Re: CVS: cvs.openbsd.org: ports
On Wed, 19 Aug 2020 08:06:51 -0600 (MDT) Charlene Wendling wrote: > CVSROOT: /cvs > Module name: ports > Changes by: c...@cvs.openbsd.org2020/08/19 08:06:51 > > Log message: > Import www/p5-Catalyst-Action-REST > > This Action handles doing automatic method dispatching for REST > requests. It takes a normal Catalyst action, and changes the > dispatch to append an underscore and method name. First it will try > dispatching to an action with the generated name, and failing that it > will try to dispatch to a regular method. > > OK afresh1@ > > Status: > > Vendor Tag: cwen > Release Tags: cwen_20200819 > > N ports/www/p5-Catalyst-Action-REST/Makefile > N ports/www/p5-Catalyst-Action-REST/distinfo > N ports/www/p5-Catalyst-Action-REST/pkg/DESCR > N ports/www/p5-Catalyst-Action-REST/pkg/PLIST > > No conflicts created by this import > Sorry i forgot to mention this comes from Wen Heping!
Re: CVS: cvs.openbsd.org: ports
On Fri, 24 Apr 2020 10:17:35 -0600 (MDT) Charlene Wendling wrote: > CVSROOT: /cvs > Module name: ports > Changes by: c...@cvs.openbsd.org2020/04/24 10:17:35 > > Modified files: > productivity/novprog: Makefile > > Log message: > novprog: drop 'atomic' from WANTLIB for powerpc, it breaks the build > and is not needed anymore with base-clang on this arch. > > OK bcallah@ (maintainer) > And OK kmos@ as well.
Re: CVS: cvs.openbsd.org: ports
On Tue, 4 Feb 2020 15:54:09 -0700 (MST)
Charlene Wendling wrote:
> CVSROOT: /cvs
> Module name: ports
> Changes by: c...@cvs.openbsd.org2020/02/04 15:54:09
>
> Modified files:
> games/frozen-bubble: Makefile
> Added files:
> games/frozen-bubble/patches: patch-c_stuff_fb_c_stuff_xs
>patch-c_stuff_lib_FBLE_pm
>patch-frozen-bubble
>
> Log message:
> Add various fixes to frozen-bubble:
>
> - allow building with base-clang (from gkoehler@, thanks!)
> - remove BROKEN-{amd64,i386} after gkoehler@'s p5-SDL fix
^ Wrong, actually the BROKEN lines were removed in the p5-SDL
commit/Makefile. But frozen-bubble is unbroken on these archs for
real.
> - replace various use of my() in false conditionals, it's fatal
> with Perl 5.30
> - deinterlace some PNGs to silence some libpng warnings
>
> OK gkoehler@ afresh1@ (who tested on sparc64 and i386)
>
Re: CVS: cvs.openbsd.org: ports
On Sun, 24 Nov 2019 03:53:42 -0700 (MST) Charlene Wendling wrote: > CVSROOT: /cvs > Module name: ports > Changes by: c...@cvs.openbsd.org2019/11/24 03:53:42 > > Modified files: > devel/libvterm/patches: patch-bin_vterm-ctrl_c > Added files: > devel/libvterm/patches: patch-Makefile > > Log message: > > (cvs did not abort on blank commit message as i expected) libvterm: drop `-Wpedantic' so it still builds on base-gcc archs with the default compiler. While here, convert a git-styled patch to what cvs(1) expect, making update-patches(1) happy. OK edd@ (maintainer)
Re: CVS: cvs.openbsd.org: ports
On Fri, 30 Aug 2019 13:54:47 -0600 (MDT) Charlene Wendling wrote: > CVSROOT: /cvs > Module name: ports > Changes by: c...@cvs.openbsd.org2019/08/30 13:54:47 > > Modified files: > databases/p5-Mojo-Pg: Makefile distinfo > > Log message: > p5-Mojo-Pg: update to 4.15 > Changelog: > https://metacpan.org/source/SRI/Mojo-Pg-4.15/Changes > > OK maintainer > Sorry! I forgot to mention it was from Wen Heping.
Re: CVS: cvs.openbsd.org: ports
On Fri, 16 Aug 2019 17:27:09 -0600 (MDT) Charlene Wendling wrote: > CVSROOT: /cvs > Module name: ports > Changes by: c...@cvs.openbsd.org2019/08/16 17:27:09 > > Modified files: > devel/angr/py-z3-solver: Makefile > Added files: > devel/angr/py-z3-solver/patches: > > patch-core_src_util_lp_permutation_matrix_h > > Log message: > py-z3-solver: unbreak with ports-gcc, by fixing an incorrect > conversion. Tested on sparc64 (by kmos@, thanks!) and macppc. > > OK kmos@ kn@ (maintainer) > And OK jasper@ who co-maintains the port.
