Re: [NEW] sysutils/pkg_depts

2020-09-23 Thread Marc Espie
On Tue, Sep 22, 2020 at 11:15:08AM +0200, m...@fn.de wrote: > > On 2020-09-22 10:26, Stuart Henderson wrote: > > > > This has put me off the idea further tbh. Once committed it can't simply > > be ignored, if it's not maintained in a timely fashion if things change > > in OpenBSD (which is not

Re: [NEW] sysutils/pkg_depts

2020-09-21 Thread Marc Espie
On Mon, Sep 21, 2020 at 02:36:36PM +0100, Stuart Henderson wrote: > It seems silly to have two tools that do basically the same thing, does this > have something that show-reverse-deps doesn't? And if so, wouldn't it be > better to improve the existing tool? > > Send tars not cvs diffs for new

Re: [NEW] sysutils/pkg_depts

2020-09-21 Thread Marc Espie
On Mon, Sep 21, 2020 at 02:52:34PM +0200, m...@fn.de wrote: > Anybody willing to OK? > No, I'm actually against it. Most specifically, you're not supposed to look within packing-lists directly. If I were to change the way @depend are recorded, your script will miserably fail. You've got lots

CVS: cvs.openbsd.org: ports

2020-09-08 Thread Marc Espie
CVSROOT:/cvs Module name:ports Changes by: es...@cvs.openbsd.org 2020/09/08 21:19:10 Modified files: multimedia/shotcut: Makefile distinfo Log message: minor update, okay tracey@

update shotcut to 2020.07.11

2020-09-08 Thread Marc Espie
tested on a small video and a few effects, things work just fine. Okay ? Index: Makefile === RCS file: /cvs/ports/multimedia/shotcut/Makefile,v retrieving revision 1.4 diff -u -p -r1.4 Makefile --- Makefile30 Jul 2020 14:40:07

Re: rsync 3.2.3

2020-08-26 Thread Marc Espie
On Wed, Aug 26, 2020 at 12:48:27PM +0100, Stuart Henderson wrote: > This updates rsync to 3.2.3, which amongst other things includes > https://github.com/WayneD/rsync/commit/9375a8c4c2b79a9d44ceae00de7284b77d79033c > which greatly reduces maxrss (https://github.com/WayneD/rsync/issues/81). > >

Re: UPDATE: TeX Live 2020

2020-08-11 Thread Marc Espie
On Tue, Aug 11, 2020 at 08:41:21PM +0100, Edd Barrett wrote: > Hi Elias, > > Thanks for the testing! I owe you one. > > On Sat, Aug 08, 2020 at 11:37:20PM -0300, Elias M. Mariani wrote: > > make port-lib-depends-check > > texlive_base-2020(print/texlive/base,-main): > > Missing: pcre.3 from

Re: LLVM 10: mail/bogofilter

2020-08-01 Thread Marc Espie
On Sat, Aug 01, 2020 at 05:13:11PM +0200, Christian Weisgerber wrote: > mail/bogofilter's build failure with LLVM 10 is another silly case > of a configure script failing to recognize the compiler and treating > is as gcc3 or older. >

CVS: cvs.openbsd.org: ports

2020-07-30 Thread Marc Espie
CVSROOT:/cvs Module name:ports Changes by: es...@cvs.openbsd.org 2020/07/30 08:40:07 Modified files: multimedia/shotcut: Makefile distinfo multimedia/shotcut/patches: patch-src_mainwindow_cpp patch-src_src_pro

Re: rsync: update to 3.2.2

2020-07-23 Thread Marc Espie
On Thu, Jul 23, 2020 at 04:31:00PM +0200, Klemens Nanni wrote: > Lots of fixes and enhancements: > https://download.samba.org/pub/rsync/NEWS#3.2.2 > > zlib patches have been merged, rest of the patches/ diff is just churn. > > For now, I've conservatively disabled lz4, zstd and xxhash support to

Re: bsd.port.mk: produce build.log

2020-07-18 Thread Marc Espie
On Sat, Jul 18, 2020 at 08:29:03PM +0200, Klemens Nanni wrote: > "make test" always produces TEST_LOGFILE (default is ${WRKDIR}/test.log) > which is really handy for analysing failures and sharing them here on > ports@. > > Diff below implements BUILD_LOGFILE so "make build" behaves the same; >

Re: to Pandoc or not

2020-07-15 Thread Marc Espie
On Wed, Jul 15, 2020 at 05:10:02PM -0700, Evan Silberman wrote: > Stuart Henderson wrote: > > On 2020/07/14 13:23, sven falempin wrote: > > > Dear readers, > > > > > > Is anyone working on that for openbsd ? > > > > > >

