Re: [patch] unbreak textproc/filepp
Sat, Dec 23, 2017 at 10:05:27 +0100, Landry Breuil may have written: > On Fri, Dec 22, 2017 at 04:29:47PM -0800, Matthew Clarke wrote: > > Hi. > > > > The textproc/filepp port has no maintainer, and the current filepp package > > installs a filepp script that won't run due to the following error: > > > > Global symbol "$DESTDIR" requires explicit package name (did you forget > > to declare "my $DESTDIR"?) at /usr/local/bin/filepp line 43. > > Execution of /usr/local/bin/filepp aborted due to compilation errors. [ snip ] > > [ ... ] Would a ports > > committer please review the following diff, and commit it if it's OK? > > > Hmm technically the filepp script isn't right since its buildsystem > replaces @moduledir@ by what ends up in moduledir, defaulting to > $prefix/share/filepp/modules, and $prefix is assigned ${DESTDIR} by the > openbsd build system. > - > The moduledir is also set in the filepp script. If you do not run > configure for whatever reason you should also change @moduledir@ on > line 40 to the directory you intend to install the modules in (or > delete the whole line if you are not using modules). > - > There should be some TRUEPREFIX here.. > > so i think the following is "more correct" in case TRUEPREFIX changes > someday: Agreed. Thanks for the review and comments. BTW, the diff in your email had spaces, not tabs, so patch(1) rejected it, but it worked fine when I applied it by hand. -- "With your own code to haunt you, who needs users?" -- Maarten Wiltink
[patch] unbreak textproc/filepp
Hi. The textproc/filepp port has no maintainer, and the current filepp package installs a filepp script that won't run due to the following error: Global symbol "$DESTDIR" requires explicit package name (did you forget to declare "my $DESTDIR"?) at /usr/local/bin/filepp line 43. Execution of /usr/local/bin/filepp aborted due to compilation errors. Line 43 of /usr/local/bin/filepp reads: push(@INC, "${DESTDIR}/usr/local/share/filepp/modules"); I've used filepp successfully on OpenBSD before, but not for a couple of years or so. Reviewing the port history, I see that revision 1.6 (October 2015) of textproc/filepp/Makefile removed the following code: post-configure: @perl -pi -e 's,\$${DESTDIR},,g;' ${WRKSRC}/filepp Reverting that change makes filepp work again, but introduces the following warning into the "configure" phase of the build: Unescaped left brace in regex is deprecated, passed through in regex; marked by <-- HERE in m/\${ <-- HERE DESTDIR}/ at -e line 1. Switching from perl to sed is one way to avoid the warning. Would a ports committer please review the following diff, and commit it if it's OK? Thanks. Index: Makefile === RCS file: /cvs/ports/textproc/filepp/Makefile,v retrieving revision 1.8 diff -u -p -r1.8 Makefile --- Makefile15 Nov 2017 21:28:05 - 1.8 +++ Makefile22 Dec 2017 23:44:37 - @@ -3,7 +3,7 @@ COMMENT = cpp-like generic file pre-processor DISTNAME = filepp-1.8.0 -REVISION = 2 +REVISION = 3 CATEGORIES = textproc devel MASTER_SITES = http://www-users.york.ac.uk/~dm26/filepp/ @@ -16,5 +16,8 @@ CONFIGURE_STYLE = gnu dest NO_BUILD = Yes PKG_ARCH = * TEST_TARGET = fulltest + +post-configure: + sed -i 's,$${DESTDIR},,g' ${WRKSRC}/filepp .include -- "On two occasions I have been asked [by members of Parliament!], `Pray, Mr. Babbage, if you put into the machine wrong figures, will the right answers come out?' I am not able rightly to apprehend the kind of confusion of ideas that could provoke such a question." -- Charles Babbage
Re: [UPDATE] shells/tcsh to 6.20.00
Thu, Aug 17, 2017 at 14:38:42 +0200, Frederic Cambus may have written: > Hi ports@, > > Here is a diff to update tcsh to 6.20.00. > > Comments? OK? I've been running with this in heavy daily use for the past week, and haven't encountered any problems. OpenBSD 6.2-beta (GENERIC.MP) #45: Sat Aug 26 10:24:08 MDT 2017 dera...@amd64.openbsd.org:/usr/src/sys/arch/amd64/compile/GENERIC.MP -- "With your own code to haunt you, who needs users?" -- Maarten Wiltink
