April 2023 - Cin - Cinelerra-GG Mailing Lists

Re: [Cin] Simplest demo of BT 638 crash
by Andrew Randrianasulu 26 Apr '23

26 Apr '23

so far I get this Backtrace: Thread 88 "cin" received signal SIGSEGV, Segmentation fault. [Switching to Thread 0xc4612b00 (LWP 16879)] 0xf4e0f37d in DelayAudio::process_realtime(long long, Samples*, Samples*) () from /usr/lib/cin/plugins/audio/delayaudio.plugin (gdb) bt full #0 0xf4e0f37d in DelayAudio::process_realtime(long long, Samples*, Samples*) () from /usr/lib/cin/plugins/audio/delayaudio.plugin No symbol table info available. #1 0x08709468 in PluginAClient::process_buffer(long long, Samples*, long long, int) () No symbol table info available. #2 0x08729ea5 in PluginServer::process_buffer(Samples**, long long, long long, long long, long long, int) () No symbol table info available. #3 0x084d393b in AAttachmentPoint::render(Samples*, int, long long, long long, long long) () No symbol table info available. #4 0x087efb03 in VirtualANode::render_as_module(Samples**, Samples*, long long, long long, long long) () No symbol table info available. #5 0x087efbaa in VirtualANode::render(Samples*, long long, long long, long long) () No symbol table info available. #6 0x087ede33 in VirtualAConsole::process_buffer(long long, long long) () No symbol table info available. #7 0x08504f84 in ARender::run() () No symbol table info available. #8 0x088f4881 in Thread::entrypoint(void*) () --Type <RET> for more, q to quit, c to continue without paging-- No symbol table info available. #9 0xf7e42328 in start_thread () from /lib/libpthread.so.0 No symbol table info available. #10 0xf6204f06 in clone () from /lib/libc.so.6 No symbol table info available. On Fri, Apr 21, 2023 at 5:08 AM Andrew Randrianasulu <randrianasulu(a)gmail.com> wrote: > > Thanks a lot, I am sorry I spend most of time watching all 7 episodes of Extrapolations (tv series) instead of doing anything at computer. > > So, it basically in "delay audio" plugin, this find narrows it quite a lot! > > пт, 21 апр. 2023 г., 04:04 Phyllis Smith <phylsmith2017(a)gmail.com>: >> >> Simplest demo to create "Delay Audio" plugin problem. I tested this with 4 random audio files or video files that included audio files and all crashed under this set of steps. So you can demo with any file but 1 small example audio file is attached. >> >> 1) load/replace file (or load to resources and then drag onto timeline) >> 2) drag the "Delay Audio" plugin to the audio track >> 3) start playing file from the beginning (or start anywhere) >> 4) without stopping first, use the "fast forward" icon in the main window to speed up the playing - CRASH >> >> ALTERNATIVE to "Delay Audio" is "F_adelay" which did not crash for me - set "delays" parameter to desired seconds and use "Apply". >> Next I will use the debugger to learn some more.

2 6

Forum
by Andrea paz 23 Apr '23

23 Apr '23

After the latest maintenance work on the server/forum, it looks cleaner, simplier to read, nicer to look at, and most importantly, I no longer have the problems of delayed post appearance that always plagued me. Great job, Sam! thank you.

2 1

"CinGG Summer of Code" Ernest Ibarra disappointment/failure
by Phyllis Smith 23 Apr '23

23 Apr '23

Last summer I put out an ad for a 10 week CinGG summer of code with the title of "Linux / C++ beginner / student" with an enticement of $400 each week for compiling, execution of well outlined steps, and make mods to the "burn" plugin to just add the gui variables. I thought someone might like the challenge and could put in a few hours each week (before and after "aging" plugin gui mods would serve as a guideline). Response from Ernest Ibarra was "he was interested" and had a "Mac Pro with Ubuntu and also had 3 years of C++ knowledge". Pointed him to the Manual on how to build CinGG and he sent me a screenshot showing a built version. So I assumed he was able to do a build which would mean he had to have done a compile?? Ibarra was paid $400 a week for 4 weeks just to experiment with coding but by the end of the 4th week he reported "having a really hard time compiling" just the plugin. I suggested he do a full compile with his modifications in (since supposedly he had compiled the whole program at least once, right?). Ernest Ibarra stopped replying to my emails after 4 weeks of PayPal payments by June 28th. Finally on July 19th, after receiving another 2 weeks (total $2400) I asked Ibarra to send me the code that he could not get to compile so I could compile it myself but never heard back. BIG DISAPPOINTMENT -- should not 3 years of C++ be enough for a competent programmer to work on a very small part of Cinelerra? So I am closing BT #62 as being a HUGE FAILURE -- at least with Ernest Ibarra.

