ah, two more patches, hopefully they will fix it for all.. On Wednesday, November 10, 2021, Andrew Randrianasulu < [email protected]> wrote:

On Wednesday, November 10, 2021, Andrew Randrianasulu < [email protected]> wrote:

...

strangely enough, this patch series builds with dv-stuff correctly now.. no build/link/runtime errors..

based on git commit 9d54c23d957

now lets try to define minimal patchset...

randrik19/ is a directory $ less randrik19/00 0001-constify-fileffmpeg.C.patch = nope 0002-TERMUX-ifdefs-in-guicast.patch = yes 0003-libzmpeg-TERMUX-ifdefs.patch = yes 0004-Add-libzmpeg-ifdef-in-fileac3-move-fireware-objects-.patch = yes 0005-configure.ac-change-for-non-x86-cflags.patch = yes 0006-configure.ac-ifdef-for-mjpegtools-on-non-x86.patch = yes 0007-add-termux-specific-include-path.patch = yes 0008-configure.ac-add-liconv-for-termux.patch = yes 0009-attempt-to-fix-_probe-function-memory-leak.patch = yes 0010-additional-XOpenIM-call-in-bcwindowbase.C.patch = yes 0011-libbthread-source-and-Makefile-patch-not-integrated-.patch = nope 0012-Additonal-mjpegtools-patches-for-termux-not-ifdefed.patch = nope 0013-More-ffmpeg-4.4-patches-renaming-B0-to-B0_-due-to-te.patch = nope 0014-add-libbthread-to-thirdparty-Makefile-do-not-install.patch = nope 0015-add-libbthread-to-configure.ac-termux-specific.patch = nope 0016-Fix-libzmpeg.h-x86-ifdef.patch = yes 0017-libzmpeg3-Makefile-add-libbthread-termux.patch = nope 0018-correctly-named-libbthread-patch.patch = nope 0019-Add-support-for-conditional-Termux-related-libraries.patch = nope 0020-Fix-mjpegtools-2.1.0.patch6-for-conditional-bthread..patch = nope 0021-Untested-add-aarch64-support-to-guicast-Makefile.patch = yes 0022-Add-timecode-to-output-mov-mxf.patch = nope 0023-Hack-pix_fmts-for-yuv4mpegpipe-muxer-encoder-allow-f.patch = nope 0024-Add-hackish-new-overlay-blend-modes.patch = nope 0025-Workaround_bugs-1-for-ffmpeg.C-decode.patch = nope 0026-Debug-add-printing-of-cpucount-from-ff_cpus.patch = nope 0027-Debug-add-cpu-count-printing-for-mwindow.C.patch = nope 0028-Hack-raise-cpu-cap-from-8-to-16-might-have-bad-effec.patch = nope 0029-Hacks-in-audiopulse-do-not-fix-audio-encoding-crash-.patch = nope 0030-HACK-try-to-speed-up-reverse-playback-for-i-only-for.patch = nope 0031-configure-script-specially-for-termux.patch = nope 0032-HACK-make-avi-demuxer-in-ffmpeg-use-cin-specific-see.patch = nope 0033-Add-x11-inc-autodetect-for-shuttle-target-using-pkg-.patch = nope 0034-additions-to-my_configure.patch = nope 0035-fix-mjpegtools-patch4-and-patch5-termux.patch = nope 0036-Try-to-fix-thirdparty-Makefile-libbthread-for-non-an.patch = nope 0037-Shut-up-warnings-on-nasm-2.15-in-x265.patch = nope 0038-Second-attempt-at-fixing-guicast-thread.h-on-non-and.patch = maybe 0039-Try-to-fix-shuttle_keys.h-includes-if-pkg-cfg-return.patch = nope 0040-fix-fireware-typo-in-cinelerra-Makefile.patch = yes 0041-more-fix-to-firewire-objs.patch = yes 0042-fix-typos-in-libdv-ifdefs-in-livevideo.patch = yes 0043-Add-two-more-libx264-patches-for-ffmpe-enable-avcint.patch = nope 0044-x265-3.5-patches-configure-update.patch = nope 0045-Add-x265-3.5-multilib-patch3-should-be-faster-at-com.patch = nope 0046-x265-patch3-was-malformed.patch = nope 0047-add-with-booby-and-without-ladspa-build-to-my_config.patch = nope 0048-sigh-i-deleted-make-in-x265_3.5.patch3-restored.patch = nope 0049-with-booby-demands-execinfo.h.patch = nope 0050-Hack-attempt-to-fix-last-flush-in-ffmpeg.c.patch = nope 0051-Attempt-at-fixing-armed-status-while-attaching-plugi.patch = nope 0052-Add-hidden-preference-armed_in_gang_mode.patch = nope 0053-Alter-RMB-attach-effect-behavior-on-disarmed-tracks.patch = nope 0054-Fix-mpeg2enc-build-install-was-broken-by-termux-seri.patch = maybe 0055-Initialize-more-in-vrender.C-constructor-cppcheck.patch = nope 0056-mem-and-resource-leaks-in-indexfile-indextate-cppche.patch = nope 0057-cppcheck-in-cursor.C.patch = nope 0058-cppcheck-in-resourcepixmap.C.patch = nope 0059-cppcheck-in-disabled-videoconfig.C.patch = nope 0060-cppcheck-in-disabled-recordengine.C.patch = nope 0061-Avoid-two-undefined-behaviors-in-vframe.C.patch = nope 0062-TMP-add-fsanitize-undefined-to-ldflags.patch = nope 0063-libaom-3.2.0-patches-update.patch = nope 0064-Can-be-built-with-dv-fireware.patch = nope

so, in summary...

