UPDATE: sysutils/entr to 3.7
Hi, Update to 3.7. All unit tests pass. OK? CC'ing maintainer/author. Changelog: - Terminate subprocess in restart mode if a file under watch disappears - Allow NOTE_ATTRIB to set '/_' only if file mode changes - New '-s' option executes commands using $SHELL -c - Print usage and exit if input is from a terminal instead of a pipe Index: Makefile === RCS file: /cvs/ports/sysutils/entr/Makefile,v retrieving revision 1.9 diff -u -p -r1.9 Makefile --- Makefile23 Oct 2016 00:21:24 - 1.9 +++ Makefile22 Mar 2017 01:59:54 - @@ -2,8 +2,8 @@ COMMENT = run arbitrary commands when files change -DISTNAME = entr-3.6 -REV = c15b0be493fc +DISTNAME = entr-3.7 +REV = c5b62bde107d CATEGORIES = sysutils Index: distinfo === RCS file: /cvs/ports/sysutils/entr/distinfo,v retrieving revision 1.6 diff -u -p -r1.6 distinfo --- distinfo23 Oct 2016 00:21:24 - 1.6 +++ distinfo22 Mar 2017 01:59:54 - @@ -1,2 +1,2 @@ -SHA256 (entr-3.6.tar.gz) = pCdG2BxUjX5VfVAPk0IrjslzHXGTCesmAbi+aa4NyOs= -SIZE (entr-3.6.tar.gz) = 22548 +SHA256 (entr-3.7.tar.gz) = lO/VDI9+nVaQYNXe6/Nmw1ZegegUqzMrlz1ymPqOoi8= +SIZE (entr-3.7.tar.gz) = 23326
Re: UPDATE: net/gpodder 3.9.1 -> 3.9.3
On Tue, March 7, 2017 11:39 pm, trondd wrote: > And pinging this one up as well. Is there anybody in there? > > On Fri, February 24, 2017 9:08 pm, trondd wrote: >> Update gPodder to 3.9.3 Various bug fixes. Requires the new >> py-podcastparser >> as they replaced py-feedparser with their own library. >> >> Tim. >> >> Index: Makefile >> === >> RCS file: /cvs/ports/net/gpodder/Makefile,v >> retrieving revision 1.10 >> diff -u -p -r1.10 Makefile >> --- Makefile 23 Oct 2016 01:59:16 - 1.10 >> +++ Makefile 25 Feb 2017 01:34:45 - >> @@ -2,7 +2,7 @@ >> >> COMMENT=media aggregator and podcast client >> >> -MODPY_EGG_VERSION= 3.9.1 >> +MODPY_EGG_VERSION= 3.9.3 >> DISTNAME= gpodder-${MODPY_EGG_VERSION} >> CATEGORIES= net www >> >> @@ -23,7 +23,7 @@ BUILD_DEPENDS= textproc/intltool >> RUN_DEPENDS=devel/desktop-file-utils \ >> devel/gettext \ >> net/mygpoclient \ >> -textproc/py-feedparser \ >> +textproc/py-podcastparser \ >> x11/dbus-python \ >> x11/gtk+3,-guic \ >> x11/py-gtk2 >> Index: distinfo >> === >> RCS file: /cvs/ports/net/gpodder/distinfo,v >> retrieving revision 1.3 >> diff -u -p -r1.3 distinfo >> --- distinfo 23 Oct 2016 01:59:16 - 1.3 >> +++ distinfo 25 Feb 2017 01:34:45 - >> @@ -1,2 +1,2 @@ >> -SHA256 (gpodder-3.9.1.tar.gz) = >> zu0Fphh368dYcAwG8paODpMlYqAgWb8nmG+nEglGDbo= >> -SIZE (gpodder-3.9.1.tar.gz) = 1009899 >> +SHA256 (gpodder-3.9.3.tar.gz) = >> LLgOxTKuOSnxPM+UY/+6mpwMfsi3I6va40umQRMYNcA= >> +SIZE (gpodder-3.9.3.tar.gz) = 1008774 >> Index: pkg/PLIST >> === >> RCS file: /cvs/ports/net/gpodder/pkg/PLIST,v >> retrieving revision 1.3 >> diff -u -p -r1.3 PLIST >> --- pkg/PLIST23 Oct 2016 01:59:16 - 1.3 >> +++ pkg/PLIST25 Feb 2017 01:34:45 - >> @@ -6,183 +6,183 @@ lib/python${MODPY_VERSION}/site-packages >> >> lib/python${MODPY_VERSION}/site-packages/gpodder-${MODPY_EGG_VERSION}-py${MODPY_VERSION}.egg-info >> lib/python${MODPY_VERSION}/site-packages/gpodder/__init__.py >> lib/python${MODPY_VERSION}/site-packages/gpodder/__init__.pyc >> -lib/python${MODPY_VERSION}/site-packages/gpodder/__init__.pyo >> +lib/python${MODPY_VERSION}/site-packages/gpodder/__init__.${MODPY_PYOEXTENSION} >> lib/python${MODPY_VERSION}/site-packages/gpodder/common.py >> lib/python${MODPY_VERSION}/site-packages/gpodder/common.pyc >> -lib/python${MODPY_VERSION}/site-packages/gpodder/common.pyo >> +lib/python${MODPY_VERSION}/site-packages/gpodder/common.${MODPY_PYOEXTENSION} >> lib/python${MODPY_VERSION}/site-packages/gpodder/config.py >> lib/python${MODPY_VERSION}/site-packages/gpodder/config.pyc >> -lib/python${MODPY_VERSION}/site-packages/gpodder/config.pyo >> +lib/python${MODPY_VERSION}/site-packages/gpodder/config.${MODPY_PYOEXTENSION} >> lib/python${MODPY_VERSION}/site-packages/gpodder/core.py >> lib/python${MODPY_VERSION}/site-packages/gpodder/core.pyc >> -lib/python${MODPY_VERSION}/site-packages/gpodder/core.pyo >> +lib/python${MODPY_VERSION}/site-packages/gpodder/core.${MODPY_PYOEXTENSION} >> lib/python${MODPY_VERSION}/site-packages/gpodder/coverart.py >> lib/python${MODPY_VERSION}/site-packages/gpodder/coverart.pyc >> -lib/python${MODPY_VERSION}/site-packages/gpodder/coverart.pyo >> +lib/python${MODPY_VERSION}/site-packages/gpodder/coverart.