Skip to content


Subversion checkout URL

You can clone with
Download ZIP

Comparing changes

Choose two branches to see what's changed or to start a new pull request. If you need to, you can also compare across forks.

Open a pull request

Create a new pull request by comparing changes across two branches. If you need to, you can also compare across forks.
base fork: xbmc/xbmc
head fork: xbmc/xbmc
Checking mergeability… Don't worry, you can still create the pull request.
This comparison is big! We're only showing the most recent 250 commits
Commits on Mar 11, 2013
@wsnipex wsnipex libdvdccs: allow use of automake >1.11, thx brianf21
this check should be reworked upstream
@phil65-2 phil65-2 Fix for AlarmClock(): Prevent freeze-up when time <=0 and loop = true e9d881b
@Voyager1 Voyager1 Fix resume playback from playlist does not work (trac 13929) 56c9973
@popcornmix popcornmix [rbp] Fix for stuttery video when seeking before zero
There are a few issues with seeking I found.
We weren't correctly setting OMX_BUFFERFLAG_TIME_UNKNOWN on the first frame after a seek which could make the GPU think video was at 0 and audio at a much larger offset.
A full video fifo (to GPU) stops any higher priority messages from being received which can stall a seek and the flush message doesn't get through. Use m_flush to discard the video packet that doesn't fit.
We get an audio frame through with unknown pts/dts after the flush, but before the GENERAL_RESYNC when seeking. This was given to GPU and was perhaps 30 seconds out from the following packets and that throws off the timing between audio and video streams. Keeping m_flush true until the GENERAL_RESYNC discards this frame. Hopefully that is safe.
@ace20022 ace20022 [PulseAEStream] Possible null pointer dereference. 8adff67
@davilla davilla fixed crash, sizeof(SDL_Event) != sizeof(XBMC_Event) 6040593
@popcornmix popcornmix Handle resolution changes during video stream.
When the resolution changes, GPU sends a port settings changed message. Host has to acknowledge it by disabling and enabling the affected ports.
@elupus elupus overlays: support ARGB image overlays 11ade98
@FernetMenta FernetMenta Always copy overlays from file parser - Fixes stuck hw resources on r…
@fritsch fritsch AE: Fix resample of e.g. 192 khz to 48 khz audio by scaling the usual…
… suspects with the src sampleRate
@popcornmix popcornmix [rbp] Enable Vsync as a default 9d37db5
@pitpompej pitpompej Prevent timeout error because of waiting for port settings change eve…
…nt on the wrong pipeline object when using deinterlace mode
@popcornmix popcornmix [rbp] Fix for audiotoggledigital on Pi.
Pi only has two audio output modes so the three way toggle gets stuck. This fix corresponds to Pi specific settings code:
@davilla davilla [osx] another mystery xcode project change 8ffa50c
Cory Fields droid: fix loading libs from outside of our apk fdb6e5f
Jim Carroll Introduce a Buffer class and use it to replace custom API definitions…
… for RenderCapture and xbmcvfs.File.
Jim Carroll [cosmetic] remove commented out code. e859e20
Jim Carroll The RenderCapture::getImage now returns a buffer that's ready to read…
… from.
Jim Carroll rename the current 'read' method to 'readBytes' and use it to impleme…
…nt a 'read' method that returns a string in order to preserve bw compatibility.
Jim Carroll change the write back to a bool for bw compatiblity with the incorrec…
…t 12.0 api.
Jim Carroll minor fix to the string length arithmetic. 2dec05a
Jim Carroll return the correct boolean value. 98fd411
Jim Carroll minor change and clarification to the File.write return value. d5b2bfe
arnova fixed: Some http servers provide different content when the range req…
…uest header is not set causing seeking to fail. To fix this use CURL_OPT_RANGE method instead of CURLOPT_RESUME_FROM_LARGE
@davilla davilla [osx] fix broken build on 10.7+, and now we find out what 8ffa50c was…
… all about
Jim Carroll [fix] build issue with std::min on some platforms. 465df62
@Montellese Montellese [win32] add boost/shared_array.hpp needed since 92c7bf0 99ee1f3
Commits on Mar 12, 2013
@opdenkamp opdenkamp [cec] extra guards around m_queryThread fc2e551
@opdenkamp opdenkamp [cec] and let's init m_queryThread to NULL so we won't crash on exit …
…when CEC has been disabled
@ulion ulion Fix code generate problem with binary string. 36a6cb8
@alanwww1 alanwww1 [lang] fixed: revert Serbian non cyrilic changes as they are wrong there 3195310
@davilla davilla Merge pull request #2429 from alanwww1/transifex-frodo-update
[lang] fixed: revert Serbian non cyrilic changes as they are wrong there
Commits on Mar 15, 2013
@Montellese Montellese fix copy-constructor of Buffer, which fixes (fixes #14…
Cory Fields rss: fix feed has been created so that we can prevent this in the
future. It is now a CNAME using the "MyBrand" feature at feedburner, which is
up and running again after the latest news entry.