try 2-10, 16, 21, 40-42

may be 38, 54

On Tuesday, November 16, 2021, Phyllis Smith via Cin < [email protected]> wrote:

Checked into GIT (from randrik18 so names different): 0003-TERMUX-ifdefs-in-guicast.patch + (now 0002 in randrik19) 0047-Second-attempt-at-fixing-guicast-thread.h-on-non-and.patch (now 0038 in randrik19) 0028-Untested-add-aarch64-support-to-guicast-Makefile.patch (now 0021 in randrik19) and this is still Untested! but looks good.

Andrew, this list is extremely helpful!

...

now lets try to define minimal patchset...

randrik19/ is a directory $ less

0002-TERMUX-ifdefs-in-guicast.patch = yes

BUT what is the purpose of changing SHMGET to SHM_GET? Was there a conflict name in android?

yes.... Very minimal testing on both 64bit and 32bit -- mostly just looked at

differences to see if they made sense.

*Andrea,* a 32-bit image is at: https://cinelerra-gg.org/ download/testing/cin-i386.AppImage

On Tuesday, November 16, 2021, Andrea paz <[email protected]> wrote:

- Tried building the new git. CinGG starts and works normally, but on terminal I get endless warnings like:

sounds like (new) set of lv2 plugins was installed on system...

[...] error: failed to open file /usr/lib/libcord.so/manifest.ttl (Not a directory) lilv_world_load_file(): error: Error loading file `file:///usr/lib/libcord.so/manifest.ttl' lilv_world_load_bundle(): error: Error reading file:///usr/lib/libcord.so/manifest.ttl error: failed to open file /usr/lib/libcord.so.1/manifest.ttl (Not a directory) lilv_world_load_file(): error: Error loading file `file:///usr/lib/libcord.so.1/manifest.ttl' lilv_world_load_bundle(): error: Error reading file:///usr/lib/libcord.so.1/manifest.ttl error: failed to open file /usr/lib/libcord.so.1.4.0/manifest.ttl (Not a directory) lilv_world_load_file(): error: Error loading file `file:///usr/lib/libcord.so.1.4.0/manifest.ttl' lilv_world_load_bundle(): error: Error reading file:///usr/lib/libcord.so.1.4.0/manifest.ttl error: failed to open file /usr/lib/libgc.so/manifest.ttl (Not a directory) lilv_world_load_file(): error: Error loading file `file:///usr/lib/libgc.so/manifest.ttl' lilv_world_load_bundle(): error: Error reading file:///usr/lib/libgc.so/manifest.ttl error: failed to open file /usr/lib/libgc.so.1/manifest.ttl (Not a directory) lilv_world_load_file(): error: Error loading file `file:///usr/lib/libgc.so.1/manifest.ttl' lilv_world_load_bundle(): error: Error reading file:///usr/lib/libgc.so.1/manifest.ttl error: failed to open file /usr/lib/libgccpp.so/manifest.ttl (Not a directory) lilv_world_load_file(): error: Error loading file `file:///usr/lib/libgccpp.so/manifest.ttl' lilv_world_load_bundle(): error: Error reading file:///usr/lib/libgccpp.so/manifest.ttl build ladspa plugin index for: /home/paz/cinelerra5/ cinelerra-5.1/bin/ladspa

- Tried the new cin-i386.AppImage in Debian 11 32-bit on a virtual machine: CinGG does not start with the usual error on libIlmImf-2_2.so.22

yeah, sounds like forcing ilmbase/openexr to be built in statically is way to go...

As always: my mistake. I put a typo in the path of lv2 (7 instead of /). Now everything is OK. Thank you.

On Wednesday, November 17, 2021, Phyllis Smith via Cin < [email protected]> wrote:

Andrew,

...

yeah, sounds like forcing ilmbase/openexr to be built in statically is way to go...

The ilmbase IlmImf was still in the i386 AppImage and worked for Andrea because it is a virtual system (at least I think so), as opposed to Debian Bullseye (a real system) where the IlmImf so file still cause an immediate crash. The only way I can get rid of it in the build is to manually edit configure.ac and comment it out. On the configure line, I have tried adding --without-openexr --without-ilmbase but it does not get rid of it. Also, tried --without-ilmBase and disable-ilmbase and I just keep getting illegal option. But now I am side-tracked again.

try to delete -dev version of openexr/ilmbase packages from system you build appimage on...?

*Question:* before I check into GIT the following termux related mods in *CONFIGURE.AC <http://CONFIGURE.AC>,* does libbthread need ifdefs around it? It is not causing a problem when I compile, but I don't know what will happen when all of the rest of the libbthread patches are added.

I think yes check 0036-Try-to-fix-thirdparty-Makefile-libbthread-for-non-an.patch in randrik19

0007-configure.ac-change-for-non-x86-cflags.patch ************randrik19/0005 0009-configure.ac-ifdef-for-mjpegtools-on-non-x86.patch ********randrik19/0006 0012-add-termux-specific-include-path.patch *****randrik19/0007 0013-configure.ac-add-liconv-for-termux.patch *****randrik19/0008 0021-add-libbthread-to-configure.ac-termux-specific.patch **************** Question? *****randrik19/0015 0026-Add-support-for-conditional-Termux-related-libraries.patch *****randrik19/0019

Also, I have to test on i386 tomorrow before checkin. Also Also, I forgot to download randrik19 but did so just now.