Re: [PATCH] make TortoiseHg work again
Thu, Feb 26, 2015 at 11:13:46 +0100, Remi Pointel may have written: Le 23-02-2015 18:31, Matthew Clarke a ??crit??: Hi. [ Sent this to the devel/tortoisehg maintainer a week ago yesterday. ] [ No response yet. Is this too late for 5.7? ] Since the Mercurial update from 3.0 to 3.2.3, TortoiseHg 3.0 refuses to run, saying it requires Mercurial 3.0.n or 3.1.n. Here's a diff to update TortoiseHg to version 3.3, which requires Mercurial 3.2.n or 3.3.n. Lightly tested on i386. Hi, sorry for the delay... I don't understand why you need the config.py part? Yeah, I should have explained that bit. This is the diff, does it works fine for you? Cheers, Remi. No, sorry. TortoiseHg 3.3 determines the location of the following files at run-time: - The license terms text file, when the License button on the About TortoiseHg dialogue is pressed. - The translation (tortoisehg.mo) files, used if a language other than English is requested (e.g., via LANG or LC_ALL). - The .png file containing the TortoiseHg logo to show at the top of the About TortoiseHg dialogue. (Actually, 3.0 did this too, but I didn't notice until testing 3.3.) The code that determines those locations is installed as lib/python/${MODPY_VERSION}/site-packages/tortoisehg/util/paths.py, and it tries to import the locations as variable settings from lib/python/${MODPY_VERSION}/site-packages/tortoisehg/util/config.py. If we don't install a config.py, the code in paths.py derives locations relative to the grandparent of the directory holding paths.py, so TortoiseHg ends up looking for the following, none of which exist. - The license terms in file lib/python/${MODPY_VERSION}/site-packages/COPYING.txt. - The translation files in sub-directories of lib/python/${MODPY_VERSION}/site-packages/locale. - The .png file containing the TortoiseHg logo in lib/python/${MODPY_VERSION}/site-packages/icons. I created config.py in post-extract, so it would exist before TortoiseHg setup.py was run, because that setup.py includes the following code: -- # Create a config.py. Distributions will need to supply their own cfgfile = os.path.join('tortoisehg', 'util', 'config.py') if not os.path.exists(cfgfile) and not os.path.exists('.hg/requires'): f = open(cfgfile, w) f.write('bin_path = /usr/bin\n') f.write('license_path = /usr/share/doc/tortoisehg/Copying.txt.gz\n') f.write('locale_path = /usr/share/locale\n') f.write('icon_path= /usr/share/pixmaps/tortoisehg/icons\n') f.write('nofork = True\n') f.close() -- With that config.py created by setup.py, TortoiseHg ends up looking for the following, which don't exist or don't contain what TortoiseHg is looking for. - The license terms in file /usr/share/doc/tortoisehg/Copying.txt.gz. - The translation files in sub-directories of /usr/share/locale. - The .png file containing the TortoiseHg logo in /usr/share/pixmaps/tortoisehg/icons. There's already a patch for setup.py, and I thought about extending it to patch the block of code that creates config.py, but I haven't learned yet how to combine ${SUBST} and make patch. A separate files/config.py and post-extract: ${SUBST_DATA} in Makefile seemed cleaner. A couple of other things I should have explained about my proposed diff: 1. RUN_DEPENDS on devel/mercurial. I proposed this change in Makefile: -- @@ -23,9 +23,19 @@ RUN_DEPENDS = ${BUILD_DEPENDS} \ editors/py-qscintilla \ - devel/mercurial=3.0 \ + devel/mercurial=3.2,3.4 \ devel/py-iniparse -- I got the Mercurial 3.0 - 3.2.3 update on a test machine a few weeks ago, but it wasn't until a couple of weeks later that I next tried to actually run hg and thg on that test machine, and found out that TortoiseHg 3.0 refused to run with Mercurial 3.2.3. Tightening the devel/mercurial RDEP in the TortoiseHg package would have made the problem visible when pkg_add updated Mercurial. Hmm. I wonder if I can automate pulling the Mercurial RDEP versions from the TortoiseHg sources. No time now, and that can wait until after 5.7 anyway. For now, I've included at the end of this message an updated proposed patch with a comment before the RUN_DEPENDS block in Makefile. 2. Packaging COPYING.txt. I proposed these changes in Makefile: -- NO_TEST = Yes + +SHAREDIR