With this set, we change we can modify the feed at any time in the future.
Commits on Mar 17, 2013
@ScudLee ScudLee Add missing audiochannels rule to PlayerSelectionRule.cpp. fixes #13536 3994d2a
@ace20022 ace20022 [AE] Initialize member vars. f5c9f04
@ulion ulion [IOS] Fix popen crash during xbmc starting up, it not work ios and ma…
…y crash.
@MartijnKaijser MartijnKaijser Bump version to Frodo 12.1 b19f31b
@davilla davilla [ios] fix versions 0d373cc
Commits on Apr 08, 2013
@MartijnKaijser MartijnKaijser Bump version to 12.2-PRE 9bdc862
@Memphiz Memphiz [atv2] - fix learned remote buttons for fwd and rewind ios 6.1 (atv2 …
…5.2) - apple changed the values for those for filling a gap in the enum i guess. Also those 2 buttons are repeatable - so handle them right for not firing 2 messages for one keypress (press, release)
@Memphiz Memphiz [CA/AE] - make the devicelist change callback more robust - should fi…
…x crashes during shutdown on some system constellations (where removing the callback seems to also triggers the callback at the same time)
@Memphiz Memphiz [airplay] - fix the playtime - itunes is very picky about the formatt…
…ing here
@Voyager1 Voyager1 [dvdplayer] fix deadlock when trying to go to disc menu 963892a
spiff fixed: prevent infinite loop in add-on dependency checks 730e312
@Voyager1 Voyager1 fixed: addoninstaller unguarded null pointer after b0825b1 95026c1
@ulion ulion Only add to preDeps when it's not in there. e44fb3f
@garbear garbear Fixed: CAddon::HasSettings() was broken after a call to CAddon::SaveS…
…ettings() on creation of user settings
arnova fixed: We can only obtain the aspect & duration this way when the Pro…
…cess() thread is running
arnova fixed: Some (live) streams no longer played since the Curl seek fix (…
…fixes #14224)
@ace20022 ace20022 [ASF Tag] Fix parsing of the year tag, it's stored as string. a58881a
@ace20022 ace20022 [ASF Tag] Fix parsing of the artist tag. More precisely: fall back to…
… the built in artist method if there's no "author" field in the tag.
@garbear garbear [gtest] Fix compile: missing include file a677bf5
@alcoheca alcoheca upnp: missing info for music videos, fixes #14239 thanks to Pulfer a49656d
@alcoheca alcoheca upnp: hide unplayable items in fiels view fixes #13526 f11d835
@alcoheca alcoheca upnp: use a better way to retain original directory ordering, fixes #…
@alcoheca alcoheca fix typo in ccdba4e 02fd788
@davilla davilla fxied - upnp: not all paths were /'d at the end 0fc5c69
wsoltys fixed memory leak on minimized due to missing deletion of unused text…
…ures (fixes parts of #14245)
@ulion ulion [Fix] work with non-ascii filter word which can not be seperated by s…
@tru tru Remove all direct calls to CFStringGetCString and CFStringGetCStringPtr
This adds a new util function DarwinCFStringToString that handles errors
and tries to first get the pointer otherwise copies the string into a
std::string. It also makes sure to use CFStringGetSystemEncoding()
everywhere so we don't get in trouble for using a hardcoded encoding.
@ulion ulion [Fix] Change from DarwinCFStringRefToString to DarwinCFStringRefToUTF…
@ulion ulion [Fix] MouseStat use button.x/button.y for mouse button event. bf26890
@ulion ulion [Fix] ACTION_GESTURE_NOTIFY should not trigger video osd. 9a952c6
@ulion ulion [Fix] xbmcvfs.exists to support check dir existence. fix #13789 483c288
@fritsch fritsch AE: possible segfault after suspend / resume (m_sink might be gone) bb30267
@popcornmix popcornmix Add support for new video codecs Latest firmware has a start_x.elf wi…
…th support for additional codecs. These are MJPEG, VP6, VP8 and Ogg Theora. They are software GPU accelerated and should be good for SD resolutions. This update should be harmless with firmware that doesn't support the new codecs - they will fail to open and behave as before.
@davilla davilla changed, assert(0) is nasty, this could just be a bad audio packet b81e981
@fritsch fritsch AE: Revisit Device Opening. Try to set periodSize of 100 ms and Buffe…
…rSize of 800 ms
@fritsch fritsch AE: Fix menu sounds by decreasing buffer(max 200 ms) and periodSize(5…
…0 ms)
@anssih anssih AE: ALSA: Try to get 200ms buffer even if we did not get 50ms periodsize df0b551
@anssih anssih AE: ALSA: Try to get a minimum of 4 periods per buffer 494a29f
@juanfont juanfont Avoid fps detection when probing the format if we will trust codec fp…