Re: to Pandoc or not

2020-07-15 Thread Marc Espie
On Wed, Jul 15 '20 at 05.50 NZST, Ingo Schwarze wrote: > I never found pandoc useful for anything - so much so that i never > even felt a need to evalute its quality for real. I dimly remember > that on rare occasions, i saw man(7) output generated by pandoc, > and if i remeber correctly, that

CVS: cvs.openbsd.org: ports

2020-07-11 Thread Marc Espie
CVSROOT:/cvs Module name:ports Changes by: es...@cvs.openbsd.org 2020/07/11 06:03:58 Modified files: sysutils/random_run: Makefile distinfo Log message: new minor version: handles "-" as standard input for option -l

Re: devel/bison: fix parallel build race

2020-07-11 Thread Marc Espie
On Sat, Jul 04, 2020 at 11:47:14PM +0200, Christian Weisgerber wrote: > Parallel building devel/bison fails. There is a race where the > build tries to run the newly compiled bison for its --help output > but doesn't wait for the binary to actually be available. > > This is one of these cases

CVS: cvs.openbsd.org: ports

2020-07-07 Thread Marc Espie
CVSROOT:/cvs Module name:ports Changes by: es...@cvs.openbsd.org 2020/07/07 04:53:27 Modified files: databases/sqlports: Makefile databases/sqlports/files/man: sqlports.5 Log message: flesh out some more stuff in the manpage. do a REV bump for once, since we

CVS: cvs.openbsd.org: ports

2020-07-06 Thread Marc Espie
CVSROOT:/cvs Module name:ports Changes by: es...@cvs.openbsd.org 2020/07/06 03:06:45 Modified files: databases/sqlports: Makefile databases/sqlports/files/man: sqlports.5 databases/sqlports/pkg: DESCR-main PLIST-main Log message: move description of

Re: UPDATE: x11/xglobe

2020-07-06 Thread Marc Espie
On Sun, Jul 05, 2020 at 05:40:31PM +0200, Rafael Sadowski wrote: > Please find below a diff to update xglobe to 0.6. The original xglobe is > dead upstream. I modernized it just for fun and with the motivation to > get rid of Qt3/KDE3 from the ports tree. This is a complete rewrite in > C++17/Qt5.

CVS: cvs.openbsd.org: ports

2020-07-05 Thread Marc Espie
CVSROOT:/cvs Module name:ports Changes by: es...@cvs.openbsd.org 2020/07/05 06:28:13 Modified files: databases/sqlports/files/man: sqlports.5 Log message: flesh out a bit more stuff from DESCR

CVS: cvs.openbsd.org: ports

2020-07-05 Thread Marc Espie
CVSROOT:/cvs Module name:ports Changes by: es...@cvs.openbsd.org 2020/07/05 06:23:24 Added files: databases/sqlports/files/man: sqlports.5 Log message: seems that people don't expect this information to be in DESC, so start converting it to mandoc, to be completed.

CVS: cvs.openbsd.org: ports

2020-07-05 Thread Marc Espie
CVSROOT:/cvs Module name:ports Changes by: es...@cvs.openbsd.org 2020/07/05 04:05:40 Modified files: security/openssl/1.1: Makefile security/openssl/1.1/pkg: PLIST Log message: fix broken conflict spec. nice thinko

CVS: cvs.openbsd.org: ports

2020-07-05 Thread Marc Espie
CVSROOT:/cvs Module name:ports Changes by: es...@cvs.openbsd.org 2020/07/05 03:45:42 Modified files: x11/i3-mousedrag: Makefile x11/i3-mousedrag/pkg: PLIST x11/i3-gaps: Makefile x11/i3-gaps/pkg: PLIST Log message: missed conflict

CVS: cvs.openbsd.org: ports

2020-07-05 Thread Marc Espie
CVSROOT:/cvs Module name:ports Changes by: es...@cvs.openbsd.org 2020/07/05 03:44:08 Modified files: lang/guile2: Makefile lang/guile2/pkg: PLIST lang/scheme-report: Makefile lang/scheme-report/pkg: PLIST Log message: missed conflict

CVS: cvs.openbsd.org: ports

2020-07-04 Thread Marc Espie
CVSROOT:/cvs Module name:ports Changes by: es...@cvs.openbsd.org 2020/07/04 12:29:21 Modified files: infrastructure/bin: pkg_check-problems Log message: do NOT check license issues for !current packages, because we don't care that much, and more importantly, because

CVS: cvs.openbsd.org: ports

2020-07-04 Thread Marc Espie
CVSROOT:/cvs Module name:ports Changes by: es...@cvs.openbsd.org 2020/07/04 10:54:22 Modified files: infrastructure/bin: pkg_check-problems Log message: seems that other people do not use this... get rid of cdrom_okay