[PATCH] make TortoiseHg work again
Hi. [ Sent this to the devel/tortoisehg maintainer a week ago yesterday. ] [ No response yet. Is this too late for 5.7? ] Since the Mercurial update from 3.0 to 3.2.3, TortoiseHg 3.0 refuses to run, saying it requires Mercurial 3.0.n or 3.1.n. Here's a diff to update TortoiseHg to version 3.3, which requires Mercurial 3.2.n or 3.3.n. Lightly tested on i386. Index: Makefile === RCS file: /cvs/ports/devel/tortoisehg/Makefile,v retrieving revision 1.16 diff -u -r1.16 Makefile --- Makefile19 May 2014 19:16:17 - 1.16 +++ Makefile16 Feb 2015 03:59:24 - @@ -2,7 +2,7 @@ COMMENT = series of applications for Mercurial -MODPY_EGG_VERSION =3.0 +MODPY_EGG_VERSION =3.3 DISTNAME = tortoisehg-${MODPY_EGG_VERSION} CATEGORIES = devel @@ -23,9 +23,19 @@ RUN_DEPENDS = ${BUILD_DEPENDS} \ editors/py-qscintilla \ - devel/mercurial=3.0 \ + devel/mercurial=3.2,3.4 \ devel/py-iniparse NO_TEST = Yes + +SHAREDIR = ${PREFIX}/share/tortoisehg + +post-extract: + @${SUBST_DATA} -c ${FILESDIR}/config.py \ + ${WRKSRC}/tortoisehg/util/config.py + +post-install: + ${INSTALL_DATA_DIR} ${SHAREDIR} + ${INSTALL_DATA} ${WRKSRC}/COPYING.txt ${SHAREDIR} .include bsd.port.mk Index: distinfo === RCS file: /cvs/ports/devel/tortoisehg/distinfo,v retrieving revision 1.13 diff -u -r1.13 distinfo --- distinfo19 May 2014 19:16:17 - 1.13 +++ distinfo14 Feb 2015 21:08:51 - @@ -1,2 +1,2 @@ -SHA256 (tortoisehg-3.0.tar.gz) = ylNx0NcgsPViEBNhLlReo4b+krXfSpxpa3LbvnFupKc= -SIZE (tortoisehg-3.0.tar.gz) = 8237239 +SHA256 (tortoisehg-3.3.tar.gz) = CHrwKLtUohYqMaLdgaFKPx8eEurd5CiQpImrEfcgHNs= +SIZE (tortoisehg-3.3.tar.gz) = 8210712 Index: files/config.py === RCS file: files/config.py diff -N files/config.py --- /dev/null 1 Jan 1970 00:00:00 - +++ files/config.py 16 Feb 2015 04:30:28 - @@ -0,0 +1,6 @@ +# $OpenBSD$ +bin_path = ${TRUEPREFIX}/bin +license_path = ${TRUEPREFIX}/share/tortoisehg/COPYING.txt +locale_path = ${TRUEPREFIX}/share/locale +icon_path= ${TRUEPREFIX}/share/pixmaps/tortoisehg/icons +nofork = False Index: pkg/PLIST === RCS file: /cvs/ports/devel/tortoisehg/pkg/PLIST,v retrieving revision 1.10 diff -u -r1.10 PLIST --- pkg/PLIST 19 May 2014 19:16:17 - 1.10 +++ pkg/PLIST 16 Feb 2015 03:45:51 - @@ -80,16 +80,14 @@ lib/python${MODPY_VERSION}/site-packages/tortoisehg/hgqt/htmldelegate.pyc lib/python${MODPY_VERSION}/site-packages/tortoisehg/hgqt/htmlui.py lib/python${MODPY_VERSION}/site-packages/tortoisehg/hgqt/htmlui.pyc -lib/python${MODPY_VERSION}/site-packages/tortoisehg/hgqt/i18n.py -lib/python${MODPY_VERSION}/site-packages/tortoisehg/hgqt/i18n.pyc lib/python${MODPY_VERSION}/site-packages/tortoisehg/hgqt/icons_rc.py lib/python${MODPY_VERSION}/site-packages/tortoisehg/hgqt/icons_rc.pyc +lib/python${MODPY_VERSION}/site-packages/tortoisehg/hgqt/infobar.py +lib/python${MODPY_VERSION}/site-packages/tortoisehg/hgqt/infobar.pyc lib/python${MODPY_VERSION}/site-packages/tortoisehg/hgqt/lexers.py lib/python${MODPY_VERSION}/site-packages/tortoisehg/hgqt/lexers.pyc lib/python${MODPY_VERSION}/site-packages/tortoisehg/hgqt/lfprompt.py lib/python${MODPY_VERSION}/site-packages/tortoisehg/hgqt/lfprompt.pyc -lib/python${MODPY_VERSION}/site-packages/tortoisehg/hgqt/license.py -lib/python${MODPY_VERSION}/site-packages/tortoisehg/hgqt/license.pyc lib/python${MODPY_VERSION}/site-packages/tortoisehg/hgqt/manifestmodel.py lib/python${MODPY_VERSION}/site-packages/tortoisehg/hgqt/manifestmodel.pyc lib/python${MODPY_VERSION}/site-packages/tortoisehg/hgqt/matching.py @@ -102,8 +100,6 @@ lib/python${MODPY_VERSION}/site-packages/tortoisehg/hgqt/modeltest.pyc lib/python${MODPY_VERSION}/site-packages/tortoisehg/hgqt/mq.py lib/python${MODPY_VERSION}/site-packages/tortoisehg/hgqt/mq.pyc -lib/python${MODPY_VERSION}/site-packages/tortoisehg/hgqt/mqutil.py -lib/python${MODPY_VERSION}/site-packages/tortoisehg/hgqt/mqutil.pyc lib/python${MODPY_VERSION}/site-packages/tortoisehg/hgqt/p4pending.py lib/python${MODPY_VERSION}/site-packages/tortoisehg/hgqt/p4pending.pyc lib/python${MODPY_VERSION}/site-packages/tortoisehg/hgqt/pbranch.py @@ -112,6 +108,8 @@ lib/python${MODPY_VERSION}/site-packages/tortoisehg/hgqt/postreview.pyc lib/python${MODPY_VERSION}/site-packages/tortoisehg/hgqt/postreview_ui.py lib/python${MODPY_VERSION}/site-packages/tortoisehg/hgqt/postreview_ui.pyc +lib/python${MODPY_VERSION}/site-packages/tortoisehg/hgqt/prune.py +lib/python${MODPY_VERSION}/site-packages/tortoisehg/hgqt/prune.pyc
Re: net/psi 0.15 from packages does not connect to any jabber server
Sun, Jul 14, 2013 at 22:37:31 +0600, Alexandr Shadchin may have written: On Fri, Jun 14, 2013 at 04:43:11PM +0500, dmitry.sensei wrote: Hi! net/psi 0.15 from packages does not connect to any jabber server. for exapmple vk.com jabber.ru xmpp.yandex.ru talkr.im xmpp.odnoklassniki.ru chat.facebook.com. [snip] I'm sorry for the long answer, too little time. This diff should solve this issue. Now _res has minimal support: not define nsaddr_list, only nscount, therefore jdns could not determine dns server. Comments ? Ok ? -- Alexandr Shadchin [ snip patch ] I was about to add a me too to dmitry.sensei's problem report when I saw the patch. Patch works for me; psi can connect to talk.google.com again. Thanks. -- On two occasions I have been asked [by members of Parliament!], `Pray, Mr. Babbage, if you put into the machine wrong figures, will the right answers come out?' I am not able rightly to apprehend the kind of confusion of ideas that could provoke such a question. -- Charles Babbage
Re: UPDATE: tcsh - 6.16.00
Thu, Jul 16, 2009 at 03:41:50PM +0100, Stuart Henderson may have written: On 2009/07/17 00:25, Ian McWilliam wrote: Stuart Henderson wrote: I don't use it myself, could any tcsh users please give this update a spin? thanks. Works fine for me on Amd64. version tcsh 6.16.00 (Astron) 2008-09-30 (amd64-amd-OpenBSD) options wide,nls,dl,al,kan,sm,rh,color,filec Version 6.17.00 was released on July 10, so maybe we should take that for a spin. ftp://ftp.astron.com/pub/tcsh/ thanks, nice of them to update the website :-) Used it all day today; no problems. tcsh 6.17.00 (Astron) 2009-07-10 (i386-intel-OpenBSD) options wide,nls,dl,al,kan,sm,rh,color,filec -- Shoot for the moon. If you miss, you're among stars.
Re: UPDATE: rdesktop 1.6.0
Wed, Jun 18, 2008 at 12:53:39PM +0200, Giovanni Bechis may have written: It works well @i386, PLIST updated and PKGNAME removed. Cheers Giovanni [ snip patch ] Works well on i386 here. - connect to XP Pro on laptop at 1024x768 - connect to XP Pro on other laptop at 1680x1050 - connect to XP Pro in qemu on local system at 1680x1050 You might want to add libsamplerate (audio/libsamplerate) to the port's dependencies, as that library will be found and used if it is installed. I could't tell that the presence of libsamplerate made any difference to the few sounds that I've heard during testing so far, though. -- Debugging is twice as hard as writing the code in the first place. Therefore, if you write the code as cleverly as possible, you are, by definition, not smart enough to debug it. -- Kernighan -- quoted by Quentyn Taylor
typo in net/yt/pkg/DESCR
No MAINTAINER listed in Makefile, so hello [EMAIL PROTECTED] Trivial typo fix: Index: net/yt/pkg/DESCR === RCS file: /cvs/ports/net/yt/pkg/DESCR,v retrieving revision 1.2 diff -u -r1.2 DESCR --- net/yt/pkg/DESCR21 Feb 2007 13:50:18 - 1.2 +++ net/yt/pkg/DESCR30 Oct 2007 19:31:46 - @@ -1,3 +1,3 @@ yt is an extensible Lua script to fetch videos from YouTube.com. You can -(and probaly should) customize it to use your preferred encoding tool. +(and probably should) customize it to use your preferred encoding tool. By default it uses graphics/ffmpeg. -- Perfection [in design] is achieved not when there is nothing left to add, but rather when there is nothing left to take away. -- Antoine de Saint-Exupery