Georgy, Thanks so much for the analysis and advice! it gives us a direction to go. I will do some checking based on my limited ability. ...Phyllis On Tue, Nov 16, 2021 at 11:33 PM Georgy Salnikov via Cin < [email protected]> wrote:

Andrew, Phyllis,

On Wed, 17 Nov 2021, Andrew Randrianasulu via Cin wrote:

...

...

...

yeah, sounds like forcing ilmbase/openexr to be built in statically is way to go...

Bullseye (a real system) where the IlmImf so file still cause an immediate crash. The only way I can get rid of it in the build is to manually edit configure.ac and comment it out. On the configure line, I have tried

try to delete -dev version of openexr/ilmbase packages from system you

I am almost sure the source of the openexr problem in cin must be an incompatibility between the openexr versions from thirdparty and that installed in the system: these two may be different, and while compiling cinelerra takes headers from /usr/include/OpenEXR, but then links with the library from thirdparty.

Here I have found my personal memo of how to compile cinelerra-4.5 (HV version):

./configure

inspect FLAGS in *_config, *akefile on -I/usr/include/OpenEXR change -I/usr/include/OpenEXR to -I$(includedir)/OpenEXR

make |& tee log

This means, after configure but before make I had to manually grep the whole Cinelerra tree like this

fgrep -l I/usr/include/OpenEXR `find . -name '*akefile' -print`

Then edit in all that autogenerated makefiles all the references to /usr/include/OpenEXR to point to the thirdparty's location, everything manually, and only then execute make.

In my opinion, even the recent Cinelerra-HV still requires such manual editing. But how in looks in Cin-GG, I did never inspect.

Of course, compilation of Cinelerra with openexr will succeed if by chance the version of openexr in the system is accidentally the same as that in thirdparty.

I am sorry, I have no time just now. Could somebody try to grep Cin-GG tree after configure, are there some refs to headers from /usr/include which could be conflicting with thirdparty or not?

Georgy

_______________________________________________________________________________

Georgy Salnikov NMR Group Novosibirsk Institute of Organic Chemistry Lavrentjeva, 9, 630090 Novosibirsk, Russia Phone +7-383-3307864 Email [email protected]

_______________________________________________________________________________

-- Cin mailing list [email protected] https://lists.cinelerra-gg.org/mailman/listinfo/cin

On Wed, 17 Nov 2021, Georgy Salnikov via Cin wrote:

inspect FLAGS in *_config, *akefile on -I/usr/include/OpenEXR change -I/usr/include/OpenEXR to -I$(includedir)/OpenEXR

fgrep -l I/usr/include/OpenEXR `find . -name '*akefile' -print`

I have grepped throuhg the build tree of some cingg git version of May, OpenEXR seem to look OK. One can find -I$(includedir)/OpenEXR, and something like thirdparty/openexr-2.4.1/usr/include/OpenEXR. But direct refs to openexr system headers are not used. There is also such a ref in the top level global_config and top level Makefile under CFG_CFLAGS but these flags perhaps are also not used anywhere. Perhaps somebody who is more familiar with the configuration sequence of the cingg build system could inspect the flags more carefully... Just to be sure, or if the compiled cingg crashes because of openexr on some system, it is possible to rename temporarily /usr/include/OpenEXR and the relevant .so files while building cin, so that only the thirdparty version will be found. _______________________________________________________________________________ Georgy Salnikov NMR Group Novosibirsk Institute of Organic Chemistry Lavrentjeva, 9, 630090 Novosibirsk, Russia Phone +7-383-3307864 Email [email protected] _______________________________________________________________________________

If you are modifying configure.ac anyway, I suggest to comment out the line about "no-unknown-warning-option" ? It causes no end of useless warnings. Also, the detection of the number of cpus (about line 143) is done by a very specific Linux method /proc/cpuinfo) , which does not work on FreeBSD and macOS. It would probably be better to replace it with "nproc", which can be made available on FreeBSD via an alias, and on macOS after installing some stuff with the "HomeBrew" package manager. MatN

On Wednesday, November 17, 2021, Phyllis Smith via Cin < [email protected]> wrote:

*MatN,*

...

If you are modifying configure.ac anyway, I suggest to comment out the line about "no-unknown-warning-option" ? It causes no end of useless warnings.

Good suggestion. It will be done.

...

Also, the detection of the number of cpus (about line 143) is done by a very specific Linux method /proc/cpuinfo) , which does not work on FreeBSD and macOS.

So I do an exact replacement from: CPUS=`grep -c "^proc" /proc/cpuinfo` to: CPUS=`nproc` Is that correct?

*Andrew,* For 0031-configure-script-specially-for-termux.patch is it acceptable to use the filename of "configure_termux" instead of "my_configure"? I was already thinking about adding "configure_32-bit" to make it easier when I create AppImages.

sure, you can also put it in blds directory like all other build scripts...