CVS: cvs.openbsd.org: ports

2020-07-04 Thread Marc Espie
CVSROOT:/cvs Module name:ports Changes by: es...@cvs.openbsd.org 2020/07/04 10:53:42 Modified files: infrastructure/lib/OpenBSD: PlistScanner.pm Log message: thinko

CVS: cvs.openbsd.org: ports

2020-06-28 Thread Marc Espie
CVSROOT:/cvs Module name:ports Changes by: es...@cvs.openbsd.org 2020/06/28 04:58:25 Modified files: infrastructure/bin: register-plist Log message: register-plist is a logical place to be able to save stuff elsewhere, set up to be able to save manpages

CVS: cvs.openbsd.org: ports

2020-06-27 Thread Marc Espie
CVSROOT:/cvs Module name:ports Changes by: es...@cvs.openbsd.org 2020/06/27 06:55:22 Modified files: graphics/unwebp: Makefile distinfo Log message: minor update: tweak usage, have -d also delete webp for already converted files as this makes more sense.

CVS: cvs.openbsd.org: ports

2020-06-26 Thread Marc Espie
CVSROOT:/cvs Module name:ports Changes by: es...@cvs.openbsd.org 2020/06/26 08:47:05 Modified files: infrastructure/bin: update-plist Log message: explain why FAKE_TREE_OWNER is not a regular option

CVS: cvs.openbsd.org: ports

2020-06-26 Thread Marc Espie
CVSROOT:/cvs Module name:ports Changes by: es...@cvs.openbsd.org 2020/06/26 07:00:07 Modified files: infrastructure/bin: update-plist Log message: add code to write conflict-related info in the cookie, so that if it changes we rerun pkglocate also check the pkglocate

CVS: cvs.openbsd.org: ports

2020-06-26 Thread Marc Espie
CVSROOT:/cvs Module name:ports Changes by: es...@cvs.openbsd.org 2020/06/26 06:00:27 Modified files: infrastructure/bin: update-plist Log message: manage a pkglocate_cookie, so that pkglocate is not run every time. Note the TODOs, there's some code to be written to

CVS: cvs.openbsd.org: ports

2020-06-26 Thread Marc Espie
CVSROOT:/cvs Module name:ports Changes by: es...@cvs.openbsd.org 2020/06/26 05:51:16 Modified files: infrastructure/mk: bsd.port.mk Log message: add a PKGLOCATE_COOKIE that will be filled in by update-plist turns out update-plist already groks -D thanks to inheriting

CVS: cvs.openbsd.org: ports

2020-06-26 Thread Marc Espie
CVSROOT:/cvs Module name:ports Changes by: es...@cvs.openbsd.org 2020/06/26 05:02:32 Modified files: infrastructure/lib/OpenBSD/PortGen/Port: Go.pm Log message: nit, help debug

Re: [portgen go] Fix issues with missing dependencies

2020-06-25 Thread Marc Espie
On Wed, Jun 24, 2020 at 09:38:38PM -0600, Aaron Bieber wrote: > + my @output = <$fh>; > + chomp @output; > + close $fh or die "Unable to close pipe '$cmd': $!"; You still missed that one: you won't get anything out of $!, it's $? that's gonna hold the child exit status.

CVS: cvs.openbsd.org: ports

2020-06-22 Thread Marc Espie
CVSROOT:/cvs Module name:ports Changes by: es...@cvs.openbsd.org 2020/06/22 06:51:12 Modified files: infrastructure/bin: portgen Log message: remove the switch, use a map name -> class this makes it easier to add languages, and also provides more decent error reports

Re: bsd.port.mk: print failed matches in "make patch"

2020-06-16 Thread Marc Espie
On Mon, Jun 15, 2020 at 01:48:46PM +0200, Klemens Nanni wrote: > On Mon, Jun 15, 2020 at 01:39:11PM +0200, Marc Espie wrote: > > The idea looks fine, I"m not quite sure it works nor is perfect. > Why would it not work? > > > [ ] && echo will only succee

Re: update print/poppler

2020-06-16 Thread Marc Espie
On Mon, Jun 15, 2020 at 02:51:06PM +0200, Matthias Kilian wrote: > Hi, > > On Sun, Jun 14, 2020 at 06:48:34PM -0700, Greg Steuck wrote: > > > here's an update to poppler-0.89.0. Survived a bulk build on amd64, > > > but I coould'nt yet test it (because of flaky internet, and I'm > > > doing the