${MODPY_PYOEXTENSION} >> lib/python${MODPY_VERSION}/site-packages/gpodder/dbsqlite.py >> lib/python${MODPY_VERSION}/site-packages/gpodder/dbsqlite.pyc >> -lib/python${MODPY_VERSION}/site-packages/gpodder/dbsqlite.pyo >> +lib/python${MODPY_VERSION}/site-packages/gpodder/dbsqlite.${MODPY_PYOEXTENSION} >> lib/python${MODPY_VERSION}/site-packages/gpodder/dbusproxy.py >> lib/python${MODPY_VERSION}/site-packages/gpodder/dbusproxy.pyc >> -lib/python${MODPY_VERSION}/site-packages/gpodder/dbusproxy.pyo >> +lib/python${MODPY_VERSION}/site-packages/gpodder/dbusproxy.${MODPY_PYOEXTENSION} >> lib/python${MODPY_VERSION}/site-packages/gpodder/directory.py >> lib/python${MODPY_VERSION}/site-packages/gpodder/directory.pyc >> -lib/python${MODPY_VERSION}/site-packages/gpodder/directory.pyo >> +lib/python${MODPY_VERSION}/site-packages/gpodder/directory.${MODPY_PYOEXTENSION} >> lib/python${MODPY_VERSION}/site-packages/gpodder/download.py >> lib/python${MODPY_VERSION}/site-packages/gpodder/download.pyc >> -lib/python${MODPY_VERSION}/site-packages/gpodder/download.pyo >> +lib/python${MODPY_VERSION}/site-packages/gpodder/download.${MODPY_PYOEXTENSION} >> lib/python${MODPY_VERSION}/site-packages/gpodder/escapist_videos.py >> lib/python${MODPY_VERSION}/site-packages/gpodder/escapist_videos.pyc >> -lib/python${MODPY_VERSION}/site-packages/
Re: NEW: textproc/py-podcastparser
On Tue, March 7, 2017 11:37 pm, trondd wrote: > On Sat, February 25, 2017 12:14 pm, trondd wrote: >> Stuart Henderson wrote: >> >>> On 2017/02/24 20:47, trondd wrote: >>> > On Thu, February 23, 2017 11:26 pm, trondd wrote: >>> > > This is new dependency for net/gpodder. They replaced >>> py-feedparser >>> with >>> > > their own parser. >>> > > >>> > > Pretty simple so I wouldn't be surprised if I've misssed something >>> or did >>> > > something I didn't need to. Dealing with the man page might be >>> iffy. >>> > > >>> > > The update to net/gpodder will follow. >>> > > >>> > > Tim. >>> > > >>> > >>> > Tweaking previous. They ship a makefile for running the tests so I >>> don't >>> > need my own do-test. >>> >>> >>> The attachment was sent as inline text/plain which doesn't work too >>> well. >> >> *sigh* It's always something, it seems. >> >> Sorry. Trying again. >> > > Pinging to the top. > Anyone...anyone? http://marc.info/?l=openbsd-ports&m=148804290503994&w=2
Re: possible chromium regression on -current
It works fine on 57.0.2987.110. On Mon, Mar 20, 2017 at 4:04 AM, Martijn Rijkeboer wrote: > On 03/17/17 09:53, Dimitris Papastamos wrote: >> Hi, >> >> It works for me now with chromium-57.0.2987.98 and the snapshot >> from the 16th of March. > > It works for me too. However, Iridium (54.0) still doesn't work. > > Kind regards, > > > Martijn Rijkeboer >
Re: change dnsdist uid/gid
On 2017/03/21 13:45, Daniel Jakots wrote: > On Tue, 21 Mar 2017 13:03:02 -0400, Daniel Jakots > wrote: > > > Hi, > > > > Pavel Korovin noticed that gitea and dnsdist use the same user [1]. > > jca@ told me that changing dnsdist should be easier because dnsdist > > doesn't/shouldn't own anything on the disk. > > > > Here are three diffs to change that. > > > > I don't cc dnsdist maintainer, I chatted with him about it and he > > didn't have any complaint about it. > > > > [1]: https://marc.info/?l=openbsd-ports&m=149009685827196&w=2 > > New current.html after tb's review. Users actually only need to change it if they want to install both of them on the same machine. > Index: faq/current.html > === > RCS file: /cvs/www/faq/current.html,v > retrieving revision 1.801 > diff -u -p -r1.801 current.html > --- faq/current.html 3 Mar 2017 11:17:32 - 1.801 > +++ faq/current.html 21 Mar 2017 17:37:41 - > @@ -631,6 +631,18 @@ It is advised to update > http://man.openbsd.org/sysctl.conf.5";>sysctl.conf(5) > to use the new name. > > + > + > +2017/03/21 - [ports] net/dnsdist UID change > + > +The net/dnsdist port now uses uid and gid 788. > +If you installed dnsdist, run the following commands. > + > + > +# groupmod -g 788 _dnsdist > +# usermod -g 788 -u 788 _dnsdist > + > + > > > >
Re: www/clearsilver
On 2017/03/21 12:05, Okan Demirmen wrote: > Anyone care if this goes? > > On Thu, Apr 28, 2016 at 12:00 PM, Okan Demirmen wrote: > > > Anyone still using or interested in this port? It's fairly old, > > unmaintained upstream and was never updated even to the 2007 version > > due to the license change to Apache2. > > > > Zap it? or any takers? > > > > Thanks. > > I'm OK with removing it, note that games/sudoku-solver depends on it.