5 7

The timelines move faster/slower when paste silence
by Stefan de Konink 16 Apr '23

16 Apr '23

Again a "Stefan is editing and gets annoyed email". We have had this discussion before. Project setup 1080p, audio 48kHz. video 1. video track 1080p25 audio 1. audio track from external recorder 96kHz. audio 2/3. audio matching video 48kHz When I cut (x) at a certain point, and then do shift-space, my audio and video looses synchronisation. Video when it happens. <http://download.stefan.konink.de/cinelerra-bugs/paste-space.mkv> So can this be resolved? When I would enable Settings > Align cursor on frames I don't have this issue. I would say, if Cinerella causes these kind of out-of-sync behavior, should we even have enable option (by default) not to edit at frame level? -- Stefan

3 4

DaVinchi Resolve on radeonsi
by Andrew Randrianasulu 13 Apr '23

13 Apr '23

https://chaos.social/@karolherbst/109982973566417227 "*karolherbst *@karolherbst <https://chaos.social/@karolherbst> DaVinci Resolve on #rusticl <https://chaos.social/tags/rusticl> :3 Big thanks to our EVoC student Antonio to implement the proper #OpenCL <https://chaos.social/tags/OpenCL> extensions to make it run (and me, because it also needs https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19855 ===== It was merged lately (around day ago) into mesa main git. I mean just opencl enable, not opencl <=> opengl interoperability. Still, this is interesting foundation stone. Too bad I am not a real developer for making any use of this!

2 1

Re: [Cin] libuuid patch
by Phyllis Smith 08 Apr '23

08 Apr '23

Changes checked into CinGG GIT are as follows: -- for clang-16 (termux) from Andrew mpeg2enc.c thirdparty/Makefile (using -std=gnu89 which may present a problem in 20 years?) libuuid-1.0.3.patch (modified to remove lines changed to be the same) -- from Andrea to have ffmpeg mimic Apple vendor in prores_ks.pro render format Thank you Andrew and Andrea (and when you have time, do a build to make sure I did not introduce any mistakes) ...Phyllis On Mon, Apr 3, 2023 at 9:45 PM Andrew Randrianasulu <randrianasulu(a)gmail.com> wrote: > Attached patch fix build of libuuid-1.0.3 on Termux current (with clang-16) > > Also attached patch I based my patch on, from NetBSD sources > > it tries to do more renaming, but I think I need another patch for > completing renaming, so this part is dropped ( hopefully only relevant on > Solaris) >

3 2

Re: [Cin] libuuid patch
by Phyllis Smith 06 Apr '23

06 Apr '23

Andrew, On Mon, Apr 3, 2023 at 9:45 PM Andrew Randrianasulu <randrianasulu(a)gmail.com> wrote: > Attached patch fix build of libuuid-1.0.3 on Termux current (with clang-16) > > Also attached patch I based my patch on, from NetBSD sources > > it tries to do more renaming, but I think I need another patch for > completing renaming, so this part is dropped ( hopefully only relevant on > Solaris) > By this do you mean "Rename random_get_bytes to avoid symbol name conflict on Solaris" which was in the patch for NetBSD and not carried over to the actual patch? So that the diff lines marked as @@ -319, @@-451, and @@-549 are not needed?

2 1

Re: [Cin] Revert c/c++ standards for thirdparty components
by Phyllis Smith 05 Apr '23

05 Apr '23

Andrew, this patch also built here with no problems. I "think" I have ways to test this some more for a couple of days too before checking into GIT. Thanks and glad you got TERMUX going again as it is sure convenient to have this working on your tablet. On Tue, Apr 4, 2023 at 12:29 AM Andrew Randrianasulu < randrianasulu(a)gmail.com> wrote: > Plus also one additional include for mpeg2enc (not tested, just fixes > build err with clang-16) > > with those patches cin builds, starts and play m2ts on aarch64/termux > again. > > вт, 4 апр. 2023 г., 08:48 Andrew Randrianasulu <randrianasulu(a)gmail.com>: > >> So clang-16 compiles them without erroring out >> >> >>