CVS: cvs.openbsd.org: ports

2020-06-15 Thread Marc Espie
CVSROOT:/cvs Module name:ports Changes by: es...@cvs.openbsd.org 2020/06/15 13:12:48 Modified files: graphics/unwebp: Makefile distinfo Log message: a few new useful options, rtfm

Re: bsd.port.mk: print failed matches in "make patch"

2020-06-15 Thread Marc Espie
On Sat, Jun 13, 2020 at 06:46:27PM +0200, Klemens Nanni wrote: > For small ports not so much, but when updating bigger ones with lots of > patches and/or churn, I always find cumbersome to go scroll back in my > terminal to look for failed hunks or cd into WRKSRC and look for .rej > files, so

CVS: cvs.openbsd.org: ports

2020-06-13 Thread Marc Espie
CVSROOT:/cvs Module name:ports Changes by: es...@cvs.openbsd.org 2020/06/13 03:14:14 Modified files: infrastructure/bin: register-plist Log message: refactor, no functional change. Start grouping and naming options

CVS: cvs.openbsd.org: ports

2020-06-11 Thread Marc Espie
CVSROOT:/cvs Module name:ports Changes by: es...@cvs.openbsd.org 2020/06/11 13:55:15 Modified files: databases/sqlports: Makefile databases/sqlports/files/man: show-reverse-deps.1 databases/sqlports/files/scripts: show-reverse-deps Log message: extend

CVS: cvs.openbsd.org: ports

2020-06-11 Thread Marc Espie
CVSROOT:/cvs Module name:ports Changes by: es...@cvs.openbsd.org 2020/06/11 10:05:52 Modified files: databases/sqlports: Makefile databases/sqlports/files/man: show-reverse-deps.1 databases/sqlports/files/scripts: show-reverse-deps Log message:

CVS: cvs.openbsd.org: ports

2020-06-11 Thread Marc Espie
CVSROOT:/cvs Module name:ports Changes by: es...@cvs.openbsd.org 2020/06/11 09:40:41 Modified files: databases/sqlports: Makefile databases/sqlports/files/scripts: show-reverse-deps databases/sqlports/pkg: PLIST-main Added files:

CVS: cvs.openbsd.org: ports

2020-06-11 Thread Marc Espie
CVSROOT:/cvs Module name:ports Changes by: es...@cvs.openbsd.org 2020/06/11 09:39:47 ports/databases/sqlports/files/man Update of /cvs/ports/databases/sqlports/files/man In directory cvs.openbsd.org:/tmp/cvs-serv85768/man Log Message: Directory

register-plist news

2020-06-10 Thread Marc Espie
I've started hacking on register-plist a bit, I will probably change internals yet some more. As far as user-visible changes go, there is a REGISTER_PLIST_OPTS in bsd.port.mk to pass options to it, and two new (potentially useful) options: -n so that register-plist will still check plist for

CVS: cvs.openbsd.org: ports

2020-06-10 Thread Marc Espie
CVSROOT:/cvs Module name:ports Changes by: es...@cvs.openbsd.org 2020/06/10 01:42:39 Modified files: infrastructure/bin: register-plist Log message: add a -s snap option that will always write plist into a separate directory if no error

CVS: cvs.openbsd.org: ports

2020-06-10 Thread Marc Espie
CVSROOT:/cvs Module name:ports Changes by: es...@cvs.openbsd.org 2020/06/10 00:54:29 Modified files: infrastructure/bin: register-plist Log message: I forgot about the mogrified part, respect -n too tweak the message to show that nothing got written

CVS: cvs.openbsd.org: ports

2020-06-09 Thread Marc Espie
CVSROOT:/cvs Module name:ports Changes by: es...@cvs.openbsd.org 2020/06/09 15:00:33 Modified files: infrastructure/bin: register-plist Log message: add option -n, as it's the simplest one to do (do all the checks, but don't write a thing. useful for bulks with test

Re: CVS: cvs.openbsd.org: ports

2020-06-09 Thread Marc Espie
On Tue, Jun 09, 2020 at 08:21:12PM +0200, Marc Espie wrote: > On Tue, Jun 09, 2020 at 07:52:53PM +0200, Rafael Sadowski wrote: > > On Tue Jun 09, 2020 at 11:35:12AM -0600, Antoine Jacoutot wrote: > > > CVSROOT: /cvs > > > Module name: ports > > > Changes

Re: CVS: cvs.openbsd.org: ports

2020-06-09 Thread Marc Espie
On Tue, Jun 09, 2020 at 07:52:53PM +0200, Rafael Sadowski wrote: > On Tue Jun 09, 2020 at 11:35:12AM -0600, Antoine Jacoutot wrote: > > CVSROOT:/cvs > > Module name:ports > > Changes by: ajacou...@cvs.openbsd.org 2020/06/09 11:35:12 > > > > Modified files: > >