Re: Update to gitea-1.1.0
On Tue, Mar 21, 2017 at 02:46:48PM +0300, Pavel Korovin wrote: > On 03/17, Daniel Jakots wrote: > > > > Here's a diff to update gitea to 1.1.0. ChangeLog is quite long and it > > says: > > > > Hi Daniel, > Works fine for me as well, thank you! > > You forgot to allocate the user/group id for _gitea user in > $PORTSDIR/infrastructure/db/user.list, and 787 is already allocated: > > 787 _dnsdist_dnsdistnet/dnsdist > > Please consider adding _gitea entry in user.list and change the gitea userid > in PLIST so there's no clash. > > Do you use it with ssh? If not, do you consider the default configuration > should allow ssh connections? > I found the defaults quite misleading with system's default sshd, gitea's SSH > port set to , _gitea home set to /nonexistent and shell set to > /sbin/nologin. ssh won't work this way. > > Maybe we'd better use OpenBSD's default port 22 and > @newgroup _gitea:788 > @newuser _gitea:788:_gitea:daemon:Gitea Account:/var/gitea:/usr/local/bin/bash > > ? No, the port is used by the builtin SSH server (for git). If the program doesn't need a shell, then nologin the correct there. -- Juan Francisco Cantero Hurtado http://juanfra.info
Re: update: lang/rust to 1.16.0
Hi, On Tue, Mar 21, 2017 at 06:22:05PM +0100, Landry Breuil wrote: > From what i understand, it's more work to keep rustdoc in the bootstrap > so that we can generate the docs at package build time. Well, time to cut my losses. The diff I had been preparing today simply added back rustdoc into the bootstrap. rustdoc is (uncompressed) an 8KB binary, so I figured it's not really a burden upon us to have it in the bootstrap. As Landry says, it's ultimately up to the maintainer and I don't feel strongly enough to argue further. FWIW, the diff follows. A new amd64 bootstrap is here: http://theunixzoo.co.uk/random/rustc-bootstrap-amd64-1.16.0-20170321.tar.gz The diff also kills some trailing whitespace. Seeya! Index: Makefile === RCS file: /home/edd/cvsync/ports/lang/rust/Makefile,v retrieving revision 1.38 diff -u -p -r1.38 Makefile --- Makefile18 Feb 2017 13:15:48 - 1.38 +++ Makefile21 Mar 2017 16:13:04 - @@ -3,16 +3,15 @@ # snapshots are only available for amd64 and i386, for now ONLY_FOR_ARCHS = amd64 i386 -COMMENT-main = compiler for Rust Language -COMMENT-doc = html documentation for rustc +COMMENT-main = compiler for Rust Language +COMMENT-doc = html documentation for rustc -V =1.15.1 +V =1.16.0 DISTNAME = rustc-${V}-src -REVISION = 0 # rustc bootstrap version -RBV-amd64 =1.15.0-20170203 -RBV-i386 = 1.15.0-20170204 +RBV-amd64 = 1.16.0-20170321 +RBV-i386 = 1.16.0-20170312 RBV = ${RBV-${MACHINE_ARCH}} # cargo bootstrap version @@ -95,9 +94,24 @@ USE_GMAKE = Yes # need for libbacktrace USE_LIBTOOL = gnu -TEST_DEPENDS +=${FULLPKGNAME-main}:${BUILD_PKGPATH} \ +TEST_DEPENDS +=${FULLPKGNAME}:${BUILD_PKGPATH} \ devel/git +TEST_TARGET ?= src/test/run-pass \ + src/test/compile-fail \ + src/test/parse-fail \ + src/test/run-fail \ + src/test/mir-opt \ + src/test/incremental \ + src/test/ui \ + src/test/compile-fail-fulldeps \ + src/test/run-make \ + src/test/rustdoc \ + src/test/pretty \ + src/test/run-pass/pretty \ + src/test/run-fail/pretty \ + path/to/nowhere + # - check datasize limit before configuring (and building) pre-configure: @if [ `ulimit -d` -lt 1572864 ]; then \ @@ -116,11 +130,14 @@ do-configure: >>${WRKBUILD}/config.toml echo 'prefix = "${LOCALBASE}"' >>${WRKBUILD}/config.toml echo 'vendor = true' >>${WRKBUILD}/config.toml - + echo '[rust]' >>${WRKBUILD}/config.toml echo 'channel = "stable"' >>${WRKBUILD}/config.toml echo 'codegen-tests = false' >>${WRKBUILD}/config.toml - + + echo '[dist]' >>${WRKBUILD}/config.toml + echo 'src-tarball = false' >>${WRKBUILD}/config.toml + .if ${PORTS_LLVM:L:Myes} echo '[target.${TRIPLE_ARCH}]' >>${WRKBUILD}/config.toml echo 'llvm-config = "${LOCALBASE}/bin/llvm-config"' \ @@ -130,7 +147,7 @@ do-configure: echo 'static-libstdcpp = false' >>${WRKBUILD}/config.toml echo 'ninja = true' >>${WRKBUILD}/config.toml .endif - + rm -rf ${WRKDIR}/modgcc-libs mkdir -p ${WRKDIR}/modgcc-libs cp ${LOCALBASE}/lib/libestdc++.so.${LIBESTDC_VERSION} \ @@ -142,10 +159,9 @@ BUILD_BIN = cd ${WRKBUILD} && exec ${SET do-build: ${BUILD_BIN} dist --verbose --jobs=${MAKE_JOBS} -COMPONENTS ?= rustc rust-std rust-docs do-install: rm -rf ${WRKBUILD}/_extractdist -.for _c in ${COMPONENTS} +.for _c in rustc rust-std rust-docs mkdir ${WRKBUILD}/_extractdist cd ${WRKBUILD}/_extractdist && tar zxf \ ${WRKBUILD}/build/dist/${_c}-${V}-${TRIPLE_ARCH}.tar.gz @@ -169,7 +185,7 @@ post-install: ${PREFIX}/lib/rustlib/manifest-* do-test: - ${BUILD_BIN} test --verbose --jobs=${MAKE_JOBS} + ${BUILD_BIN} test --verbose --jobs=${MAKE_JOBS} ${TEST_TARGET} # bootstrap target permits to regenerate the bootstrap archive BOOTSTRAPDIR=${WRKDIR}/rustc-bootstrap-${MACHINE_ARCH}-${V}-new @@ -179,10 +195,9 @@ bootstrap: build ${MAKE} clean=fake ${MAKE} fake \ PREFIX="${BOOTSTRAPDIR}" \ - COMPONENTS="rustc rust-std" \ FAKE_SETUP="" rm -rf ${BOOTSTRAPDIR}/{man,share} \ - ${BOOTSTRAPDIR}/bin/{rust-gdb,rust
update zkt to 1.1.4
Hi, zkt-1.1.4 is avaiable since 09 May 2016. It fixes some bug and is able to use stronger algorithms. I'm completely new on ports so please excuse me if I don't follow habits. Regards. --- Index: Makefile === RCS file: /cvs/ports/security/zkt/Makefile,v retrieving revision 1.11 diff -u -p -r1.11 Makefile --- Makefile3 Sep 2014 08:54:41 - 1.11 +++ Makefile21 Mar 2017 18:56:37 - @@ -2,7 +2,7 @@ COMMENT= DNSsec Zone Key Tool -DISTNAME= zkt-1.1.0 +DISTNAME= zkt-1.1.4 REVISION= 0 CATEGORIES=security Index: distinfo === RCS file: /cvs/ports/security/zkt/distinfo,v retrieving revision 1.7 diff -u -p -r1.7 distinfo --- distinfo18 Jan 2015 03:15:08 - 1.7 +++ distinfo21 Mar 2017 18:56:37 - @@ -1,2 +1,2 @@ -SHA256 (zkt-1.1.0.tar.gz) = iH4EzlqoXYSvV4UnlNI4jL1nUWpeXJ/NG4Jw0vQeeiY= -SIZE (zkt-1.1.0.tar.gz) = 343811 +SHA256 (zkt-1.1.4.tar.gz) = cEBlpHKhwgnEfm0NAAOzQ0yqEBwaXQhEaIlkP8q6fMc= +SIZE (zkt-1.1.4.tar.gz) = 382629 cvs server: Diffing patches cvs server: Diffing pkg signature.asc Description: PGP signature
Re: productivity/wyrd (remove?)
