Re: [Musicpd-dev-team] FS improvements + refinements in PollGroupWinSelect
On Thu, Dec 05, 2013 at 19:17:56 +0600, Denis Krjuchkov wrote: g_get_home_dir() works differently depending on GLib version. Newer GLib versions check $HOME first and if it's empty use getpwnam_r() to obtain home directory for currently user. Old versions go straight into using getpwnam_r(). What approach should we use? Please use $HOME then fall back to getpwnam_r(). I have a user with the same uid/gid, but a different name and $HOME (mainly to put games into a special root without a separate login), but getpwnam_r() gets my main user's $HOME because it's earlier in the passwd file. --Ben -- Sponsored by Intel(R) XDK Develop, test and display web and hybrid apps with a single code base. Download it for free now! http://pubads.g.doubleclick.net/gampad/clk?id=111408631iu=/4140/ostg.clktrk ___ Musicpd-dev-team mailing list Musicpd-dev-team@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/musicpd-dev-team
Re: [Musicpd-dev-team] Tags in OGG httpd stream
On Sat, Oct 19, 2013 at 15:47:25 +0200, Max Kellermann wrote: On 2013/10/11 05:12, Ben Boeckel maths...@gmail.com wrote: the tag chunk is dropped because chunk-length is 0 for it. A patch which fixes the issue is attached. Merged, thanks. Sorry for letting you wait for so long. Thanks! No problem; I'm getting other patches through to get these tags visible on the player end too (ffmpeg and mpv need patched; mplayer had its own ogg decoder); quite the adventure! --Ben -- October Webinars: Code for Performance Free Intel webinars can help you accelerate application performance. Explore tips for MPI, OpenMP, advanced profiling, and more. Get the most from the latest Intel processors and coprocessors. See abstracts and register http://pubads.g.doubleclick.net/gampad/clk?id=60135031iu=/4140/ostg.clktrk ___ Musicpd-dev-team mailing list Musicpd-dev-team@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/musicpd-dev-team
Re: [Musicpd-dev-team] Tags in OGG httpd stream
On Wed, Dec 05, 2012 at 16:51:54 -0500, Ben Boeckel wrote: @@ -524,31 +532,7 @@ decoder_replay_gain(struct decoder *decoder, float return_db = 0; assert(decoder != NULL); - if (replay_gain_info != NULL) { - static unsigned serial; - if (++serial == 0) - serial = 1; - - if (REPLAY_GAIN_OFF != replay_gain_mode) { - return_db = 20.0 * log10f( - replay_gain_tuple_scale( - replay_gain_info-tuples[replay_gain_get_real_mode()], - replay_gain_preamp, replay_gain_missing_preamp, - replay_gain_limit)); - } - - decoder-replay_gain_info = *replay_gain_info; - decoder-replay_gain_serial = serial; - - if (decoder-chunk != NULL) { - /* flush the current chunk because the new -replay gain values affect the following -samples */ - decoder_flush_chunk(decoder); So this line here was the issue. This flushes the chunk into the stream with just the tag and no data on it. Combined with this code right before the chunk leaves for the sources: static bool play_chunk(player_control pc, Song *song, struct music_chunk *chunk, MusicBuffer buffer, const AudioFormat format, Error error) { assert(chunk-CheckFormat(format)); if (chunk-tag != nullptr) update_song_tag(song, *chunk-tag); if (chunk-length == 0) { buffer.Return(chunk); return true; } pc.Lock(); pc.bit_rate = chunk-bit_rate; pc.Unlock(); /* send the chunk to the audio outputs */ if (!audio_output_all_play(chunk, error)) return false; pc.total_play_time += (double)chunk-length / format.GetTimeToSize(); return true; } the tag chunk is dropped because chunk-length is 0 for it. A patch which fixes the issue is attached. - g_cond_signal(decoder-dc-client_cond); - } - } else - decoder-replay_gain_serial = 0; + replay_gain_state_set_info(decoder-replay_gain, replay_gain_info); --Ben From 9f7690767838052661160de5a30d28acc6ef3473 Mon Sep 17 00:00:00 2001 From: Ben Boeckel maths...