CVS: cvs.openbsd.org: ports

2020-06-09 Thread Marc Espie
CVSROOT:/cvs Module name:ports Changes by: es...@cvs.openbsd.org 2020/06/09 05:01:08 Modified files: infrastructure/mk: bsd.port.mk Log message: add REGISTER_PLIST_OPTS to be used to allow various things

Re: pkg_check on 6.7: texmf packages have too many dependencies

2020-06-04 Thread Marc Espie
On Thu, Jun 04, 2020 at 02:29:06PM +0100, Edd Barrett wrote: > Hi Folks, > > I recently received an email from a user (in CC) who reported that if > you install texlive_texmf-full on 6.7 then pkg_check will complain. > > I installed 6.7 in a VM and tried. Sure enough: > > ``` > openbsd67#

CVS: cvs.openbsd.org: ports

2020-06-03 Thread Marc Espie
CVSROOT:/cvs Module name:ports Changes by: es...@cvs.openbsd.org 2020/06/03 06:06:05 Modified files: sysutils/random_run: Makefile distinfo Log message: new release, with a new option

CVS: cvs.openbsd.org: ports

2020-06-02 Thread Marc Espie
CVSROOT:/cvs Module name:ports Changes by: es...@cvs.openbsd.org 2020/06/02 09:29:59 Modified files: graphics/libpano13: Makefile Log message: are you kidding me ? remove useless lines from Makefile

Re: NOW IN CURRENT: shotcut

2020-06-01 Thread Marc Espie
The big issue is reproducibility. I can get shotcut to crash *eventually* with something that does not look anywhere like what you guys are getting, but by default, it starts perfectly fine and I have no trouble reading videos nor playing with the timeline. Haven't figured out what's different

Re: NOW IN CURRENT: shotcut

2020-06-01 Thread Marc Espie
On Mon, Jun 01, 2020 at 03:20:34PM +0200, prx wrote: > Hi, > after trying the new shotcut port, I wonder if there is any tip to read > video files with it. Do you convert to any format befort importing into > shotcut? > > Find attached messages printed by shotcut recorded with script(1). This

CVS: cvs.openbsd.org: ports

2020-06-01 Thread Marc Espie
CVSROOT:/cvs Module name:ports Changes by: es...@cvs.openbsd.org 2020/06/01 02:41:36 Modified files: infrastructure/mk: bsd.port.mk Log message: fix some PORTS_PRIVSEP tidbits: UPDATE_COOKIES and BULK_COOKIES are not generated by dpb, but if you build stuff manually

Re: games/ioquake3 cvs current diffs adding ppc support

2020-06-01 Thread Marc Espie
On Mon, Jun 01, 2020 at 09:38:32AM +0200, Landry Breuil wrote: > On Mon, Jun 01, 2020 at 09:27:25AM +0200, Charlene Wendling wrote: > > On Mon, 1 Jun 2020 05:24:19 +0200 > > Alex Free wrote: > > > > +.if ${MACHINE_ARCH} == "powerpc" > > +QUAKE_ARCH=ppc > > +.else > > QUAKE_ARCH=

CVS: cvs.openbsd.org: ports

2020-05-31 Thread Marc Espie
CVSROOT:/cvs Module name:ports Changes by: es...@cvs.openbsd.org 2020/05/31 13:24:26 Modified files: infrastructure/lib/DPB: PkgPath.pm Log message: work a bit harder so that pkgpaths do not gain a bogus info by autovivification this allows to use is_stub once again

Re: Inconsolata 3.0 is no longer fixed width.

2020-05-30 Thread Marc Espie
On Sat, May 30, 2020 at 01:00:17PM -0600, Anthony J. Bentley wrote: > Hi, > > Matthieu Herrb writes: > > This github issue does pretty much explain what's going on: > > https://github.com/googlefonts/Inconsolata/issues/42 : > > I see nothing as extreme as the screenshots in that link in my

Re: Inconsolata 3.0 is no longer fixed width.