Changes to configure.ac (and 1 small one to autogen.sh) have been checked into GIT to accommodate Termux/Android + Mat's suggestion to remove "no-unknown-warning-option". Anyone who has time and can build, please test. I had no problems on 64 or 32-bit. Thanks, Phyllis

On Thursday, November 18, 2021, mnieuw--- via Cin < [email protected]> wrote:

...

On Wed, 17 Nov 2021 01:07:05 -0700 Phyllis Smith via Cin <[email protected]> wrote:

...

...

Also, the detection of the number of cpus (about line 143) is done by a very specific Linux method /proc/cpuinfo) , which does not work on FreeBSD and macOS.

So I do an exact replacement from: CPUS=`grep -c "^proc" /proc/cpuinfo` to: CPUS=`nproc` Is that correct?

Yes, please change it to CPUS=nproc

The single quotes are not needed, those are to prevent shell expansion within the quotes.

nproc properly reports number of threads on macOS (after installing development tools which are needed anyway), and FreeBSD via an alias.

on linux it should works since coreutils 8.1 (nov 2009?)

FreeBSD does not build as-is, e.g. it uses gmake which doesn't exist on FreeSD version 13.

https://freebsd.pkgs.org/13/freebsd-amd64/gmake-4.3_2.txz.html ?

nproc needs an alias but that is easy done in the bsd.bld file. FreeBSD is also very limited in included options, like is excludes all of thirdparty. I wonder if that is still relevant, for example I did see LV2 plugins for FreeBSD. Was it downloaded a lot when there still was a FreeBSD version, or put it another way, is the effort useful to resurrect it?

All builds report during configure time that with gcc -V or clang -V the upper case -V is not known; it probably tries to retrieve the version number. In both cases, lower case -v gives the version number. I did see a while ago a msg about outdated configure.ac . Maybe that would help, I'll see what I can find.

I don't know whether freeBSD or macOS would be good platforms for CinGG, but (trying) to build for other platforms is a good way to weed out hidden problems.

macos apparently can run x11 apps: https://www.xquartz.org/releases/index.html not sure about sound...

MatN

-- Cin mailing list [email protected] https://lists.cinelerra-gg.org/mailman/listinfo/cin

Checked into GIT just now: 1) MatN improvement to configure.ac with CPUS=nproc 2) more Andrew mostly related to Termux on Android from RANDRIK19 - 0003-libzmpeg-TERMUX-ifdefs.patch (0004 randrik18) - added new file blds/termux.bld which was previously named my_config in the following patches: a) 0031-configure-script-specially-for-termux.patch b) 0034-additions-to-my_configure.patch c) 0047-add-with-booby-and-without-ladspa-build-to-my_config.patc d) 0049-with-booby-demands-execinfo.h.patch e) 0064-Can-be-built-with-dv-fireware.patch - 0040-fix-fireware-typo-in-cinelerra-Makefile.patch (0049 randrik18) - 0041-more-fix-to-firewire-objs.patch (0050 randrik18) *Andrew PLEASE CHECK THESE:* - 0004-Add-libzmpeg-ifdef-in-fileac3-move-fireware-objects-.patch BUT changed #if HAVE_LIBZMPEG to #ifdef HAVE_LIBZMPEG BECAUSE that is what file.C and formatpopup.C used. ??? - 0066-Fix-libzmpeg3-x86-ifdefs-again.patch (randrik18) BUT CONFLICT 0016-Fix-libzmpeg.h-x86-ifdef.patch (randrik19) BECAUSE that is the way it was done in commercials.h, dbwindow.h, and mediadb.h. ??? - 0006-more-ifdefs-for-file.C-ac3-demand-libzmpeg.patch FROM RANDRIK18 has no equivalent in RANDRIK19 so did not add??? I think that is all added -- trying to finish up TERMUX related mods. Yes, please change it to

CPUS=nproc

Done.

FreeBSD does not build as-is, e.g. it uses gmake which doesn't exist on FreeSD version 13. nproc needs an alias but that is easy done in the bsd.bld file. FreeBSD is also very limited in included options, like is excludes all of thirdparty. I wonder if that is still relevant, for example I did see LV2 plugins for FreeBSD. Was it downloaded a lot when there still was a FreeBSD version, or put it another way, is the effort useful to resurrect it?

Not sure if need to update to version 13 but version 12 of BSD should still work. It would be nice for someone to look into this as there was a contact there who was tasked with working it. The last time contacted, he did not respond so may have moved on.

All builds report during configure time that with gcc -V or clang -V the upper case -V is not known; it probably tries to retrieve the version number. In both cases, lower case -v gives the version number. I did see a while ago a msg about outdated configure.ac . Maybe that would help, I'll see what I can find.

Thank you -- let us know if you find anything that can be improved.

I don't know whether freeBSD or macOS would be good platforms for CinGG, but (trying) to build for other platforms is a good way to weed out hidden problems.

Absolutely agree!

On Thu, 18 Nov 2021, mnieuw--- via Cin wrote:

...

...

CPUS=`nproc` Is that correct?

Yes, please change it to CPUS=nproc

The single quotes are not needed, those are to prevent shell expansion within the quotes.