On Tue 2017.03.21 at 12:22 -0400, kwesterb...@gmail.com wrote: > > > > On Mar 21, 2017, at 12:09 PM, Jeremie Courreges-Anglas > > wrote: > > > > Daniel Dickman writes: > > > >>> On Mar 13, 2017, at 11:47 AM, Okan Demirmen wrote: > >>> > >>> Hi, > >>> > >>> Quick poll - anyone want to take this? It's dead upstream for the past > >>> few years and I have no interest in it either, especially the way > >>> ocaml is moving. > >>> > >>> Removing is the idea otherwise. > >>> > >>> Thanks, > >>> Okan > >>> > >> > >> removing sounds like a good idea, as you say "especially given the way > >> ocaml is moving". > >> > >> surprised ken hasn't already axed this one yet. > > > > Just to chime in about this one: I have no problem with > > productivity/wyrd going away, but this generic statement: > > > > "especially given the way ocaml is moving" > > > > bugs me. A port written in OCaml will likely depend on OCaml libraries. > > The fact that those libraries can be installed through opam doesn't mean > > that we should remove the port and the aforementioned libraries. > > Exactly. My goal is only to remove unused libraries, libraries that make it > hard to use modern ocaml programming tools and intermediate ports. NOT any > end-user ports and the libraries they use. > > If wyrd is being used then I have no desire to remove it. > > I took "the way ocaml is moving" to be a comment on ocaml language and > ecosystem changes that made wyrd harder to keep running since it died > upstream. Yes, I was off-base there in where I was going. Really wyrd is going nowhere and that's the only point I meant to make.
archivers/unzip CVE
Hi, There are a bunch of CVE and other fixes for unzip in debian and redhat bug tracker. I added the links to the patch files. The fix for CVE-2014-9636 got an update, so we have a diff onto of a diff. Unfortunately unzip did its last release in 2009 and they do not offer patches on their web site. ok? bluhm Index: archivers/unzip/Makefile === RCS file: /data/mirror/openbsd/cvs/ports/archivers/unzip/Makefile,v retrieving revision 1.59 diff -u -p -r1.59 Makefile --- archivers/unzip/Makefile13 Sep 2016 11:44:06 - 1.59 +++ archivers/unzip/Makefile21 Mar 2017 15:24:39 - @@ -7,7 +7,7 @@ COMMENT = extract, list & test files in VERSION = 6.0 DISTNAME = unzip${VERSION:S/.//} PKGNAME = unzip-${VERSION} -REVISION = 9 +REVISION = 10 CATEGORIES = archivers MASTER_SITES = ${MASTER_SITE_SOURCEFORGE:=infozip/} \ ftp://ftp.info-zip.org/pub/infozip/src/ Index: archivers/unzip/patches/patch-crypt_c === RCS file: archivers/unzip/patches/patch-crypt_c diff -N archivers/unzip/patches/patch-crypt_c --- /dev/null 1 Jan 1970 00:00:00 - +++ archivers/unzip/patches/patch-crypt_c 21 Mar 2017 16:24:58 - @@ -0,0 +1,28 @@ +$OpenBSD$ + +Fix CVE-2015-7696: upstream fix for heap overflow +https://bugs.debian.org/802162 +https://bugzilla.redhat.com/show_bug.cgi?id=1260944 +https://bugzilla.redhat.com/attachment.cgi?id=1073002 + +--- crypt.c.orig Fri Jan 5 16:47:36 2007 crypt.cTue Mar 21 16:10:27 2017 +@@ -465,7 +465,17 @@ int decrypt(__G__ passwrd) + GLOBAL(pInfo->encrypted) = FALSE; + defer_leftover_input(__G); + for (n = 0; n < RAND_HEAD_LEN; n++) { +-b = NEXTBYTE; ++/* 2012-11-23 SMS. (OUSPG report.) ++ * Quit early if compressed size < HEAD_LEN. The resulting ++ * error message ("unable to get password") could be improved, ++ * but it's better than trying to read nonexistent data, and ++ * then continuing with a negative G.csize. (See ++ * fileio.c:readbyte()). ++ */ ++if ((b = NEXTBYTE) == (ush)EOF) ++{ ++return PK_ERR; ++} + h[n] = (uch)b; + Trace((stdout, " (%02x)", h[n])); + } Index: archivers/unzip/patches/patch-extract_c === RCS file: /data/mirror/openbsd/cvs/ports/archivers/unzip/patches/patch-extract_c,v retrieving revision 1.1 diff -u -p -r1.1 patch-extract_c --- archivers/unzip/patches/patch-extract_c 6 Feb 2015 21:37:04 - 1.1 +++ archivers/unzip/patches/patch-extract_c 21 Mar 2017 16:24:58 - @@ -1,11 +1,20 @@ $OpenBSD: patch-extract_c,v 1.1 2015/02/06 21:37:04 naddy Exp $ +Fix CVE-2015-7696: prevent unsigned overflow on invalid input +https://bugzilla.redhat.com/attachment.cgi?id=1075942 +https://bugzilla.redhat.com/show_bug.cgi?id=1260944 Fix CVE-2014-8139: CRC32 verification heap-based overflow +https://bugzilla.redhat.com/show_bug.cgi?id=1174844 +https://bugzilla.redhat.com/attachment.cgi?id=989833 Fix CVE-2014-8140: out-of-bounds write issue in test_compr_eb() Fix CVE-2014-9636: out-of-bounds read/write in test_compr_eb() +Fix CVE-2015-7697: infinite loop when extracting empty bzip2 data +https://bugs.debian.org/802160 +https://bugzilla.redhat.com/show_bug.cgi?id=1260944 +https://bugzilla.redhat.com/attachment.cgi?id=1073339 --- extract.c.orig Sat Mar 14 02:32:52 2009 -+++ extract.c Thu Feb 5 18:58:23 2015 extract.c Tue Mar 21 16:10:27 2017 @@ -1,5 +1,5 @@ /* - Copyright (c) 1990-2009 Info-ZIP. All rights reserved. @@ -22,7 +31,26 @@ Fix CVE-2014-9636: out-of-bounds read/wr static ZCONST char Far InvalidComprDataEAs[] = " invalid compressed data for EAs\n"; # if (defined(WIN32) && defined(NTSD_EAS)) -@@ -2023,7 +2025,8 @@ static int TestExtraField(__G__ ef, ef_len) +@@ -1255,8 +1257,17 @@ static int extract_or_test_entrylist(__G__ numchunk, + if (G.lrec.compression_method == STORED) { + zusz_t csiz_decrypted = G.lrec.csize; + +-if (G.pInfo->encrypted) ++if (G.pInfo->encrypted) { ++if (csiz_decrypted < 12) { ++/* handle the error now to prevent unsigned overflow */ ++Info(slide, 0x401, ((char *)slide, ++ LoadFarStringSmall(ErrUnzipNoFile), ++ LoadFarString(InvalidComprData), ++ LoadFarStringSmall2(Inflate))); ++return PK_ERR; ++} + csiz_decrypted -= 12; ++} + if (G.lrec.ucsize != csiz_decrypted) { + Info(slide, 0x401, ((char *)slide, + LoadFarStringSmall2(WrnStorUCSizCSizDiff), +@@ -2023,7 +2034,8 @@ static int TestExt
Re: change dnsdist uid/gid