2020-05-30 Thread Marc Espie
On Sat, May 30, 2020 at 12:41:16PM +0200, Matthieu Herrb wrote: > On Mon, May 25, 2020 at 01:08:52AM -0600, Anthony J. Bentley wrote: > > CVSROOT:/cvs > > Module name:ports > > Changes by: bent...@cvs.openbsd.org 2020/05/25 01:08:52 > > > > Modified files: > >

Re: Creating new port issue - including WANTLIB in package

2020-05-29 Thread Marc Espie
On Fri, May 29, 2020 at 01:12:42PM -0600, Steve Williams wrote: > It applies patches and compiles clean.  However, when I try to create a > package, it seems to want to include the libraries that I have in WANTLIB. WANTLIB doesn't work by itself. It also requires LIB_DEPENDS/RUN_DEPENDS

Re: NOW IN CURRENT: shotcut

2020-05-29 Thread Marc Espie
One important note for non-technical readers: *porting* shotcut to OpenBSD meant fixing *three bugs* in the ported software: - mlt (the framework shotcut is based on) uses mutexes extensively, but does NOT initialize them completely *even though the mutex default type is unspecified by posix -

NOW IN CURRENT: shotcut

2020-05-29 Thread Marc Espie
A few days ago, I imported shotcut, then fixed quite a few bugs so that it's now usable. Now that snapshots have rolled over, it is available as a package. shotcut is modern video editing software. With the current medical situation, I had needs to prepare online video courses, and having to go

CVS: cvs.openbsd.org: ports

2020-05-28 Thread Marc Espie
CVSROOT:/cvs Module name:ports Changes by: es...@cvs.openbsd.org 2020/05/28 14:53:40 Modified files: infrastructure/lib/DPB: PkgPath.pm Log message: revert temporarily there's a bug elsewhere (autovivification of $w->{info}{BUILD_PACKAGES}) that I need to track down

CVS: cvs.openbsd.org: ports

2020-05-26 Thread Marc Espie
CVSROOT:/cvs Module name:ports Changes by: es...@cvs.openbsd.org 2020/05/26 13:11:22 Modified files: multimedia/shotcut: Makefile Added files: multimedia/shotcut/patches: patch-src_mltcontroller_cpp

Re: CVS: cvs.openbsd.org: ports

2020-05-26 Thread Marc Espie
On Tue, May 26, 2020 at 11:05:42AM -0600, Marc Espie wrote: > CVSROOT: /cvs > Module name: ports > Changes by: es...@cvs.openbsd.org 2020/05/26 11:05:42 > > Added files: > multimedia/shotcut/patches: patch-src_qmltypes_qmlfilter_cpp > > Log message: >

CVS: cvs.openbsd.org: ports

2020-05-26 Thread Marc Espie
CVSROOT:/cvs Module name:ports Changes by: es...@cvs.openbsd.org 2020/05/26 11:06:34 Modified files: multimedia : Makefile Log message: +shotcut

CVS: cvs.openbsd.org: ports

2020-05-26 Thread Marc Espie
CVSROOT:/cvs Module name:ports Changes by: es...@cvs.openbsd.org 2020/05/26 11:05:42 Added files: multimedia/shotcut/patches: patch-src_qmltypes_qmlfilter_cpp Log message: as usual, this was a REALLY stupid bug turns out name.toUf8().constData() is a temporary, so

Re: shotcut core

2020-05-25 Thread Marc Espie
On Mon, May 25, 2020 at 04:10:49PM +0200, Marc Espie wrote: > Here's another one... > I went back to 19.12.31 > > and installed the debug-qt* packages (somehow they weren't around last > time). > > > It's still the exact same bug. > > > By adding printing,

Re: shotcut core

2020-05-25 Thread Marc Espie
Here's another one... I went back to 19.12.31 and installed the debug-qt* packages (somehow they weren't around last time). It's still the exact same bug. By adding printing, that "name" parameter is complete garbage each and every time. Sometime, we get "lucky" and it doesn't segfault. It

Re: about BUILD_PACKAGES and no-strip

2020-05-24 Thread Marc Espie
On Sun, May 24, 2020 at 11:40:34PM +0200, Rafael Sadowski wrote: > On Sun May 24, 2020 at 03:09:42PM +0200, Marc Espie wrote: > > There are some special needs ports like cmake and qmake with labyrinthine > > options to figure out what to build "release" and you get

shotcut core

2020-05-24 Thread Marc Espie
Here's the most frequent backtrace now. It's very easy to obtain: start shotcut, load a video, try to apply a filter that requires parameters (say size and position), watch shotcut segfault. It seems to come from qml code. Somewhat a pain to try to figure it out further, since we don't have

about BUILD_PACKAGES and no-strip

2020-05-24 Thread Marc Espie
There are some special needs ports like cmake and qmake with labyrinthine options to figure out what to build "release" and you get no debug info. "debug" and you get wy too much stuff. the tweak I committed a few days ago does cut through the red tape... Tool wants special treatment to

extend PORTS_PRIVSEP

2020-05-22 Thread Marc Espie
While working manually on dpb goo, I found out BULK_COOKIES_DIR and UPDATE_COOKIES_DIR are not handled (this results in error messages at the end of logs, actually, not actual errors, since this is behind a -rm -f Also, the empty(UPDATE_COOKIES_DIR) goo is wrong Could you test/check, please ?