Attention, it is absolutely not the same, whether in backquotes or unquoted. Backquotes mean that the shell tries to execute the quoted contents as a command and substitute the standard output of that command as the result. Please compare and see the difference: mri.sge:~>nproc 2 mri.sge:~>CPUS=nproc mri.sge:~>echo $CPUS nproc mri.sge:~>CPUS=`nproc` mri.sge:~>echo $CPUS 2 (there are 2 cores on the test computer, mri.sge:~> being the shell prompt) _______________________________________________________________________________ Georgy Salnikov NMR Group Novosibirsk Institute of Organic Chemistry Lavrentjeva, 9, 630090 Novosibirsk, Russia Phone +7-383-3307864 Email [email protected] _______________________________________________________________________________

Checked into GIT this morning.

0009-attempt-to-fix-_probe-function-memory-leak.patch = yes 0010-additional-XOpenIM-call-in-bcwindowbase.C.patch = yes 0042-fix-typos-in-libdv-ifdefs-in-livevideo.patch = yes

now lets try to define minimal patchset...

so, in summary...

try 2-10, 16, 21, 40-42

may be 38, 54

Minimal patchset has all been checked into GIT after very minimal testing on my part. EXCEPT for 54 because there are very many mjpegtools patches that precede it and I can not tell if there will be user impact for non-Termux builds and not sure how to adequately test. *Concerns for Andrew *on these 3 sets of patches: ? 0009 ? is freeing this frame always safe or what if it does not exist? I am thinking it would have already crashed if it is empty. ? 0010 ? does XSetLocaleModifiers affect future locale settings that might cause issues ? or is it just temporary? ? 0042 ? I changed all of the "if HAVE_DVf" to "ifdef HAVE_LIBDV" because that is what is used everywhere else ? ? 0054 ? + 12, 14, 20, 35 ? is there a really good set of tests for patches in mjpegtools which I have not added in yet ?

UPD: Phyllis, can we add 0022-Add-timecode-to-output-mov-mxf.patch ifdefed to 0, so it will be not lost? I think DVR (Resolve) has similar functionality, just not sure how to wire it as user-selected option... https://www.reddit.com/r/davinciresolve/comments/11mty4y/sony_timecode_how_t... ср, 10 нояб. 2021 г., 18:39 Andrew Randrianasulu <[email protected]>:

On Wednesday, November 10, 2021, Andrew Randrianasulu < [email protected]> wrote:

...

strangely enough, this patch series builds with dv-stuff correctly now.. no build/link/runtime errors..

based on git commit 9d54c23d957

now lets try to define minimal patchset...

randrik19/ is a directory $ less randrik19/00 0001-constify-fileffmpeg.C.patch = nope 0002-TERMUX-ifdefs-in-guicast.patch = yes 0003-libzmpeg-TERMUX-ifdefs.patch = yes 0004-Add-libzmpeg-ifdef-in-fileac3-move-fireware-objects-.patch = yes 0005-configure.ac-change-for-non-x86-cflags.patch = yes 0006-configure.ac-ifdef-for-mjpegtools-on-non-x86.patch = yes 0007-add-termux-specific-include-path.patch = yes 0008-configure.ac-add-liconv-for-termux.patch = yes 0009-attempt-to-fix-_probe-function-memory-leak.patch = yes 0010-additional-XOpenIM-call-in-bcwindowbase.C.patch = yes 0011-libbthread-source-and-Makefile-patch-not-integrated-.patch = nope 0012-Additonal-mjpegtools-patches-for-termux-not-ifdefed.patch = nope 0013-More-ffmpeg-4.4-patches-renaming-B0-to-B0_-due-to-te.patch = nope 0014-add-libbthread-to-thirdparty-Makefile-do-not-install.patch = nope 0015-add-libbthread-to-configure.ac-termux-specific.patch = nope 0016-Fix-libzmpeg.h-x86-ifdef.patch = yes 0017-libzmpeg3-Makefile-add-libbthread-termux.patch = nope 0018-correctly-named-libbthread-patch.patch = nope 0019-Add-support-for-conditional-Termux-related-libraries.patch = nope 0020-Fix-mjpegtools-2.1.0.patch6-for-conditional-bthread..patch = nope 0021-Untested-add-aarch64-support-to-guicast-Makefile.patch = yes 0022-Add-timecode-to-output-mov-mxf.patch = nope 0023-Hack-pix_fmts-for-yuv4mpegpipe-muxer-encoder-allow-f.patch = nope 0024-Add-hackish-new-overlay-blend-modes.patch = nope 0025-Workaround_bugs-1-for-ffmpeg.C-decode.patch = nope 0026-Debug-add-printing-of-cpucount-from-ff_cpus.patch = nope 0027-Debug-add-cpu-count-printing-for-mwindow.C.patch = nope 0028-Hack-raise-cpu-cap-from-8-to-16-might-have-bad-effec.patch = nope 0029-Hacks-in-audiopulse-do-not-fix-audio-encoding-crash-.patch = nope 0030-HACK-try-to-speed-up-reverse-playback-for-i-only-for.patch = nope 0031-configure-script-specially-for-termux.patch = nope 0032-HACK-make-avi-demuxer-in-ffmpeg-use-cin-specific-see.patch = nope 0033-Add-x11-inc-autodetect-for-shuttle-target-using-pkg-.patch = nope 0034-additions-to-my_configure.patch = nope 0035-fix-mjpegtools-patch4-and-patch5-termux.patch = nope 0036-Try-to-fix-thirdparty-Makefile-libbthread-for-non-an.patch = nope 0037-Shut-up-warnings-on-nasm-2.15-in-x265.patch = nope 0038-Second-attempt-at-fixing-guicast-thread.h-on-non-and.patch = maybe 0039-Try-to-fix-shuttle_keys.h-includes-if-pkg-cfg-return.patch = nope 0040-fix-fireware-typo-in-cinelerra-Makefile.patch = yes 0041-more-fix-to-firewire-objs.patch = yes 0042-fix-typos-in-libdv-ifdefs-in-livevideo.patch = yes 0043-Add-two-more-libx264-patches-for-ffmpe-enable-avcint.patch = nope 0044-x265-3.5-patches-configure-update.patch = nope 0045-Add-x265-3.5-multilib-patch3-should-be-faster-at-com.patch = nope 0046-x265-patch3-was-malformed.patch = nope 0047-add-with-booby-and-without-ladspa-build-to-my_config.patch = nope 0048-sigh-i-deleted-make-in-x265_3.5.patch3-restored.patch = nope 0049-with-booby-demands-execinfo.h.patch = nope 0050-Hack-attempt-to-fix-last-flush-in-ffmpeg.c.patch = nope 0051-Attempt-at-fixing-armed-status-while-attaching-plugi.patch = nope 0052-Add-hidden-preference-armed_in_gang_mode.patch = nope 0053-Alter-RMB-attach-effect-behavior-on-disarmed-tracks.patch = nope 0054-Fix-mpeg2enc-build-install-was-broken-by-termux-seri.patch = maybe 0055-Initialize-more-in-vrender.C-constructor-cppcheck.patch = nope 0056-mem-and-resource-leaks-in-indexfile-indextate-cppche.patch = nope 0057-cppcheck-in-cursor.C.patch = nope 0058-cppcheck-in-resourcepixmap.C.patch = nope 0059-cppcheck-in-disabled-videoconfig.C.patch = nope 0060-cppcheck-in-disabled-recordengine.C.patch = nope 0061-Avoid-two-undefined-behaviors-in-vframe.C.patch = nope 0062-TMP-add-fsanitize-undefined-to-ldflags.patch = nope 0063-libaom-3.2.0-patches-update.patch = nope 0064-Can-be-built-with-dv-fireware.patch = nope