…s during playback.
@juanfont juanfont Improved code legibility. 5beb13d
@elupus elupus dvdplayer: don't use OpenDefaultStreams() if something else dictate
This works for both dvd and bluray navigation
@FernetMenta FernetMenta dvdplayer: make HasVideo return true if a video stream was opened sin…
…ce playback has started - fixes missing video info on channel change
@FernetMenta FernetMenta dvdplayer: align HasAudio with HasVideo 4d76051
@elupus elupus dvdplayer: a disabled stream could be re-selected for bluray and dvd de13762
@elupus elupus dvdplayer: allow 200ms of automatic time update instead of 1ms
It was originally meant to allow 1 second. But has at some point
been broken since the unit has been changed from ms to dvd time
@elupus elupus dvdplayer: fixed time_offset is difference between adjusted pts and time 4394389
@sraue sraue fixed: Compile warning (unused label) Based on 132730e in master and …
…added in 28de0da
@sraue sraue dvdplayer: don't use OpenDefaultStreams() if something else dictate B…
…ased on dda59f0 in master and added in 28de0da
@sraue sraue dvdplayer: make HasVideo return true if a video stream was opened sin…
…ce playback has started - fixes missing video info on channel change Based on bd49c32 in master and added in 28de0da
@sraue sraue dvdplayer: align HasAudio with HasVideo Based on cfb14c3 in master an…
…d added in 28de0da
@sraue sraue dvdplayer: a disabled stream could be re-selected for bluray and dvd …
…Based on b2fc050 in master and added in 28de0da
@sraue sraue dvdplayer: ff/rw/seek was broken for inputs implementing IDisplayTime…
… Based on 4c90033 in master and added in 28de0da
@sraue sraue dvdplayer: allow 200ms of automatic time update instead of 1ms Based on
a489fb5 in master and added in 28de0da
@sraue sraue fix deadlock when trying to go to disc menu Based on befdcfa in maste…
…r and added in 28de0da
@sraue sraue dvdplayer: another fix for seeking/fw/rw in PVR Based on 676b9e7 in m…
…aster and added in 28de0da
@sraue sraue dvdplayer: fixed time_offset is difference between adjusted Based on 4…
…06ef63 in master and added in 28de0da
@sraue sraue fixed: We can only obtain the aspect & duration this way when the Pro…
…cess() thread is running Based on 9db853e in master and added in 28de0da
@sraue sraue cosmetic changes: sync with dvdplayer (changes) this was added in 28d… 51d5a18
@davilla davilla [rbp] Set mediatime on GPU after a seek. 5b52212
@FernetMenta FernetMenta dvdplayer: reevaluate HasVideo/Audio after a stream change c707a7c
@FernetMenta FernetMenta pvr: remove check for different types of channels on SwitchChannel ca7d05f
@FernetMenta FernetMenta pvr: fix too early finish of recordings 70a9005
@FernetMenta FernetMenta pvr: align NexStream and SkipNext with logic in dvdplayer 05405ba
Commits on Apr 09, 2013
@alcoheca alcoheca upnp: missing playlist & musicvideo nodes for music library f1a3644
@alcoheca alcoheca upnp: smartplaylists had incorrect upnp:class d78eefb
@alcoheca alcoheca upnp: fixed missing fanart and thumbs for playlist items. fixes #13945 114f255
arnova changed: Cosmetics 301dbd9
arnova fixed: We should only try to extract/show thumbs/flags for file-like …
…items (fixes #14176)
arnova fixed: The check for determining whether we're done was wrong 6007931
arnova fixed: Auto flags didn't work for items without a VideoInfoTag 73ae781
arnova changed: Improve logic in GetStreamDetails() for items without a (com…
…plete) VideoInfoTag
arnova fixed: Setting streamdetails for auto thumbs/flags didn't work for it…
…ems with incomplete/empty infotags
arnova fixed: Make sure item has a VideoInfoTag before accessing it b7e10d8
arnova added: Doxygen comments to CFileItem:IsAudio/IsVideo/IsPicture 7c318a4
@pieh pieh GUIScrollBarControl: fix onright/ondown navigation when master contro…
…l doesn't have enough items to fill at least 1 page
@popcornmix popcornmix Ensure GPU does resize when caching thumbnails
On a Pi, imageres and fanartres are set quite low (typically 512 and 720 or lower).
So, when caching the artwork there is a resize step involved.
The OMXImage jpeg decoder has the option to resize "for free".
Unfortunately this currently only limits the resize to current GUI resolution,
so there is often a second resize done by the ARM. This is very slow.
The larger decoded image also requires more RAM.