@gmail.com Date: Thu, 10 Oct 2013 23:00:52 -0400 Subject: [PATCH] PlayerThread: Only drop 0 length packets without tags Fixes a regression from 752dfb3d95482c562e5d24c6ea839c4815de9a6d which caused the current chunk to be flushed as soon as new replaygain information was found. If this occurs on a tag chunk, it has no data (length 0) and is then skipped before pushing it to all of the outputs. This change allows 0-length chunks through if they contain a tag and they are now appearing in mplayer and mpv properly. --- src/PlayerThread.cxx | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/src/PlayerThread.cxx b/src/PlayerThread.cxx index 9ad37ea..534a24e 100644 --- a/src/PlayerThread.cxx +++ b/src/PlayerThread.cxx @@ -726,7 +726,7 @@ play_chunk(player_control pc, if (chunk-tag != nullptr) update_song_tag(song, *chunk-tag); - if (chunk-length == 0) { + if (chunk-length == 0 chunk-tag == nullptr) { buffer.Return(chunk); return true; } -- 1.8.3.1 -- October Webinars: Code for Performance Free Intel webinars can help you accelerate application performance. Explore tips for MPI, OpenMP, advanced profiling, and more. Get the most from the latest Intel processors and coprocessors. See abstracts and register http://pubads.g.doubleclick.net/gampad/clk?id=60134071iu=/4140/ostg.clktrk___ Musicpd-dev-team mailing list Musicpd-dev-team@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/musicpd-dev-team
Re: [Musicpd-dev-team] Tags in OGG httpd stream
On Thu, Oct 10, 2013 at 23:12:27 -0400, Ben Boeckel wrote: So this line here was the issue. This flushes the chunk into the stream with just the tag and no data on it. Combined with this code right before the chunk leaves for the sources: It seems this fix isn't complete actually. It works properly if I do 'mpd next', but if the song changes as usual, the tags don't show up again. This will have to wait until next week for me to dig in further. --Ben -- October Webinars: Code for Performance Free Intel webinars can help you accelerate application performance. Explore tips for MPI, OpenMP, advanced profiling, and more. Get the most from the latest Intel processors and coprocessors. See abstracts and register http://pubads.g.doubleclick.net/gampad/clk?id=60134071iu=/4140/ostg.clktrk ___ Musicpd-dev-team mailing list Musicpd-dev-team@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/musicpd-dev-team
Re: [Musicpd-dev-team] Tags in OGG httpd stream
On Fri, Oct 11, 2013 at 00:14:46 -0400, Ben Boeckel wrote: On Thu, Oct 10, 2013 at 23:12:27 -0400, Ben Boeckel wrote: So this line here was the issue. This flushes the chunk into the stream with just the tag and no data on it. Combined with this code right before the chunk leaves for the sources: It seems this fix isn't complete actually. It works properly if I do 'mpd next', but if the song changes as usual, the tags don't show up again. This will have to wait until next week for me to dig in further. Or this is just me not checking mplayer, which is fine in both cases. mpv/ffmpeg isn't seeing the new tags, which is a separate issue. Time for bed if I'm missing this stuff... --Ben -- October Webinars: Code for Performance Free Intel webinars can help you accelerate application performance. Explore tips for MPI, OpenMP, advanced profiling, and more. Get the most from the latest Intel processors and coprocessors. See abstracts and register http://pubads.g.doubleclick.net/gampad/clk?id=60134071iu=/4140/ostg.clktrk ___ Musicpd-dev-team mailing list Musicpd-dev-team@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/musicpd-dev-team
[Musicpd-dev-team] Crash on master on exit/stream disconnect
When disconnecting a stream or stopping (SIGINT) mpd, I'm seeing the following backtrace with master (17c6db6c33b9b14a8627b3f9b3da8580a5f932a2): #0 EventLoop::RemoveFD (this=0x0, _fd=0, m=...) at src/event/Loop.cxx:64 #1 0x0043f0d7 in SocketMonitor::Schedule (this=0x1a74c28, flags=flags@entry=0) at src/event/SocketMonitor.cxx:172 #2 0x0043f13b in Cancel (this=optimized out) at src/event/SocketMonitor.hxx:136 #3 SocketMonitor::Dispatch (this=optimized out, flags=optimized out) at src/event/SocketMonitor.cxx:43 #4 0x0044066b in EventLoop::Run (this=0x14f3c60) at src/event/Loop.cxx:176 #5 0x0041b436 in mpd_main (argc=optimized out, argv=optimized out) at src/Main.cxx:503 #6 0x7fd8cc0d0f75 in __libc_start_main (main=0x409a80 main(int, char**), argc=4, argv=0x7fff4df90918, init=optimized out, fini=optimized out, rtld_fini=optimized out, stack_end=0x7fff4df90908) at libc-start.c:285 #7 0x00409c49 in _start () I can bisect if needed, but it's pretty easy to reproduce here with just connect/disconnecting an mplayer or mpv instance a couple of times. --Ben -- October Webinars: Code for Performance Free Intel webinars can help you accelerate application performance. Explore tips for MPI, OpenMP, advanced profiling, and more. Get the most from the latest Intel processors and coprocessors. See abstracts and register http://pubads.g.doubleclick.net/gampad/clk?id=60134071iu=/4140/ostg.clktrk ___ Musicpd-dev-team mailing list Musicpd-dev-team@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/musicpd-dev-team