so, in summary...

try 2-10, 16, 21, 40-42

may be 38, 54

вт, 6 июн. 2023 г., 21:17 Phyllis Smith <[email protected]>:

OK, I will check that out and let you know. This was a "nope" in your original list; are there others that I missed or you want in?

I'll recheck today. This is not really important patch (no one requested this feature!) but with possibility of hdd/tablet crash & burn (drown) I think I better to put it into codebase, because I tend to forgot even things I wrote some time ago!

On Tue, Jun 6, 2023 at 11:46 AM Andrew Randrianasulu via Cin < [email protected]> wrote:

...

UPD:

Phyllis, can we add

0022-Add-timecode-to-output-mov-mxf.patch

ifdefed to 0, so it will be not lost?

I think DVR (Resolve) has similar functionality, just not sure how to wire it as user-selected option...

https://www.reddit.com/r/davinciresolve/comments/11mty4y/sony_timecode_how_t...

ср, 10 нояб. 2021 г., 18:39 Andrew Randrianasulu <[email protected]

...

:

...

On Wednesday, November 10, 2021, Andrew Randrianasulu < [email protected]> wrote:

...

strangely enough, this patch series builds with dv-stuff correctly now.. no build/link/runtime errors..

based on git commit 9d54c23d957

now lets try to define minimal patchset...