On Tue, 21 Mar 2017 13:03:02 -0400, Daniel Jakots wrote: > Hi, > > Pavel Korovin noticed that gitea and dnsdist use the same user [1]. > jca@ told me that changing dnsdist should be easier because dnsdist > doesn't/shouldn't own anything on the disk. > > Here are three diffs to change that. > > I don't cc dnsdist maintainer, I chatted with him about it and he > didn't have any complaint about it. > > [1]: https://marc.info/?l=openbsd-ports&m=149009685827196&w=2 New current.html after tb's review. Index: faq/current.html === RCS file: /cvs/www/faq/current.html,v retrieving revision 1.801 diff -u -p -r1.801 current.html --- faq/current.html3 Mar 2017 11:17:32 - 1.801 +++ faq/current.html21 Mar 2017 17:37:41 - @@ -631,6 +631,18 @@ It is advised to update http://man.openbsd.org/sysctl.conf.5";>sysctl.conf(5) to use the new name. + + +2017/03/21 - [ports] net/dnsdist UID change + +The net/dnsdist port now uses uid and gid 788. +If you installed dnsdist, run the following commands. + + +# groupmod -g 788 _dnsdist +# usermod -g 788 -u 788 _dnsdist + +
Re: update: lang/rust to 1.16.0
On Tue, Mar 21, 2017 at 08:28:53AM +0100, Sebastien Marie wrote: > On Mon, Mar 20, 2017 at 10:11:29PM +, Edd Barrett wrote: > > On Mon, Mar 20, 2017 at 02:43:58PM +0100, Sebastien Marie wrote: > > > On Sun, Mar 19, 2017 at 03:55:26PM +, Edd Barrett wrote: > > > > > > > > How about a separate port for docs that depends on lang/rust? Would that > > > > be more palatable? > > > > > > I assume it could be possible, but I am unsure if rustbuild will be able > > > to > > > generate only the documentation (and not just binary + doc). Testing is > > > required :) > > > > I've been prodding around. There's a 'doc' target in the top-level > > Makefile.in, but we don't have a Makefile, presumably because we didn't > > use the configure script. > > > > If I run the configure script, then run `make doc` then it will try to > > build again (including failing to get a bootstrap binary from rustup). I > > think we'd need to use the configure script from the start, and I > > appreciate there is probably a good reason not to do that. > > configure script and Makefile are the legacy way to build rustc. There > are only keep for compatibility reason. The Makefile is done in a way to > call rustbuild. > > The lang/rust port uses rustbuild directly (it is more flexible for the > build options). The configuration is done by generating a config.toml > (see 'configure' target in the port). And call rustbuild directly for > the build (see 'do-build' target in the port). > > > But wait, earlier you said: > > > > > the main problem with shipping doc subpackages is now it requires > > > rustdoc binary to be present in the bootstrap archive too. > > > > Are you sure? We are already removing the rustdoc binary from the > > bootstrap tarball in the in-tree 0.15.x version, so how does the in-tree > > version make docs? It must be using it's own rustdoc? > > > > The bootstrap is generated in the same way that the port itself (and if > you read 'bootstrap' target in the port, you will see it reused the fake > framework but in different directory). So all files are installed in the > bootstrap directory, and it removes all unneeded files. Since the > beginning, rustdoc is removed from the required files to be present in > bootstrap. > > In 1.15.1, the rustbuild used the generated rustdoc binary for compiling > docs. It isn't the case anymore and it is why it is a problem. >From what i understand, it's more work to keep rustdoc in the bootstrap so that we can generate the docs at package build time, so to make things simpler sebastien wants to drop the -docs subpackage, which i fully agree with (less work for the maintainer). I havent tested the diff yet but the direction in which this is going makes sense to me. We're in 2017, docs are available online - and if they're really needed, maybe provide a separate port for them, without all the bootstrapping stuff that makes the rust port already quite complicated. Landry
change dnsdist uid/gid
Hi, Pavel Korovin noticed that gitea and dnsdist use the same user [1]. jca@ told me that changing dnsdist should be easier because dnsdist doesn't/shouldn't own anything on the disk. Here are three diffs to change that. I don't cc dnsdist maintainer, I chatted with him about it and he didn't have any complaint about it. [1]: https://marc.info/?l=openbsd-ports&m=149009685827196&w=2 Comments? OK? Cheers, DanielIndex: Makefile === RCS file: /cvs/ports/net/dnsdist/Makefile,v retrieving revision 1.2 diff -u -p -r1.2 Makefile --- Makefile 22 Jan 2017 12:13:09 - 1.2 +++ Makefile 21 Mar 2017 17:01:32 - @@ -5,6 +5,8 @@ COMMENT= highly DNS-, DoS- and abuse-aw DISTNAME= dnsdist-1.1.0 EXTRACT_SUFX= .tar.bz2 +REVISION= 0 + CATEGORIES= net HOMEPAGE= http://dnsdist.org/ Index: pkg/PLIST === RCS file: /cvs/ports/net/dnsdist/pkg/PLIST,v retrieving revision 1.1.1.1 diff -u -p -r1.1.1.1 PLIST --- pkg/PLIST 15 Jan 2017 17:13:33 - 1.1.1.1 +++ pkg/PLIST 21 Mar 2017 17:01:32 - @@ -1,7 +1,7 @@ @comment $OpenBSD: PLIST,v 1.1.1.1 2017/01/15 17:13:33 sthen Exp $ @extra ${SYSCONFDIR}/dnsdist/dnsdist.conf -@newgroup _dnsdist:787 -@newuser _dnsdist:787:_dnsdist:daemon:dnsdist DNS load balancer:/nonexistent:/sbin/nologin +@newgroup _dnsdist:788 +@newuser _dnsdist:788:_dnsdist:daemon:dnsdist DNS load balancer:/nonexistent:/sbin/nologin @bin bin/dnsdist @man man/man1/dnsdist.1 share/doc/dnsdist/ Index: user.list === RCS file: /cvs/ports/infrastructure/db/user.list,v retrieving revision 1.289 diff -u -p -r1.289 user.list --- user.list 12 Jan 2017 22:03:02 - 1.289 +++ user.list 21 Mar 2017 16:38:48 - @@ -295,4 +295,5 @@ id user group port options 784 _hping _hping net/hping 785 _uucp _uucp net/uucp 786 _riemann _riemann sysutils/riemann -787 _dnsdist _dnsdist net/dnsdist +787 _gitea _gitea www/gitea +788 _dnsdist _dnsdist net/dnsdist Index: faq/current.html === RCS file: /cvs/www/faq/current.html,v retrieving revision 1.801 diff -u -p -r1.801 current.html --- faq/current.html 3 Mar 2017 11:17:32 - 1.801 +++ faq/current.html 21 Mar 2017 16:08:51 - @@ -631,6 +631,17 @@ It is advised to update http://man.openbsd.org/sysctl.conf.5";>sysctl.conf(5) to use the new name. + +2017/03/21 - [ports] net/dnsdist UID change +The net/dnsdist port now uses uid and gid 788. If you installed dnsdist, +run the following commands. + + +# groupmod -g 788 _dnsdist +# usermod -u 788 _dnsdist +# usermod -g 788 _dnsdist + +
Re: productivity/wyrd (remove?)