CVS: cvs.openbsd.org: ports

2020-05-21 Thread Marc Espie
CVSROOT:/cvs Module name:ports Changes by: es...@cvs.openbsd.org 2020/05/21 06:38:46 Modified files: devel/dwz : Makefile Log message: forgotten piece of the "no-strip" commit, noticed by aja@ this is actually less tricky: just use the built dwz

Re: kdenlive update

2020-05-20 Thread Marc Espie
On Tue, May 19, 2020 at 03:34:41PM +0200, Erling Westenvik wrote: > On Sun, May 17, 2020 at 04:48:27PM +0200, Marc Espie wrote: > > Seems I can now remove the BROKEN marker. > > This does depend on rttr and the mlt fixed update > > Awesome. A working video editor! Comp

CVS: cvs.openbsd.org: ports

2020-05-20 Thread Marc Espie
CVSROOT:/cvs Module name:ports Changes by: es...@cvs.openbsd.org 2020/05/20 08:54:23 Log message: shotcut is a full-featured video editor loosely based on WIP by bentley@ the current port is not finished, but it can do a few things: play videos, cut

CVS: cvs.openbsd.org: ports

2020-05-20 Thread Marc Espie
CVSROOT:/cvs Module name:ports Changes by: es...@cvs.openbsd.org 2020/05/20 06:57:10 Modified files: infrastructure/mk: bsd.port.mk Added files: infrastructure/bin: no-strip Log message: neuter strip if DEBUG_PACKAGES is set tweak the path for copy-debug-info

CVS: cvs.openbsd.org: ports

2020-05-20 Thread Marc Espie
CVSROOT:/cvs Module name:ports Changes by: es...@cvs.openbsd.org 2020/05/20 04:03:45 Modified files: x11/kde-applications/kdenlive: Makefile x11/kde-applications/kdenlive/pkg: PLIST Log message: kdenlive should work again

CVS: cvs.openbsd.org: ports

2020-05-20 Thread Marc Espie
CVSROOT:/cvs Module name:ports Changes by: es...@cvs.openbsd.org 2020/05/20 03:17:04 Modified files: multimedia/mlt : Makefile multimedia/mlt/patches: patch-src_framework_mlt_property_c Log message: neuter ALL setlocale LC_NUMERIC. That code is not

Re: devel/autoconf/2.69, add runstatedir

2020-05-19 Thread Marc Espie
On Tue, May 19, 2020 at 12:26:45PM +0100, Stuart Henderson wrote: > This adds the runstatedir patch from autoconf git. It was added upstream > in 2013 but there hasn't been a newer release; some Linux distros have been > patching it in their 2.69 packages for years. > > I don't _really_ like

CVS: cvs.openbsd.org: ports

2020-05-19 Thread Marc Espie
CVSROOT:/cvs Module name:ports Changes by: es...@cvs.openbsd.org 2020/05/19 02:50:55 Modified files: databases/sqlports: Makefile databases/sqlports/files: Info.pm Log message: handling of DEBUG_PACKAGES/DEBUG_CONFIGURE_ARGS is bogus just ignore them for now

CVS: cvs.openbsd.org: ports

2020-05-19 Thread Marc Espie
CVSROOT:/cvs Module name:ports Changes by: es...@cvs.openbsd.org 2020/05/19 02:44:44 Modified files: databases/sqlports/files: TreeWalker.pm Log message: likewise, don't try to use incomplete hash

CVS: cvs.openbsd.org: ports

2020-05-19 Thread Marc Espie
CVSROOT:/cvs Module name:ports Changes by: es...@cvs.openbsd.org 2020/05/19 02:38:03 Modified files: databases/sqlports/files: TreeWalker.pm Log message: parse errors in dump-vars will lead to $h being invalid, push past that so that it properly errors out without a

CVS: cvs.openbsd.org: ports

2020-05-18 Thread Marc Espie
CVSROOT:/cvs Module name:ports Changes by: es...@cvs.openbsd.org 2020/05/18 13:33:43 Modified files: multimedia/mlt/patches: patch-src_framework_mlt_property_c Log message: add comment explaining what's going on and why the current fix is fine

Re: bsd.port.mk: Add EXTRACT_FILES

2020-05-18 Thread Marc Espie
On Mon, May 18, 2020 at 01:46:38PM +0200, Klemens Nanni wrote: > On Mon, May 18, 2020 at 12:17:01PM +0100, Stuart Henderson wrote: > > I like it, can you work up something for the bsd.port.mk(5) manual please? > Diff below tries to lay out the inner workings without being too > specific; a bit