1 0

Re: [Cin] libuuid patch
by Phyllis Smith 05 Apr '23

05 Apr '23

Andrew, I built with the attached 0001-Fix-libuuid ... patch just fine. I will execute the binary for a couple of days and compare this patch with the netBSD similar one that was attached. There should not be any issues, but since I have no idea how "UUIDs generated by this library" are used in CinGG, will test some more. On Mon, Apr 3, 2023 at 9:45 PM Andrew Randrianasulu <randrianasulu(a)gmail.com> wrote: > Attached patch fix build of libuuid-1.0.3 on Termux current (with clang-16) > > Also attached patch I based my patch on, from NetBSD sources > > it tries to do more renaming, but I think I need another patch for > completing renaming, so this part is dropped ( hopefully only relevant on > Solaris) >

1 0

Re: [Cin] Checked updated x264 into GIT
by Andrew Randrianasulu 05 Apr '23

05 Apr '23

вт, 4 апр. 2023 г., 19:15 Phyllis Smith <phylsmith2017(a)gmail.com>: > Thanks Andrew -- I fixed the GIT checkin just now. The 4 bandages on my > fingers are causing a lot of typing errors and I need 1 more bandage too. > :( What happened? > according to this you just named files 264-something, not x264-something >> >> so ffmpeg build fail .... because it can't find x264 >> >> but I have more problems wit mjpegtools due to clang-16 defaulting to >> c++17 and libuuid also complains about flock function >> > Are these new problems due to the March library upgrades or previous > patches? > no, just termux is rolling distro, and it just managed to roll over such landmine as clang-16 ... >> ==== >> >> ../utils/fastintfns.h:5:27: error: ISO C++17 does not allow 'register' >> storage class specifier [-Wregister] >> static inline int >> intmax( register int x, register int y ) >> ^~~~~~~~~ >> ../utils/fastintfns.h:5:43: error: ISO C++17 >> does not allow 'register' storage class specifier [-Wregister] >> >> static inline int intmax( register int x, register int y ) >> >> ^~~~~~~~~ >> ../utils/fastintfns.h:10:27: error: ISO C++17 does not allow 'register' >> storage class specifier [-Wregister] >> static inline int intmin( >> register int x, register int y ) >> ^~~~~~~~~ >> ../utils/fastintfns.h:10:43: error: ISO C++17 does not >> allow 'register' storage class specifier [-Wregister] >> static >> inline int intmin( register int x, register int y ) >> ^~~~~~~~~ >> In file included from >> ratectl.cc:57: >> ../utils/fastintfns.h:5:27: error: ISO C++17 does not allow 'register' >> storage class specifier [-Wregister] >> static inline int >> intmax( register int x, register int y ) >> ^~~~~~~~~ >> ../utils/fastintfns.h:5:43: error: ISO C++17 >> does not allow 'register' storage class specifier [-Wregister] >> >> static inline int intmax( register int x, register int y ) >> >> ^~~~~~~~~ >> ../utils/fastintfns.h:10:27: error: ISO C++17 does not allow 'register' >> storage class specifier [-Wregister] >> static inline int intmin( >> register int x, register int y ) >> ^~~~~~~~~ >> ../utils/fastintfns.h:10:43: error: ISO C++17 does not >> allow 'register' storage class specifier [-Wregister] >> static >> inline int intmin( register int x, register int y ) >> >> ==== >> >> gen_uuid.c:297:10: error: call to undeclared function 'flock'; ISO C99 >> and later do not support implicit function declarations >> [-Wimplicit-function-declaration] >> while (flock(state_fd, LOCK_EX) < 0) { >> ^ >> gen_uuid.c:297:10: note: did you >> mean 'clock'? >> /data/data/com.termux/files/usr/include/time.h:96:9: note: 'clock' >> declared here clock_t clock(void); >> ^ >> >> gen_uuid.c:369:3: error: call to undeclared function 'flock'; ISO C99 and >> later do not support implicit function declarations >> [-Wimplicit-function-declaration] >> flock(state_fd, LOCK_UN); >> ^ >> 2 errors generated. >> >> ==== >> >> I am trying to apply netbsd patch to it, but no success so far >> >> >> вт, 4 апр. 2023 г., 01:03 Phyllis Smith <phylsmith2017(a)gmail.com>: >> >>> Tested an upgraded x264 stable version (r3106) with CinGG and checked >>> into GIT. Hopefully, Andrea has time to test a little too. ...Phyllis >>> >>

3 2