With fanartres=512, you currently get resize during decode of 1920x1080 to 1280x720 (gui res).
Then before encode there is a resize to 910x512. The two stage resize reduces image quality.

This fix takes a bit more care to include the imageres/fanartres resize during the decode.
I needed to make the 16:9 check a little more lax to also detect a 16:9 image after it's been resized.
I've gone for differs from 16:9 by 1%, rather than the existing exact match of 16:9.
@popcornmix popcornmix Always acquire lock when using OMXMediaTime
I don't believe the lock=false call should ever be done unless caller already has lock
@dagwieers dagwieers Name threads on Linux properly
I implemented this as part of the OpenELEC project.

It helps debug issues if we can easily identify the XBMC thread by name. This patch implements just that.

This second incarnation now implements a proper configure check to ensure we are using the capabilities of the library on various platforms (esp. when using glibc < 1.12).
@popcornmix popcornmix [rbp] Add reporting of fifo fullness for use with hdmi_status_show in…
… debug builds
@popcornmix popcornmix [rbp] Move m_messenger.Init from constructor to OnStartup
This makes omxplayer constistent with dvdplayer.
It avoids an initial stray message to immediately try and go normal speed when it should be buffering.
@popcornmix popcornmix [rbp] Expose GetAudioRenderingLatency
This could be useful for detecting underrun conditions
@popcornmix popcornmix [rbp] Add OMXLateCount to query if packets have arrived late
Could be useful for detecting underrun conditions
@popcornmix popcornmix [rbp] Use CurrentHostCounter consistently 51c136e
@popcornmix popcornmix [rbp] Use correct size when determining fifo has space in it
This can cause player to get a decode error from audio_decode component
closing stream when packet doesn't actually fit.
@popcornmix popcornmix [rbp] Add more debug logging 498b697
@popcornmix popcornmix [rbp] Cosmetics a2f35a0
@popcornmix popcornmix [rbp] Avoid coming out of stalled state due to uninitialised variable
m_endtime is only initialised when we underrun. Use m_av_clock->OMXAudioBuffer() to protect access to uninitialised variable.
@popcornmix popcornmix [rbp] The stillframe logic doesn't work on Pi and causes stuttering cf8a307
@elupus elupus CoreAE: fix invalid logging about mixing matrix efc25de
@elupus elupus CoreAE: mixing unit must have same output count as mixing map 1200cb6
@elupus elupus CoreAE: fix null dereference if no mixer map was created 1cee5e1
@elupus elupus CoreAE: mixerunits input is the inputunits output
Not this still doesn't resolve all issues. We are still not
setting up a correct channel map for the output
@elupus elupus dvdplayer: complete the update of ffmpeg interrupt interface
The old interface could not handle being called from another thread
@michaelni michaelni udp: Fix sign of error codes.
Signed-off-by: Michael Niedermayer <>
Nicolas George udp: fix non-blocking and interrupt handling.
In non-blocking mode, lowest-level read protocols are
supposed block only for a short amount of time to let
retry_transfer_wrapper() check for interrupts.