CVS: cvs.openbsd.org: ports

2020-05-18 Thread Marc Espie
CVSROOT:/cvs Module name:ports Changes by: es...@cvs.openbsd.org 2020/05/18 06:29:17 Modified files: devel : Makefile Log message: +rttr

CVS: cvs.openbsd.org: ports

2020-05-18 Thread Marc Espie
CVSROOT:/cvs Module name:ports Changes by: es...@cvs.openbsd.org 2020/05/18 06:28:33 Log message: Runtime Reflection Library for C++ mostly a dependency for kdenlive tweaks and okay rsadowski@ Status: Vendor Tag: espie Release Tags:

Re: NEW PORT: rttr (dependency for kdenlive)

2020-05-18 Thread Marc Espie
On Sun, May 17, 2020 at 04:43:33PM +0200, Marc Espie wrote: > kdenlive lists this as "optional" but in actuality > it cheats and tries to download it on the side (which > obviously fails for us) > > The tests take forever to build, and some actually don't build with

FETCH_PACKAGES fix committed

2020-05-18 Thread Marc Espie
I found the reason for a FETCH_PACKAGES bug, linked to pkg_add optimizations: with faster package updates, unchanged files would mean pkg_add would not always read to the end of the archive. Unfortunately, this also applies to symlinks at the end of the plist, as the meta-info is enough, AND

CVS: cvs.openbsd.org: ports

2020-05-18 Thread Marc Espie
CVSROOT:/cvs Module name:ports Changes by: es...@cvs.openbsd.org 2020/05/18 04:54:15 Modified files: multimedia/mlt : Makefile Log message: missed --disable-rubberband in jca@'s mail

CVS: cvs.openbsd.org: ports

2020-05-18 Thread Marc Espie
CVSROOT:/cvs Module name:ports Changes by: es...@cvs.openbsd.org 2020/05/18 04:33:17 Modified files: multimedia/mlt : Makefile distinfo multimedia/mlt/patches: patch-src_framework_Makefile multimedia/mlt/pkg: PLIST-gpl2 PLIST-main Added files:

Re: UPDATE: multimedia/mlt

2020-05-18 Thread Marc Espie
On Sun, May 17, 2020 at 01:12:58PM -0600, Tracey Emery wrote: > On Sun, May 17, 2020 at 08:19:05PM +0200, Marc Espie wrote: > > On Sun, May 17, 2020 at 04:47:10PM +0200, Marc Espie wrote: > > > Now with better runtime behavior > > > > > > Some of the mutexes do

Re: UPDATE: multimedia/mlt

2020-05-17 Thread Marc Espie
On Sun, May 17, 2020 at 04:47:10PM +0200, Marc Espie wrote: > Now with better runtime behavior > > Some of the mutexes don't work with our stringent defaults, > this prevents some aborts on shotcut. > > and I can actually start a clip in kdenlive with this. New version.

WIP: multimedia/shotcut

2020-05-17 Thread Marc Espie
With the new mlt (mutex fixes), this at least starts and can import clips to the timeline, and can try to edit stuff There are more bugs to be found, the beastie tends to segfault on a whim, but assuming new mlt gets in, we could probably work in-tree. Note that DEBUG_PACKAGES working properly

kdenlive update

2020-05-17 Thread Marc Espie
Seems I can now remove the BROKEN marker. This does depend on rttr and the mlt fixed update Index: Makefile === RCS file: /cvs/ports/x11/kde-applications/kdenlive/Makefile,v retrieving revision 1.9 diff -u -p -r1.9 Makefile ---

Re: UPDATE: multimedia/mlt

2020-05-17 Thread Marc Espie
Now with better runtime behavior Some of the mutexes don't work with our stringent defaults, this prevents some aborts on shotcut. and I can actually start a clip in kdenlive with this. Index: Makefile === RCS file:

NEW PORT: rttr (dependency for kdenlive)

2020-05-17 Thread Marc Espie
kdenlive lists this as "optional" but in actuality it cheats and tries to download it on the side (which obviously fails for us) The tests take forever to build, and some actually don't build with clang/libc++, so I've removed them, as the goal is only to build kdenlive rttr.tgz Description:

Re: UPDATE: multimedia/mlt

2020-05-16 Thread Marc Espie
On Sat, May 16, 2020 at 05:38:20PM +0200, Marc Espie wrote: > On Sat, May 16, 2020 at 05:10:23PM +0200, Marc Espie wrote: > > On Sat, May 16, 2020 at 08:20:55AM -0600, Tracey Emery wrote: > > > On Sat, May 16, 2020 at 02:38:22PM +0200, Marc Espie wrote: > > > > Wi

  1   2   3   4   5   6   7   8   9   10   >