> On Mar 21, 2017, at 12:09 PM, Jeremie Courreges-Anglas > wrote: > > Daniel Dickman writes: > >>> On Mar 13, 2017, at 11:47 AM, Okan Demirmen wrote: >>> >>> Hi, >>> >>> Quick poll - anyone want to take this? It's dead upstream for the past >>> few years and I have no interest in it either, especially the way >>> ocaml is moving. >>> >>> Removing is the idea otherwise. >>> >>> Thanks, >>> Okan >>> >> >> removing sounds like a good idea, as you say "especially given the way ocaml >> is moving". >> >> surprised ken hasn't already axed this one yet. > > Just to chime in about this one: I have no problem with > productivity/wyrd going away, but this generic statement: > > "especially given the way ocaml is moving" > > bugs me. A port written in OCaml will likely depend on OCaml libraries. > The fact that those libraries can be installed through opam doesn't mean > that we should remove the port and the aforementioned libraries. Exactly. My goal is only to remove unused libraries, libraries that make it hard to use modern ocaml programming tools and intermediate ports. NOT any end-user ports and the libraries they use. If wyrd is being used then I have no desire to remove it. I took "the way ocaml is moving" to be a comment on ocaml language and ecosystem changes that made wyrd harder to keep running since it died upstream. Ken > > -- > jca | PGP : 0x1524E7EE / 5135 92C1 AD36 5293 2BDF DDCC 0DFA 74AE 1524 E7EE
Re: www/clearsilver
I say junk it. But that's my 2 cents. On Mar 21, 2017 11:06 AM, "Okan Demirmen" wrote: Anyone care if this goes? On Thu, Apr 28, 2016 at 12:00 PM, Okan Demirmen wrote: > Anyone still using or interested in this port? It's fairly old, > unmaintained upstream and was never updated even to the 2007 version > due to the license change to Apache2. > > Zap it? or any takers? > > Thanks. >
Re: productivity/wyrd (remove?)
Daniel Dickman writes: >> On Mar 13, 2017, at 11:47 AM, Okan Demirmen wrote: >> >> Hi, >> >> Quick poll - anyone want to take this? It's dead upstream for the past >> few years and I have no interest in it either, especially the way >> ocaml is moving. >> >> Removing is the idea otherwise. >> >> Thanks, >> Okan >> > > removing sounds like a good idea, as you say "especially given the way ocaml > is moving". > > surprised ken hasn't already axed this one yet. Just to chime in about this one: I have no problem with productivity/wyrd going away, but this generic statement: "especially given the way ocaml is moving" bugs me. A port written in OCaml will likely depend on OCaml libraries. The fact that those libraries can be installed through opam doesn't mean that we should remove the port and the aforementioned libraries. -- jca | PGP : 0x1524E7EE / 5135 92C1 AD36 5293 2BDF DDCC 0DFA 74AE 1524 E7EE
Re: www/clearsilver
Anyone care if this goes? On Thu, Apr 28, 2016 at 12:00 PM, Okan Demirmen wrote: > Anyone still using or interested in this port? It's fairly old, > unmaintained upstream and was never updated even to the 2007 version > due to the license change to Apache2. > > Zap it? or any takers? > > Thanks. >
Re: productivity/wyrd (remove?)
> On Mar 13, 2017, at 11:47 AM, Okan Demirmen wrote: > > Hi, > > Quick poll - anyone want to take this? It's dead upstream for the past > few years and I have no interest in it either, especially the way > ocaml is moving. > > Removing is the idea otherwise. > > Thanks, > Okan > removing sounds like a good idea, as you say "especially given the way ocaml is moving". surprised ken hasn't already axed this one yet.