Also, checking the interrupt_callback in the receiving thread is
wrong, as interrupt_callback is not guaranteed to be thread-safe
and the job is already done by retry_transfer_wrapper(). The error
code was also incorrect.

Bug reported by Andrey Utkin.
@elupus elupus dvdplayer: make sure we can also abort the open of a ffmpeg input stream 5e4b832
@elupus elupus dvdplayer: fixed crash introduced with change in handling of abort 1ca77bb
@Memphiz Memphiz [airplay] - handle empty session ids from itunes b683b08
@Memphiz Memphiz [airplay] - refactor the playstate announcement by using the IAnnounc…
…er interface

- also adapt to some new findings on revers engineering itunes airplay traffic (add sessioncounter to reverse event and fix http request formatting)
- fixes #14191


Commits on Apr 10, 2013
@davilla davilla Merge pull request #2581 from Memphiz/frodo_aefixes
[CA/AE] - fixes
@davilla davilla Merge pull request #2582 from Memphiz/frodo_airplayconflicts
Frodo airplayconflicts
@davilla davilla Merge pull request #2558 from elupus/abort_frodo
dvdplayer: fix abort of ffmpeg streams
@davilla davilla Merge pull request #2569 from alcoheca/upnp-playlists-frodo
Upnp Playlist bug fixes for Frodo
Commits on Apr 12, 2013
@ScudLee ScudLee Only override episode/season numbers if the file name has them ed4d224
@mooninite mooninite AE: Pulse Audio fix crash on player stop.
PA is calling back functions on stream destruction and may try
to read a free'd stream variable. (trac bug 13903)

Signed-off-by: Michael Cronenworth <>
@mooninite mooninite AE: Pulse Audio fix buffer underruns.
The cache buffer size was being initialized to the write size
of a PA stream, which is not the buffer size of the PA server.
The cache buffer size will now use the PA server buffer size.
(trac bug 13902)

Signed-off-by: Michael Cronenworth <>
Commits on Apr 14, 2013
@davilla davilla [m3u8/airplay] - strip off any url options before checking
 for m3u8 filename - fixes stream quality selection (e.x. for airplay ^^) -
 right way this time
Commits on Apr 15, 2013
@alcoheca alcoheca [fix] invalid mediaitems trumped upnp listings, resolves #14268 0a74686
Commits on Apr 16, 2013
@davilla davilla Merge pull request #2618 from alcoheca/upnp-handle-non-mediafiles
[fix] non-media items trumped upnp listings, fixes #14268
@ulion ulion StringUtils::FindWords fix dead loop. 1e5d6e9
@pieh pieh GUIBaseContainer: remove unneeded call from CGUIBaseContainer::Render…
…() - we are calling it in CGUIBaseContainer::Process() earlier
@pieh pieh GUI containers: adjust page control offset when scrolling down to mat…
…ch behaviour when scrolling up
@ScudLee ScudLee Only start checking the nfo file from the first episodedetails tag f1016fc
Commits on Apr 18, 2013
@alcoheca alcoheca [fix] upnp: parentID of the root container should be -1, fixes #14279 804898f
Commits on Apr 19, 2013
@davilla davilla Merge pull request #2630 from alcoheca/upnp-root-parent
[fix] upnp: parentID of the root container should be -1, fixes #14279
Commits on Apr 23, 2013
huceke make sure we can also abort the open of a ffmpeg input stream 6ca8056
huceke reevaluate HasVideo/Audio after a stream change 79c9dd9
@popcornmix popcornmix Fix elapsed time not updating near end of file
Currently, once demuxer has reached EOF, we send the EOF messages to audio/video players
and set their inited/started flags to false. But if started is false we ignore any
PLAYER_DISPLAYTIME coming back from players, which stops the elapsed time from updating.
This can affect the final ~8 seconds of the file.