Re: [Musicpd-dev-team] FileSystem: new library for Path-friendly file system routines
On Sun, Jan 20, 2013 at 19:10:28 +0600, Denis Krjuchkov wrote: Pushed slightly improved version: http://git.musicpd.org/cgit/dk/mpd.git/commit/?id=e6ed592b8aeb5025be0893ee99ff44e46a9ffd1c For CheckExists, you should use lstat instead of stat. stat will error if the path is a broken symlink. --Ben -- Master Visual Studio, SharePoint, SQL, ASP.NET, C# 2012, HTML5, CSS, MVC, Windows 8 Apps, JavaScript and much more. Keep your skills current with LearnDevNow - 3,200 step-by-step video tutorials by Microsoft MVPs and experts. ON SALE this month only -- learn more at: http://p.sf.net/sfu/learnmore_123012 ___ Musicpd-dev-team mailing list Musicpd-dev-team@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/musicpd-dev-team
Re: [Musicpd-dev-team] UTF-8 file names support for Windows
On Sun, Jan 20, 2013 at 16:20:13 +0100, Max Kellermann wrote: C++ provides wrappers for standard C headers (e.g. cstdio for stdio.h). Are there any rules on them? I tend to use them, however they have serious disadvantage: it's hard to predict whenever some function would appear to std namespace. So should we dismiss those headers? I have never used the C++ wrapper headers, I don't know why they exist and whether using them has an advantage. Until somebody explains it to me, I'll keep on using standard C headers, but I wouldn't mind patches using C++ headers and std::. Things like std::min and std::max are implemented using overloads instead of as macros. Same with pretty much all of the math functions. Looking at GCC 4.7.2's cstdlib, there are quite a few functions which are implemented using real functions instead of macros (so you can pass them as arguments to std::for_each and other higher-level functions from algorithm). --Ben -- Master Visual Studio, SharePoint, SQL, ASP.NET, C# 2012, HTML5, CSS, MVC, Windows 8 Apps, JavaScript and much more. Keep your skills current with LearnDevNow - 3,200 step-by-step video tutorials by Microsoft MVPs and experts. ON SALE this month only -- learn more at: http://p.sf.net/sfu/learnmore_123012 ___ Musicpd-dev-team mailing list Musicpd-dev-team@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/musicpd-dev-team
Re: [Musicpd-dev-team] FileSystem: new library for Path-friendly file system routines
On Sun, Jan 20, 2013 at 17:14:12 +0100, Max Kellermann wrote: - slightly disagree with Ben (but thanks for pointing out that potential problem): there should be some way to specify whether symbolic links shall be followed. Some callers may want to follow symlinks, some do not, and the database update has that configurable at runtime. Maybe have a CheckFileExists (stat) and a CheckPathExists (lstat)? --Ben -- Master Visual Studio, SharePoint, SQL, ASP.NET, C# 2012, HTML5, CSS, MVC, Windows 8 Apps, JavaScript and much more. Keep your skills current with LearnDevNow - 3,200 step-by-step video tutorials by Microsoft MVPs and experts. ON SALE this month only -- learn more at: http://p.sf.net/sfu/learnmore_123012 ___ Musicpd-dev-team mailing list Musicpd-dev-team@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/musicpd-dev-team
Re: [Musicpd-dev-team] FileSystem: new library for Path-friendly file system routines
On Sun, Jan 20, 2013 at 17:25:26 +0100, Max Kellermann wrote: bool IsRegularFile(Path, bool follow_symlinks=true) That works as well. --Ben -- Master Visual Studio, SharePoint, SQL, ASP.NET, C# 2012, HTML5, CSS, MVC, Windows 8 Apps, JavaScript and much more. Keep your skills current with LearnDevNow - 3,200 step-by-step video tutorials by Microsoft MVPs and experts. ON SALE this month only -- learn more at: http://p.sf.net/sfu/learnmore_123012 ___ Musicpd-dev-team mailing list Musicpd-dev-team@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/musicpd-dev-team