Re: phpmyadmin config and readme
On Tue, Mar 21, 2017 at 01:02:47PM +0200, Paul Irofti wrote: > On Tue, Mar 21, 2017 at 07:56:55AM +, Stuart Henderson wrote: > > On 2017/03/20 15:03, Paul Irofti wrote: > > > phpMyAdmin has been installed into ${INSTDIR}. > > > > > > If using Apache, you can make this accessible to clients by enabling > > > -the configuration file: > > > +mod_rewrite in ${SYSCONFDIR}/apache2/httpd2.conf, the mysql php module, > > > +and the configuration file: > > > > > > # ln -s ../modules.sample/phpmyadmin.conf /var/www/conf/modules/ > > > +# cp /etc/php-5.6.sample/mysqli.ini /etc/php-5.6/ > > > > Please symlink rather than cp, and use variables so this doesn't > > get stale after a PHP update: > > > > # ln -s ../php-${MODPHP_VERSION}.sample/mysqli.ini > > ${SYSCONFDIR}/php-${MODPHP_VERSION}/ > > Good point on the variables! > > I actually wanted to discus ln -s vs. cp. Isn't this a bad idea in > general? For the php bit I agree that it is useless to copy, but for the > phpmyadmin.conf I think it's actually required to cp instead of link: if > I customize the config I also modify the sample file which will get > overwritten by an update or pkg_add will scream about a bad checksum, > right? No, because phpmyadmin.conf is a @sample. It will only be overwritten if it was not locally modified. -- Antoine
Re: Update net/syncthing to 0.14.23
On Sun, Mar 19, 2017 at 02:23:51PM -0400, trondd wrote: > I was going to send this today, too. I upgraded yesterday. And here's the new release 0.14.25. OK anyone? Index: Makefile === RCS file: /home/edd/cvsync/ports/net/syncthing/Makefile,v retrieving revision 1.13 diff -u -p -r1.13 Makefile --- Makefile26 Dec 2016 09:59:16 - 1.13 +++ Makefile21 Mar 2017 11:25:23 - @@ -2,7 +2,7 @@ COMMENT = open decentralized synchronization utility -V =0.14.16 +V =0.14.25 DISTNAME = syncthing-${V} DISTFILES =syncthing-source-v${V}${EXTRACT_SUFX} Index: distinfo === RCS file: /home/edd/cvsync/ports/net/syncthing/distinfo,v retrieving revision 1.9 diff -u -p -r1.9 distinfo --- distinfo26 Dec 2016 09:59:16 - 1.9 +++ distinfo21 Mar 2017 11:26:36 - @@ -1,2 +1,2 @@ -SHA256 (syncthing-source-v0.14.16.tar.gz) = 6eO8psrox12AJCq3HUb6G3bZ1akL+fmAJvVBiFxWKI4= -SIZE (syncthing-source-v0.14.16.tar.gz) = 6902293 +SHA256 (syncthing-source-v0.14.25.tar.gz) = kVKNDMu1qX39LkBez4JfCeHmK7LOYwLV1DxY3rIQUyA= +SIZE (syncthing-source-v0.14.25.tar.gz) = 8038422 Index: patches/patch-build_go === RCS file: /home/edd/cvsync/ports/net/syncthing/patches/patch-build_go,v retrieving revision 1.3 diff -u -p -r1.3 patch-build_go --- patches/patch-build_go 30 Aug 2016 21:05:50 - 1.3 +++ patches/patch-build_go 4 Mar 2017 17:42:17 - @@ -1,7 +1,7 @@ $OpenBSD: patch-build_go,v 1.3 2016/08/30 21:05:50 edd Exp $ build.go.orig Tue Aug 23 02:45:26 2016 -+++ build.go Sat Aug 27 18:28:04 2016 -@@ -391,7 +391,7 @@ func install(target target, tags []string) { +--- build.go.orig Wed Mar 1 21:27:02 2017 build.go Wed Mar 1 21:27:02 2017 +@@ -436,7 +436,7 @@ func install(target target, tags []string) { log.Fatal(err) } os.Setenv("GOBIN", filepath.Join(cwd, "bin")) @@ -10,7 +10,7 @@ $OpenBSD: patch-build_go,v 1.3 2016/08/3 if len(tags) > 0 { args = append(args, "-tags", strings.Join(tags, " ")) } -@@ -411,7 +411,7 @@ func build(target target, tags []string) { +@@ -456,7 +456,7 @@ func build(target target, tags []string) { tags = append(target.tags, tags...) rmr(target.binaryName) -- Best Regards Edd Barrett http://www.theunixzoo.co.uk
Re: Update to gitea-1.1.0
On 03/17, Daniel Jakots wrote: > > Here's a diff to update gitea to 1.1.0. ChangeLog is quite long and it > says: > Hi Daniel, Works fine for me as well, thank you! You forgot to allocate the user/group id for _gitea user in $PORTSDIR/infrastructure/db/user.list, and 787 is already allocated: 787 _dnsdist_dnsdistnet/dnsdist Please consider adding _gitea entry in user.list and change the gitea userid in PLIST so there's no clash. Do you use it with ssh? If not, do you consider the default configuration should allow ssh connections? I found the defaults quite misleading with system's default sshd, gitea's SSH port set to , _gitea home set to /nonexistent and shell set to /sbin/nologin. ssh won't work this way. Maybe we'd better use OpenBSD's default port 22 and @newgroup _gitea:788 @newuser _gitea:788:_gitea:daemon:Gitea Account:/var/gitea:/usr/local/bin/bash ? -- With best regards, Pavel Korovin
Re: phpmyadmin config and readme
On 03/21/17 12:11, Giovanni Bechis wrote: > On 03/21/17 12:02, Paul Irofti wrote: >> On Tue, Mar 21, 2017 at 07:56:55AM +, Stuart Henderson wrote: >>> On 2017/03/20 15:03, Paul Irofti wrote: phpMyAdmin has been installed into ${INSTDIR}. If using Apache, you can make this accessible to clients by enabling -the configuration file: +mod_rewrite in ${SYSCONFDIR}/apache2/httpd2.conf, the mysql php module, +and the configuration file: # ln -s ../modules.sample/phpmyadmin.conf /var/www/conf/modules/ +# cp /etc/php-5.6.sample/mysqli.ini /etc/php-5.6/ >>> >>> Please symlink rather than cp, and use variables so this doesn't >>> get stale after a PHP update: >>> >>> # ln -s ../php-${MODPHP_VERSION}.sample/mysqli.ini >>> ${SYSCONFDIR}/php-${MODPHP_VERSION}/ >> >> Good point on the variables! >> >> I actually wanted to discus ln -s vs. cp. Isn't this a bad idea in >> general? For the php bit I agree that it is useless to copy, but for the >> phpmyadmin.conf I think it's actually required to cp instead of link: if >> I customize the config I also modify the sample file which will get >> overwritten by an update or pkg_add will scream about a bad checksum, >> right? >> > exactly, config.inc.php should not be a symlink. > Giovanni > s/config\.inc\.php/phpmyadmin\.conf/
Re: phpmyadmin config and readme
On 03/21/17 12:02, Paul Irofti wrote: > On Tue, Mar 21, 2017 at 07:56:55AM +, Stuart Henderson wrote: >> On 2017/03/20 15:03, Paul Irofti wrote: >>> phpMyAdmin has been installed into ${INSTDIR}. >>> >>> If using Apache, you can make this accessible to clients by enabling >>> -the configuration file: >>> +mod_rewrite in ${SYSCONFDIR}/apache2/httpd2.conf, the mysql php module, >>> +and the configuration file: >>> >>> # ln -s ../modules.sample/phpmyadmin.conf /var/www/conf/modules/ >>> +# cp /etc/php-5.6.sample/mysqli.ini /etc/php-5.6/ >> >> Please symlink rather than cp, and use variables so this doesn't >> get stale after a PHP update: >> >> # ln -s ../php-${MODPHP_VERSION}.sample/mysqli.ini >> ${SYSCONFDIR}/php-${MODPHP_VERSION}/ > > Good point on the variables! > > I actually wanted to discus ln -s vs. cp. Isn't this a bad idea in > general? For the php bit I agree that it is useless to copy, but for the > phpmyadmin.conf I think it's actually required to cp instead of link: if > I customize the config I also modify the sample file which will get > overwritten by an update or pkg_add will scream about a bad checksum, > right? > exactly, config.inc.php should not be a symlink. Giovanni
Re: phpmyadmin config and readme
On Tue, Mar 21, 2017 at 07:56:55AM +, Stuart Henderson wrote: > On 2017/03/20 15:03, Paul Irofti wrote: > > phpMyAdmin has been installed into ${INSTDIR}. > > > > If using Apache, you can make this accessible to clients by enabling > > -the configuration file: > > +mod_rewrite in ${SYSCONFDIR}/apache2/httpd2.conf, the mysql php module, > > +and the configuration file: > > > > # ln -s ../modules.sample/phpmyadmin.conf /var/www/conf/modules/ > > +# cp /etc/php-5.6.sample/mysqli.ini /etc/php-5.6/ > > Please symlink rather than cp, and use variables so this doesn't > get stale after a PHP update: > > # ln -s ../php-${MODPHP_VERSION}.sample/mysqli.ini > ${SYSCONFDIR}/php-${MODPHP_VERSION}/ Good point on the variables! I actually wanted to discus ln -s vs. cp. Isn't this a bad idea in general? For the php bit I agree that it is useless to copy, but for the phpmyadmin.conf I think it's actually required to cp instead of link: if I customize the config I also modify the sample file which will get overwritten by an update or pkg_add will scream about a bad checksum, right?