The fix delays changing these flags until audio/video players have signalled EOS.
@popcornmix popcornmix Fix audio volume jumps when switching tracks
When the global volume has been set low, and a sequence of tracks are being played, there are complaints of occasional jumps to full volume.
This is down to a race condition where the volume request can arrive at OMXAudio before it has been initialised.
The fix is simple, don't send the volume change until m_CurrentAudio.started.
@ScudLee ScudLee Allow dashes in music videos on the first pass 888f540
Commits on Apr 24, 2013
@davilla davilla Merge pull request #2648 from ScudLee/musicdashfixFrodo
Allow dashes in music videos on the first pass
@koying koying FIX: Allow "Scan for new content" on folders 5328fe6
@popcornmix popcornmix Don't block waiting for EOS in audio/video players
Currently we block in OMXPlayerAudio/OMXPlayerVideo from the point we see EOF from demuxer,
until the last frame/audio sample has been played out. This can be a few seconds.
It means no more messages (such as abort) can be received during this period.
This results in a bug where if you press stop after the demuxer EOF has occurred it takes
a long time to stop. You would expect this to be the few seconds of queued data,
but it actually turns out to be 30 seconds, as the clocks get stopped by the stop message,
but the players never find out and we hit a timeout.
It also stops seek/pause working during the playout period.
It also stops (graphical) subtitles from being rendered during this time.
The fix involves not blocking for the EOS, but allowing the polling from OMXPlayer to catch it.
@davilla davilla Merge pull request #2653 from sraue/Frodo-fixes.1
[Frodo] Don't block waiting for EOS in audio/video players
Commits on Apr 25, 2013
@alanwww1 alanwww1 [lang] update of internal addon language files 205b643
@alanwww1 alanwww1 [lang] update of core language files 151f530
@alanwww1 alanwww1 [lang] update of skin.confluence language files 41fee92
Commits on Apr 27, 2013
@alanwww1 alanwww1 [lang] added langinfo.xml for new language Latvian 89ff51c
@alanwww1 alanwww1 [lang] added langinfo.xml for new language Faroese 5f0445c
@alanwww1 alanwww1 [lang] added langinfo.xml for new language Welsh 3fa5c24
Commits on Apr 29, 2013
@davilla davilla Merge pull request #2392 from arnova/Frodo_dvdplayerfix
fixed: We can only obtain the aspect & duration this way when the Proces...
@davilla davilla Merge pull request #2662 from alanwww1/transifex-frodo-update
[lang] Update of language files from Transifex
Commits on Apr 30, 2013
@ronie ronie correct date formats. fixes #14286 dc78b8f
@MartijnKaijser MartijnKaijser sync scrapers with repo 2564cd4
arnova fixed: The db check for streamdetails was wrong ed6e20e
arnova fixed: Streamdetails & resume-bookmark saving etc. for bluray folders…
… didn't work (fixes #14288)
@davilla davilla Merge pull request #2672 from arnova/Frodo_bluray_fix
fixed: Streamdetails & resume-bookmark saving etc. for bluray folders di...
arnova fixed: Freezing when player fails to queue the next song (fixes #1429…
…7 #14231 #14229 #13919)
arnova fixed: DVDPlayer GUI display time not updating near the end of file (…
…fixes #14299)
@davilla davilla [vda] disable use of cvBufferRef for 12.2+ 4dfa1aa
Commits on May 02, 2013
@MartijnKaijser MartijnKaijser bump version to 12.2 32b1a5e
Commits on Sep 12, 2013
@readmanr readmanr Frodo English spelling corrections English/strings.po acd6a56
@MartijnKaijser MartijnKaijser Merge pull request #3242 from readmanr/Frodo-english-check
Frodo English spelling corrections English/strings.po
@MartijnKaijser MartijnKaijser Revert "Merge pull request #3242 from readmanr/Frodo-english-check"
This reverts commit d5c4887, reversing
changes made to 32b1a5e.
Commits on Oct 20, 2013
@MartijnKaijser MartijnKaijser [release] bump version to 12.3 885dfbc
Commits on Oct 26, 2013
@popcornmix popcornmix Remove unnecessary gl ifdef
I'm not sure why this was originally added.
I couldn't think of a reason why the Pi would want things done differently here,
so I tried without the ifdef. I can't see any difference in behaviour,
so I think we're better off removing it.
@elupus elupus dvdplayer: move CanSeek/CanPause to seekable interface 5d3f9cb
@elupus elupus dvdplayer: disable seeking and pause for udp/rtp and seek for tcp dd6209b
@whaupt whaupt dvdplayer: move navigator state into IMenus 2a284eb
@ScudLee ScudLee Re-get details from nfo file after advancing to first episodedetails 3738408
@ulion ulion Fix color index overflow by reuse existed color in the vector. Fix #1…
@ulion ulion [OSX] Fix always on top after restore from fullscreen. c347f4a
@ulion ulion Fix undefined reference caused by wrong detection of gcc builtin atom…
…ic functions.
@ulion ulion avoid detect folder.jpg under 'add' item. 4f0b344
@ulion ulion Check exists before listing dir, to avoid produce error log, fix #14210 e6a3ead
@bobo1on1 bobo1on1 fixed: when doing a format conversion from float to 32 bit int, multi…
…ply with INT32_MAX - 127, since this is the maximum value that can be stored in 32 bit float and int, if INT32_MAX gets converted to float, it gets rounded to INT32_MAX + 1 which can cause wrap around distortion
@hmis hmis Read CD audio USB drive bug fixed
libcdio seems to allow read no more than about 10 audio sectors at once when CD audio device is connected via USB.
This patch makes XBMC read small number of sectors if default one fails. It uses more CPU but allows to use USB CD devices.
Tested on GNU/Linux x86 and RPi. (On Rpi OMXPlayer does not play CD, but I can rip tracks).
@FernetMenta FernetMenta backport some constructor initializations from 92e8bc4 c2c97cf
@ulion ulion Fixed: do not send Range request header when encounter error. 7eb2e03
@Montellese Montellese CGUISliderControl: only switch between selectors on <Enter> if there …
…are more than one
@ulion ulion Fix color tag didn't hide bug introduced by PR2725. a654be5
@FernetMenta FernetMenta Linux: fix handling of zombie processes efbb36f
@FernetMenta FernetMenta pvr: fix channel switch for addons using other stream 2284a23
@Memphiz Memphiz - bump libnfs to 1.6 ae838d8
@Memphiz Memphiz - bump libnfs to 1.6 731704f
@Memphiz Memphiz - bump libnfs to 1.6 c45ee93
@Memphiz Memphiz - bump libnfs to 1.6 5babf57
@opdenkamp opdenkamp sync add-ons * opdenkamp/xbmc-pvr-addons#178 * opdenkamp/xbmc-pvr-add… e8a590a
@ulion ulion [OSX] Kickout screensaver if it's running e66c544
@davilla davilla Generate stable DVD disk IDs by performing CSS first fd88e70
@jmbreuer jmbreuer Get and use correct (removable://) file record from database Fixes DV…
…D resume after resume details already exist
@Memphiz Memphiz Revert "[AE/CA] - ooppps - typo in last commit (note to self - commit…
… on the same host you compiled it)"