randrik19/ is a directory $ less randrik19/00 0001-constify-fileffmpeg.C.patch = nope 0002-TERMUX-ifdefs-in-guicast.patch = yes 0003-libzmpeg-TERMUX-ifdefs.patch = yes 0004-Add-libzmpeg-ifdef-in-fileac3-move-fireware-objects-.patch = yes 0005-configure.ac-change-for-non-x86-cflags.patch = yes 0006-configure.ac-ifdef-for-mjpegtools-on-non-x86.patch = yes 0007-add-termux-specific-include-path.patch = yes 0008-configure.ac-add-liconv-for-termux.patch = yes 0009-attempt-to-fix-_probe-function-memory-leak.patch = yes 0010-additional-XOpenIM-call-in-bcwindowbase.C.patch = yes 0011-libbthread-source-and-Makefile-patch-not-integrated-.patch = nope 0012-Additonal-mjpegtools-patches-for-termux-not-ifdefed.patch = nope 0013-More-ffmpeg-4.4-patches-renaming-B0-to-B0_-due-to-te.patch = nope 0014-add-libbthread-to-thirdparty-Makefile-do-not-install.patch = nope 0015-add-libbthread-to-configure.ac-termux-specific.patch = nope 0016-Fix-libzmpeg.h-x86-ifdef.patch = yes 0017-libzmpeg3-Makefile-add-libbthread-termux.patch = nope 0018-correctly-named-libbthread-patch.patch = nope 0019-Add-support-for-conditional-Termux-related-libraries.patch = nope 0020-Fix-mjpegtools-2.1.0.patch6-for-conditional-bthread..patch = nope 0021-Untested-add-aarch64-support-to-guicast-Makefile.patch = yes 0022-Add-timecode-to-output-mov-mxf.patch = nope 0023-Hack-pix_fmts-for-yuv4mpegpipe-muxer-encoder-allow-f.patch = nope 0024-Add-hackish-new-overlay-blend-modes.patch = nope 0025-Workaround_bugs-1-for-ffmpeg.C-decode.patch = nope 0026-Debug-add-printing-of-cpucount-from-ff_cpus.patch = nope 0027-Debug-add-cpu-count-printing-for-mwindow.C.patch = nope 0028-Hack-raise-cpu-cap-from-8-to-16-might-have-bad-effec.patch = nope 0029-Hacks-in-audiopulse-do-not-fix-audio-encoding-crash-.patch = nope 0030-HACK-try-to-speed-up-reverse-playback-for-i-only-for.patch = nope 0031-configure-script-specially-for-termux.patch = nope 0032-HACK-make-avi-demuxer-in-ffmpeg-use-cin-specific-see.patch = nope 0033-Add-x11-inc-autodetect-for-shuttle-target-using-pkg-.patch = nope 0034-additions-to-my_configure.patch = nope 0035-fix-mjpegtools-patch4-and-patch5-termux.patch = nope 0036-Try-to-fix-thirdparty-Makefile-libbthread-for-non-an.patch = nope 0037-Shut-up-warnings-on-nasm-2.15-in-x265.patch = nope 0038-Second-attempt-at-fixing-guicast-thread.h-on-non-and.patch = maybe 0039-Try-to-fix-shuttle_keys.h-includes-if-pkg-cfg-return.patch = nope 0040-fix-fireware-typo-in-cinelerra-Makefile.patch = yes 0041-more-fix-to-firewire-objs.patch = yes 0042-fix-typos-in-libdv-ifdefs-in-livevideo.patch = yes 0043-Add-two-more-libx264-patches-for-ffmpe-enable-avcint.patch = nope 0044-x265-3.5-patches-configure-update.patch = nope 0045-Add-x265-3.5-multilib-patch3-should-be-faster-at-com.patch = nope 0046-x265-patch3-was-malformed.patch = nope 0047-add-with-booby-and-without-ladspa-build-to-my_config.patch = nope 0048-sigh-i-deleted-make-in-x265_3.5.patch3-restored.patch = nope 0049-with-booby-demands-execinfo.h.patch = nope 0050-Hack-attempt-to-fix-last-flush-in-ffmpeg.c.patch = nope 0051-Attempt-at-fixing-armed-status-while-attaching-plugi.patch = nope 0052-Add-hidden-preference-armed_in_gang_mode.patch = nope 0053-Alter-RMB-attach-effect-behavior-on-disarmed-tracks.patch = nope 0054-Fix-mpeg2enc-build-install-was-broken-by-termux-seri.patch = maybe 0055-Initialize-more-in-vrender.C-constructor-cppcheck.patch = nope 0056-mem-and-resource-leaks-in-indexfile-indextate-cppche.patch = nope 0057-cppcheck-in-cursor.C.patch = nope 0058-cppcheck-in-resourcepixmap.C.patch = nope 0059-cppcheck-in-disabled-videoconfig.C.patch = nope 0060-cppcheck-in-disabled-recordengine.C.patch = nope 0061-Avoid-two-undefined-behaviors-in-vframe.C.patch = nope 0062-TMP-add-fsanitize-undefined-to-ldflags.patch = nope 0063-libaom-3.2.0-patches-update.patch = nope 0064-Can-be-built-with-dv-fireware.patch = nope

so, in summary...

try 2-10, 16, 21, 40-42

may be 38, 54

-- Cin mailing list [email protected] https://lists.cinelerra-gg.org/mailman/listinfo/cin

вс, 11 июн. 2023 г., 18:35 Phyllis Smith <[email protected]>:

Andrew, I checked in the "if defed 0" today so it does not get lost (along with your latest dvdcreate.C deinterlaced fix). Could you verify that I did the "if def'ed 0" correctly and put the correct lines within?

Checked visually by comparing to my old patch seems to be ok, thanks! Also, I hope things will still work as intended for big (local) renderfarm DVD rendering .....

On Tue, Jun 6, 2023 at 12:20 PM Andrew Randrianasulu < [email protected]> wrote:

...

вт, 6 июн. 2023 г., 21:17 Phyllis Smith <[email protected]>:

...

OK, I will check that out and let you know. This was a "nope" in your original list; are there others that I missed or you want in?

I'll recheck today. This is not really important patch (no one requested this feature!) but with possibility of hdd/tablet crash & burn (drown) I think I better to put it into codebase, because I tend to forgot even things I wrote some time ago!

...

On Tue, Jun 6, 2023 at 11:46 AM Andrew Randrianasulu via Cin < [email protected]> wrote:

...

UPD:

Phyllis, can we add

0022-Add-timecode-to-output-mov-mxf.patch

ifdefed to 0, so it will be not lost?

I think DVR (Resolve) has similar functionality, just not sure how to wire it as user-selected option...

https://www.reddit.com/r/davinciresolve/comments/11mty4y/sony_timecode_how_t...

ср, 10 нояб. 2021 г., 18:39 Andrew Randrianasulu < [email protected]>:

...