Re: CVS: cvs.openbsd.org: ports
On Sun, 7 Jul 2019 10:58:46 -0600 (MDT) Charlene Wendling wrote: > CVSROOT: /cvs > Module name: ports > Changes by: c...@cvs.openbsd.org2019/07/07 10:58:46 > > Modified files: > audio/taglib : Makefile > Added files: > audio/taglib/patches: patch-tests_test_synchdata_cpp > > Log message: > taglib: fix narrowing issues on arm and ppc, unbreak the build. > Tested on macppc. I've forgotten to mention that i moved to PERMIT_PACKAGE. > OK jca@ >
Re: CVS: cvs.openbsd.org: ports
On Fri, 17 May 2019 14:23:24 -0600 (MDT) Charlene Wendling wrote: > CVSROOT: /cvs > Module name: ports > Changes by: c...@cvs.openbsd.org2019/05/17 14:23:24 > > Modified files: > devel/p5-Term-Encoding: Makefile distinfo > > Log message: > p5-Term-Encoding: update to 0.03 > Changelog: > https://metacpan.org/release/MIYAGAWA/Term-Encoding-0.03 > *sigh* This one wasn't OK'ed, i've taken the wrong list. Sorry.
Re: CVS: cvs.openbsd.org: ports
On Wed, 15 May 2019 07:34:58 -0600 (MDT) Charlene Wendling wrote: > CVSROOT: /cvs > Module name: ports > Changes by: c...@cvs.openbsd.org2019/05/15 07:34:58 > > Modified files: > www/p5-HTTP-Headers-Fast: Makefile distinfo > > Log message: > p5-HTTP-Headers-Fast: update to 0.22 > Changelog: > https://metacpan.org/source/TOKUHIROM/HTTP-Headers-Fast-0.22/Changes > > tweaks and OK afresh1@ > I've totally forgotten to mention that it's an original submission from Wen Heping (thanks!), sorry.
Re: untangling some bsd.port.mk permissions
On Thu, 9 May 2019 22:35:10 +0200
Marc Espie wrote:
> On Thu, May 09, 2019 at 10:22:48PM +0200, Charlène Wendling wrote:
> > make update-plist
> > =
> >
> > => doas /usr/bin/perl /usr/ports/infrastructure/bin/update-plist
> >
> > In mk/pkgpath.mk there is this:
> >
> > _PERLSCRIPT = /usr/bin/perl ${PORTSDIR}/infrastructure/bin
> >
> > After removing the perl invocation the problem is getting fishier:
> >
> > - /usr/ports/infrastructure/bin/update-plist is executable, fine
> > - /usr/ports/infrastructure/bin/port-getpkgpath-helper is not
> > executable, i chmoded +x it to see how deep it goes
> > - then it calls 'doas /usr/ports/infrastructure/bin/update-plist'
> > - and it croaks:
>
> Those are a problem, because you can't really enforce +x under source
> directories. Removing the _PERLSCRIPT everywhere is a bad idea.
>
> (note that src suffers from the same issue and explicitly prefix with
> /bin/sh every script it's going to run).
>
> The only script that really wants root is update-plist. Entering a
> password for that one looks admissible.
I agree, as you propose it, i think it's already nice enough, this
reduces massively the count of password prompts with PORT_PRIVSEP
and "permit keepenv $USER", and anyway if you set doas as this,
you know you're gonna type your password often.
> > make clean=packages
> > ===
> >
> > - my user account wants to rm:
> > /usr/ports/packages/amd64/cache/portname.tgz
> > and it belongs to _pfetch, so i needed this in bsd.port.mk:
> >
> > @@ -3120,10 +3120,10 @@ _internal-clean:
> > .endif
> > .if ${_clean:Mpackages} || ${_clean:Mpackage} && ${_clean:Msub}
> > ${_PBUILD} rm -f ${_PACKAGE_COOKIES}
> > - rm -f ${_UPDATE_COOKIES} ${_CACHE_PACKAGE_COOKIES}
> > + ${_PFETCH} rm -f ${_UPDATE_COOKIES} $
> > {_CACHE_PACKAGE_COOKIES} .elif ${_clean:Mpackage}
> > ${_PBUILD} rm -f ${_PACKAGE_COOKIES${SUBPACKAGE}}
> > - rm -f ${_UPDATE_COOKIE${SUBPACKAGE}}
> > + ${_PFETCH} rm -f ${_UPDATE_COOKIE${SUBPACKAGE}}
>
>
> Err, this is weird. ${_CACHE_PACKAGE_COOKIES} will indeed belong to
> _pfetch. Not so for ${_UPDATE_COOKIES}
>
It was a wild guess, after the mess i did to get up to this point, i
understood that there was no future for this, and cut to the chase.