This reverts commit 1c7791d.
@Memphiz Memphiz - fix regression (stuck ringbuffer after 46min, 62min, 180min - depen…
…ding on the stream setup). Revert "[AE/CA] - fix a possible overflow in the ringbuffer during readsize calculation (yes we hit this in pvr branch in internalflush - when zapping channels)"

This reverts commit e447c4b.
@Memphiz Memphiz - fix race in 2 threads consuming the ringbuffer (dvdplayeraudio call…
…s flush and ca thread still pulls frames) by moving the flush to the ca thread instead - fixes crashing when switching tv channels (this was the real issue the 2 reverted commits before wanted to fix...)
@anssih anssih SoftAE: Fix A/V sync issues caused by wrong buffer time calculation
CSoftAE::GetDelay(), CSoftAE::GetCacheTime(), and
CSoftAE::GetCacheTotal() assume in their calculations that m_buffer
contains frames that have sink/encoder frame size. However, m_buffer
actually contains frames in the format received from CSoftAEStream.

This causes varying levels of A/V de-sync depending on other conditions.
On my Linux system I encounter this always when using using HDMI audio
and having the display refresh rate set at 24Hz.

Fix those functions to use the correct frame size.
@Memphiz Memphiz - fix bt keyboard support (fixes space, tab, esc, enter and modifier …
@taxigps taxigps fixed bug with arial.ttf a5b3131
@wsoltys wsoltys fixed resource leak. a5fc05e
@wsoltys wsoltys fixed: fixed memory leaks in ximadsp.cpp 9cf5e5f
@wsoltys wsoltys fixed memory leak in ximatif.cpp 2166aee
@mkortstiege mkortstiege [linux] - fix wrong local FileTime due to not honoring daylight savin…
…gs time
Jonathan Marshall Use round() rather than MathUtils::round_int() for time, and modulate…
… to unsigned int. Fixes #14205
@fritsch fritsch AE: Workaround (ugly) non existing channel maps in old ffmpeg (workar…
…ounds: #14407)
@ulion ulion [Fix] Event server code for remote mouse with event client protocol 0c233ac
@ScudLee ScudLee [Fix] Only use episode number in nfo check if it's an episode 42a648e
@amet amet fix app start on 10.9 e232f83
@Montellese Montellese CGUIWindowFoo: properly initialize m_dlgProgress pointer to NULL (fix…
…es #14635)
ronie add weather provider logo 70ebc49
arnova fixed: Video chapter seek handling didn't work properly for some cont…
…ent handled by ffmpeg