On Wednesday, November 10, 2021, Andrew Randrianasulu < [email protected]> wrote:

...

strangely enough, this patch series builds with dv-stuff correctly now.. no build/link/runtime errors..

based on git commit 9d54c23d957

now lets try to define minimal patchset...

randrik19/ is a directory $ less randrik19/00 0001-constify-fileffmpeg.C.patch = nope 0002-TERMUX-ifdefs-in-guicast.patch = yes 0003-libzmpeg-TERMUX-ifdefs.patch = yes 0004-Add-libzmpeg-ifdef-in-fileac3-move-fireware-objects-.patch = yes 0005-configure.ac-change-for-non-x86-cflags.patch = yes 0006-configure.ac-ifdef-for-mjpegtools-on-non-x86.patch = yes 0007-add-termux-specific-include-path.patch = yes 0008-configure.ac-add-liconv-for-termux.patch = yes 0009-attempt-to-fix-_probe-function-memory-leak.patch = yes 0010-additional-XOpenIM-call-in-bcwindowbase.C.patch = yes 0011-libbthread-source-and-Makefile-patch-not-integrated-.patch = nope 0012-Additonal-mjpegtools-patches-for-termux-not-ifdefed.patch = nope 0013-More-ffmpeg-4.4-patches-renaming-B0-to-B0_-due-to-te.patch = nope 0014-add-libbthread-to-thirdparty-Makefile-do-not-install.patch = nope 0015-add-libbthread-to-configure.ac-termux-specific.patch = nope 0016-Fix-libzmpeg.h-x86-ifdef.patch = yes 0017-libzmpeg3-Makefile-add-libbthread-termux.patch = nope 0018-correctly-named-libbthread-patch.patch = nope 0019-Add-support-for-conditional-Termux-related-libraries.patch = nope 0020-Fix-mjpegtools-2.1.0.patch6-for-conditional-bthread..patch = nope 0021-Untested-add-aarch64-support-to-guicast-Makefile.patch = yes 0022-Add-timecode-to-output-mov-mxf.patch = nope 0023-Hack-pix_fmts-for-yuv4mpegpipe-muxer-encoder-allow-f.patch = nope 0024-Add-hackish-new-overlay-blend-modes.patch = nope 0025-Workaround_bugs-1-for-ffmpeg.C-decode.patch = nope 0026-Debug-add-printing-of-cpucount-from-ff_cpus.patch = nope 0027-Debug-add-cpu-count-printing-for-mwindow.C.patch = nope 0028-Hack-raise-cpu-cap-from-8-to-16-might-have-bad-effec.patch = nope 0029-Hacks-in-audiopulse-do-not-fix-audio-encoding-crash-.patch = nope 0030-HACK-try-to-speed-up-reverse-playback-for-i-only-for.patch = nope 0031-configure-script-specially-for-termux.patch = nope 0032-HACK-make-avi-demuxer-in-ffmpeg-use-cin-specific-see.patch = nope 0033-Add-x11-inc-autodetect-for-shuttle-target-using-pkg-.patch = nope 0034-additions-to-my_configure.patch = nope 0035-fix-mjpegtools-patch4-and-patch5-termux.patch = nope 0036-Try-to-fix-thirdparty-Makefile-libbthread-for-non-an.patch = nope 0037-Shut-up-warnings-on-nasm-2.15-in-x265.patch = nope 0038-Second-attempt-at-fixing-guicast-thread.h-on-non-and.patch = maybe 0039-Try-to-fix-shuttle_keys.h-includes-if-pkg-cfg-return.patch = nope 0040-fix-fireware-typo-in-cinelerra-Makefile.patch = yes 0041-more-fix-to-firewire-objs.patch = yes 0042-fix-typos-in-libdv-ifdefs-in-livevideo.patch = yes 0043-Add-two-more-libx264-patches-for-ffmpe-enable-avcint.patch = nope 0044-x265-3.5-patches-configure-update.patch = nope 0045-Add-x265-3.5-multilib-patch3-should-be-faster-at-com.patch = nope 0046-x265-patch3-was-malformed.patch = nope 0047-add-with-booby-and-without-ladspa-build-to-my_config.patch = nope 0048-sigh-i-deleted-make-in-x265_3.5.patch3-restored.patch = nope 0049-with-booby-demands-execinfo.h.patch = nope 0050-Hack-attempt-to-fix-last-flush-in-ffmpeg.c.patch = nope 0051-Attempt-at-fixing-armed-status-while-attaching-plugi.patch = nope 0052-Add-hidden-preference-armed_in_gang_mode.patch = nope 0053-Alter-RMB-attach-effect-behavior-on-disarmed-tracks.patch = nope 0054-Fix-mpeg2enc-build-install-was-broken-by-termux-seri.patch = maybe 0055-Initialize-more-in-vrender.C-constructor-cppcheck.patch = nope 0056-mem-and-resource-leaks-in-indexfile-indextate-cppche.patch = nope 0057-cppcheck-in-cursor.C.patch = nope 0058-cppcheck-in-resourcepixmap.C.patch = nope 0059-cppcheck-in-disabled-videoconfig.C.patch = nope 0060-cppcheck-in-disabled-recordengine.C.patch = nope 0061-Avoid-two-undefined-behaviors-in-vframe.C.patch = nope 0062-TMP-add-fsanitize-undefined-to-ldflags.patch = nope 0063-libaom-3.2.0-patches-update.patch = nope 0064-Can-be-built-with-dv-fireware.patch = nope

so, in summary...

try 2-10, 16, 21, 40-42

may be 38, 54

-- Cin mailing list [email protected] https://lists.cinelerra-gg.org/mailman/listinfo/cin