of 44kHz/16bit FLAC files stored on the device. During playback it could happen that the next track starts while the current one isn't even close to finished. This can be in the first few seconds or in the middle of the song.

This usually doesn't happen to the same song twice. It's generally a different song each time. When skipping back to the track that was skipped, it usually plays without any problems.

The memory has been checked with CHKDSK and no errors were found. The FLACs play fine in foobar2000 and I'm certain that they are okay.

I do believe the error occurs most often after roughly quarter to half an hour of playback.

This may be related to FS#12208

In  FS#8894  (the original timestrech task), an updated version of the algorithm was uploaded.

The author said it performs and sounds better, so we should try to incorporate it.

I've tried building it, but I can't get a useful result, but I'm also not sure if I used it correctly. I used ./scale 2 < in.wav > out.wav. The result wasn't streched, but had crackling noises every few seconds.

This adds asplash and asplashf (and vasplashf for internal use) which normal splashes but can be aborted with button presses.

Additionally, they return the action received (for CONTEXT_STD) so that the caller may handle aborts differently depending on the button.

Also I converted a few annoying splashes to use it (when loading themes, wps, fonts etc and the resume failed splash)

bsearch() is a general purpose binary search function for arrays.
It's supposedly faster than looping over arrays.
The array needs to be sorted in ascending order under the provided
comparison function. If the key and array element are of the same kind,
then the same compare function can be used for qsort() and bsearch().

Code taken from glibc.

—

So, is anyone interested?

The setting works similar to the autoresume dirs: Directories are seperated
by colon, e.g. "/Music:/Podcasts". Default is "/sdcard" on android, "/" on
all other targets.

A maximum of 8 dirs can be selected, the setting cannot be longer than 80 chars.

Note: There's one problem. If you specifiy "/Music:/" (or any folder and one if its parents afterwards), then /Music is scanned twice. It's tricky to fix, and I'm not sure if it's worth it. This does not happen for "/:/Music" as it can be detected more easily in this order (and this is already implemented).

All settings should be listed there (even more so settings which cannot be set through the UI), but root menu order is only described in advanced_topic/main.tex as of now.

The %pb tag has a small problem. If an image is specified but no height, the tag draws the size of a line even if larger than the the image. This leads to pixel garbage below the progressbar.

The test.wps reproduces the problem in the 320x480 sdl app The default font, 27-adobe, has to be selected probably. The wps works if a height, i.e. 20 in this case as the image height, is specified

Test.png shows the problem. test.zip is the small test theme.

The 2bit greyscale LCD drivers ignore the current viewport’s drawmode. This makes you unable to draw icons with a non-clear background.

I have only tested this on my H120 with an 18007 build. I have been using the APE codec for some time now and never had a problem in the past.

After reformatting my hard drive and installing the current build I found that rockbox has 2 problems with my APE files

1. The parser for APE files no longer recognizes my APE files encoded with version 8,9 of Jriver Media Center. Result is that files don't play when selected in the file browser and these files do net get added to the database.

2. When switching APE tracks one of two possible behavior occur. a) the Iriver freezes completely. b) On switching tracks there is a loud hiss and then playback of the next track resumes.

Cheers

  Thomas

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.

This is an almost complete rewrite of the current dict plugin.

It is based on the stardict dictionary file format and acts as a viewer for .idx files.

Most important features are:
- fuzzy search by showing all articles that start with a certain word
- quite fast search (binary search, same as the current dict plugin)
- scrolling for long descriptions
- handles even huge dictionaries (e.g. Wikipedia)

When first opening a dictionary, an offset table is created to provide random access to the index file.

Stardict dictionaries can be used almost directly but all files have to be uncompressed.

I created another patch, which adds a “Dicionaries” entry to the root menu (FS# 6630). It shows all dictionaries placed in “/dicts”.

Some more information may be found in a thread about an older version on MisticRiver http://www.misticriver.net/showthread.php?p=546738

Some dictionary ressources:
- Stardict Website: http://stardict.sourceforge.net/Dictionaries.php - Wikipedia dumps:

1. english: http://srv2.filepoint.de/download/1457/enwiki-20061130.7z
2. german: http://srv2.filepoint.de/download/407/dewiki.7z

On page 51, which is in Section 6.8, there is a table for keys/actions. The Key column is completely empty, though.

On page 77, in Section 11.1.1, the blackjack game plugin, there table is similarly broken. Same for BrickMania in Section 11.1.2. Chessbox, in section 11.1.4 has a similar problem with only the "direction keys" being in the table.
There are many more games with this problem.

This is for a manual I downloaded 2010-10-11.

Thank you for your attention.