Re: untangling some bsd.port.mk permissions
Hi,
On Wed, 8 May 2019 21:00:57 +0200
Marc Espie wrote:
> In a few places, bsd.port.mk does
> doas /usr/bin/env -i ${_TERM_ENV}
> TRUSTED_PKG_PATH=... /usr/sbin/pkg_add
>
> On the one hand, env allows virtually everything to execute;
> on the other hand, swapping things around means keepenv has to be used
> correctly.
>
> Looking closer at the actual usage pattern, the env variables
> concerned are:
>
> - TERM: necessary for correct progressmeter
> - TERMCAP: good for people with bad terminal configuration.
> Definitely not something to trust in doas.conf
> - ftp_proxy/http_proxy: useful in general, but those pkg_add
> invocations are actually local
> - TRUSTED_PKG_PATH: *TOTALLY* necessary. This prevents pkg_add from
> looking in other locations, and replaces a former -Dunsigned which
> did remove signature handling from everywhere and not just the
> correct directory.
>
> Inspired by Charlene's idea of fixing the path to touch, I think we
> want the patch that follows.
As you already know, it works fine for me, thanks!
> Plus: people will have a full list of what's needed to run as root for
> ports work.
I've found more (i'm describing issues as i have met them) to make it
"seamless":
make update-plist
=
=> doas /usr/bin/perl /usr/ports/infrastructure/bin/update-plist
In mk/pkgpath.mk there is this:
_PERLSCRIPT = /usr/bin/perl ${PORTSDIR}/infrastructure/bin
After removing the perl invocation the problem is getting fishier:
- /usr/ports/infrastructure/bin/update-plist is executable, fine
- /usr/ports/infrastructure/bin/port-getpkgpath-helper is not
executable, i chmoded +x it to see how deep it goes
- then it calls 'doas /usr/ports/infrastructure/bin/update-plist'
- and it croaks:
DON'T BUILD PORTS AS ROOT!
(or make sure you pass env variables PORTS_TREE_OWNER and
FAKE_TREE_OWNER thru doas to root)
- So i've changed my doas.conf accordingly, after reading
update-plist, i noticed it requires PORTSDIR to be set as
well, as seen in strip_dependency_directories() iiuc,
or i have this:
update-plist: Prefix required
- /usr/ports/infrastructure/bin/port-resolve-lib-helper needs
also to be chmoded +x
make install/reinstall
==
- /usr/ports/infrastructure/bin/register-plist requires to be
chmoded +x as well
make clean=packages
===
- my user account wants to rm:
/usr/ports/packages/amd64/cache/portname.tgz
and it belongs to _pfetch, so i needed this in bsd.port.mk:
@@ -3120,10 +3120,10 @@ _internal-clean:
.endif
.if ${_clean:Mpackages} || ${_clean:Mpackage} && ${_clean:Msub}
${_PBUILD} rm -f ${_PACKAGE_COOKIES}
- rm -f ${_UPDATE_COOKIES} ${_CACHE_PACKAGE_COOKIES}
+ ${_PFETCH} rm -f ${_UPDATE_COOKIES} ${_CACHE_PACKAGE_COOKIES}
.elif ${_clean:Mpackage}
${_PBUILD} rm -f ${_PACKAGE_COOKIES${SUBPACKAGE}}
- rm -f ${_UPDATE_COOKIE${SUBPACKAGE}}
+ ${_PFETCH} rm -f ${_UPDATE_COOKIE${SUBPACKAGE}}
doas.conf
=
permit keepenv charlene
[...]
permit nopass keepenv charlene as _pbuild
permit nopass keepenv charlene as _pfetch
permit nopass setenv { TRUSTED_PKG_PATH TERM } charlene cmd /usr/bin/touch
permit nopass setenv { TRUSTED_PKG_PATH TERM } charlene cmd /usr/sbin/pkg_delete
permit nopass setenv { TRUSTED_PKG_PATH TERM } charlene cmd /usr/sbin/pkg_add
permit nopass setenv { PORTSDIR PORTS_TREE_OWNER FAKE_TREE_OWNER
TRUSTED_PKG_PATH TERM } \
charlene cmd /usr/ports/infrastructure/bin/update-plist
> Minus: if you don't keepenv TRUSTED_PKG_PATH, things will stop
> working. If you don't keepenv TERM, pkg_add will lose its
> progressmeter.
>
> (that said, pkg_delete already has the same issue and it doesn't look
> like people protest)
>
> okay, objections ?
None of these two in my case, i just wanted to report the issues i met
if one doesn't want to input root password during port works. I may
have missed some stuff, or lacking foresight though.
I had fun anyway!
Charlène.