Re: npm in lang/node is outdated
Facepalm! For some reason I thought to update this for 6.1 not 6.0 Stable. It's working now. Regards, -- Oriol Demaria 0x1FEF5D72 20 de marzo del 2017 18:21, "Juan Francisco Cantero Hurtado" escribió: > On Mon, Mar 20, 2017 at 05:18:24PM +, Oriol Demaria wrote: > >> In stable we have 4.4.5 and it segfaults for me. Also I can't compile it >> from ports. Anyone else >> with this problem? > > Did you mount the partitions with the option wxallowed? > >> Regards, >> -- >> Oriol Demaria >> 0x1FEF5D72 >> 10 de febrero del 2017 15:38, "Aaron Bieber" escribió: >>> On Thu, Feb 09, 2017 at 06:09:47PM +0100, Leo Unglaub wrote: >>> Hey friends, i noticed that npm (gets shiped with lang/node) is very outdated. Currently version 3.10.10 is shipped, but 4.1.2 should be included. >>> >>> It's not outdated. That's the version that ships with this version of >>> node. Upstream development keeps happening on npm, so there are new >>> versions. >>> >>> https://github.com/nodejs/node/tree/f098f8295c407e5a5126b501ed5f1aa80bd86106/deps >>> >>> ^ That is the commit from the node v6.9.4 tag. >>> I looked at the Makefile but I don't see a reason for node containing this old version of npm. Is that on purpose? Or is this a bug? >>> >>> You can install it yourself, currently 4.3.0 is the latest version. >>> Thanks and greetings Leo >>> >>> -- >>> PGP: 0x1F81112D62A9ADCE / 3586 3350 BFEA C101 DB1A 4AF0 1F81 112D 62A9 ADCE > > -- > Juan Francisco Cantero Hurtado http://juanfra.info
Re: [UPDATE] emulators/fuse to 1.3.4
On Tue, Mar 21, 2017 at 10:25:23AM +0100, Frederic Cambus wrote: > I'm not seeing this error message though. Does fuse 1.3.3 also displays > the same message, or did it appear with the 1.3.4 update? 1.3.3 says the same. It's likely that one of the many uhidev devices I have plugged in thinks it is a joystick in some capacity. -- Best Regards Edd Barrett http://www.theunixzoo.co.uk
Re: [UPDATE] emulators/fuse to 1.3.4
On Mon, Mar 20, 2017 at 10:22:18PM +, Edd Barrett wrote: > > Here is a diff to update fuse to 1.3.4. > > Seems to work, looks good portwise. I get this message as I start fuse: > > fuse: error: sorry, joystick 2 is inadequate! > > Maybe that's expected, I don't have any joysticks plugged in. Did you > try joystick support? I didn't try joystick support as I do not have any. I'm not seeing this error message though. Does fuse 1.3.3 also displays the same message, or did it appear with the 1.3.4 update?
Re: phpmyadmin config and readme
On 2017/03/20 15:03, Paul Irofti wrote: > phpMyAdmin has been installed into ${INSTDIR}. > > If using Apache, you can make this accessible to clients by enabling > -the configuration file: > +mod_rewrite in ${SYSCONFDIR}/apache2/httpd2.conf, the mysql php module, > +and the configuration file: > > # ln -s ../modules.sample/phpmyadmin.conf /var/www/conf/modules/ > +# cp /etc/php-5.6.sample/mysqli.ini /etc/php-5.6/ Please symlink rather than cp, and use variables so this doesn't get stale after a PHP update: # ln -s ../php-${MODPHP_VERSION}.sample/mysqli.ini ${SYSCONFDIR}/php-${MODPHP_VERSION}/
Re: update: lang/rust to 1.16.0
On Mon, Mar 20, 2017 at 10:11:29PM +, Edd Barrett wrote: > On Mon, Mar 20, 2017 at 02:43:58PM +0100, Sebastien Marie wrote: > > On Sun, Mar 19, 2017 at 03:55:26PM +, Edd Barrett wrote: > > > > > > How about a separate port for docs that depends on lang/rust? Would that > > > be more palatable? > > > > I assume it could be possible, but I am unsure if rustbuild will be able to > > generate only the documentation (and not just binary + doc). Testing is > > required :) > > I've been prodding around. There's a 'doc' target in the top-level > Makefile.in, but we don't have a Makefile, presumably because we didn't > use the configure script. > > If I run the configure script, then run `make doc` then it will try to > build again (including failing to get a bootstrap binary from rustup). I > think we'd need to use the configure script from the start, and I > appreciate there is probably a good reason not to do that. configure script and Makefile are the legacy way to build rustc. There are only keep for compatibility reason. The Makefile is done in a way to call rustbuild. The lang/rust port uses rustbuild directly (it is more flexible for the build options). The configuration is done by generating a config.toml (see 'configure' target in the port). And call rustbuild directly for the build (see 'do-build' target in the port). > But wait, earlier you said: > > > the main problem with shipping doc subpackages is now it requires > > rustdoc binary to be present in the bootstrap archive too. > > Are you sure? We are already removing the rustdoc binary from the > bootstrap tarball in the in-tree 0.15.x version, so how does the in-tree > version make docs? It must be using it's own rustdoc? > The bootstrap is generated in the same way that the port itself (and if you read 'bootstrap' target in the port, you will see it reused the fake framework but in different directory). So all files are installed in the bootstrap directory, and it removes all unneeded files. Since the beginning, rustdoc is removed from the required files to be present in bootstrap. In 1.15.1, the rustbuild used the generated rustdoc binary for compiling docs. It isn't the case anymore and it is why it is a problem. But now it is enough. Please show your code or shut up. -- Sebastien Marie