This mainly applies to several MP4 files where we currently can't reliably detect in which chapter we are when
we perform chapter seeking (probably an ffmpeg bug?). To workaround this make sure current chapter != 0 and
fallback to large stepping where appropriate.
@FernetMenta FernetMenta kaitoast: fix deadlock caused by grabbing the gfx lock while other lo…
…cks are held
@koying koying FIX: crash with missing controls in DialogKaiToast 1222b7a
@koying koying fixup kaitoast 4cce50a
@garbear garbear Fix infinite loop while viewing "EPG: Timeline" window
With invalid begin/end times, the for loop in GUIEPGContainer.cpp L#825 ( loops endlessly, freezing XBMC when the user views the "EPG: Timeline" window. Solve this by adding check for invalid end times.
@Montellese Montellese webserver: fix crash on HTTP GET arguments without a "=" (fixes #14650) dfbbc5e
@davilla davilla fix the long standing coreaudio mixmap bugs with ae 55ba63d
Commits on Oct 27, 2013
Arne Morten Kvarving fixed: scraper fallback logic
code ended up selecting the default album scraper for artists
Jonathan Marshall invalid use of genre ID when checking for a scraper set on an album 9d75530
Commits on Oct 28, 2013
@popcornmix popcornmix omxplayer: move CanSeek/CanPause to seekable interface fdd81c7
@popcornmix popcornmix omxplayer: move navigator state into IMenus 4ff0707
Commits on Oct 30, 2013
@Memphiz Memphiz [osx/mavericks] - ugly hacked fix for the mouse bug (mouse wasn't abl…
…e to move over the full screen) - intermediated isolated hacked solution until new windowing is ready
@Memphiz Memphiz [osx/mavericks] - hide the menubar not only when on mainscreen - mave…
…ricks has menubars on all screens - always hide them
Commits on Oct 31, 2013
@davilla davilla Merge pull request #3524 from Memphiz/Frodo
[osx/mavericks] - fixups for mavericks 10.9
@MartijnKaijser MartijnKaijser sync addons with repo 25d29c3
Commits on Nov 01, 2013
wsoltys [WIN32] fixed: make frodo branch jenkins ready 940c1c8
@davilla davilla Merge pull request #3538 from wsoltys/frodo3
[WIN32] fixed: make frodo branch jenkins ready
@davilla davilla regression with hebrew characters.
Revert "fixed bug with arial.ttf"

This reverts commit a5b3131.
@davilla davilla Merge pull request #3518 from popcornmix/Frodo_fix_rpi
[rbp/omxplayer] Add missing omxplayer changes to previous commits
Commits on Nov 05, 2013
@opdenkamp opdenkamp [pvr] bump add-ons a981447
@davilla davilla Merge pull request #3569 from opdenkamp/frodopvr
[pvr] bump add-ons
Commits on Dec 01, 2013
Jonathan Marshall fix crash loading JPEGs where m_cinfo.scale_num was set to 9 due to n…
…ot finding a suitably sized scale factor. Thanks to tscheck
@davilla davilla [osx/sdl] - fix possible NPE - fixes #14676 8eed85b
Commits on Dec 05, 2013
@Memphiz Memphiz - fix crashing with libnfs 1.6 by bumping to 1.6.1 (which is 1.6.0 + …
…a compile fix - 32bit time_t define was missing in the release target of vs project for libnfs)
Commits on Dec 12, 2013
@Memphiz Memphiz [darwin] - move isMavericks helper to DarwinUtils 3277ba4
@Memphiz Memphiz [osx/mavericks] - on mavericks don't use pbo rendering with intel gpu…
…s as it might result in distorted rendering (ripple)
@davilla davilla Merge pull request #3820 from Memphiz/Frodo
[osx/mavericks] fix rippled rendering with intel gpu
@davilla davilla Merge pull request #3821 from MartijnKaijser/12.3
[release] bump version to 12.3