Excerpts from Dagobert Michelsen's message of Thu Sep 18 17:33:12 -0400 2008:
> The packages can be downloaded from
><http://mirror.opencsw.org/testing/>
And I just dropped in ruby* 1.8.7p72.
-Ben
--
Ben Walton
Systems Programmer - CHASS
University of Toronto
C:416.407.5610
reat so far. Nice work! :)
-Ben
--
Ben Walton
Systems Programmer - CHASS
University of Toronto
C:416.407.5610 | W:416.978.4302
GPG Key Id: 8E89F6D2; Key Server: pgp.mit.edu
Contact me to arrange for a CAcert assurance meeting.
signature.asc
Des
ild fully.
It's just taken far longer than I anticipated...
Thanks
-Ben
--
Ben Walton
Systems Programmer - CHASS
University of Toronto
C:416.407.5610 | W:416.978.4302
GPG Key Id: 8E89F6D2; Key Server: pgp.mit.edu
Contact me to arrange for a CAcert assurance m
not futz with it over the holidays (but I will if I need to)...
Thanks
-Ben
--
Ben Walton
Systems Programmer - CHASS
University of Toronto
C:416.407.5610 | W:416.978.4302
GPG Key Id: 8E89F6D2; Key Server: pgp.mit.edu
Contact me to arrange for a CAcert assurance meeting.
pec files in the list
was breaking my PKGFILES_ regex globs below (eg: picking their files
in the wrong order). I later discovered that I simply wasn't
squishing the old prototype files...this likely explains why I didn't
notice this breakage too! :)
As always, thanks!
-Ben
--
Ben
s as to how this
would interact in that area are welcome.
Thanks
-Ben
--
Ben Walton
Systems Programmer - CHASS
University of Toronto
C:416.407.5610 | W:416.978.4302
GPG Key Id: 8E89F6D2; Key Server: pgp.mit.edu
Contact me to arrange for a CAcert assurance meeting.
__
e-foo to do that. If the file were still stored
external to the Makefile (not included at all), there might be a way
to pull it in [$(shell cat ...) or something], but I don't see a way
to then have it leave shell variables alone while evaluating the $()
Make variables. Suggestions welcome.
. That may be, in part,
due to the directory structure that rhel uses (where version numbers
are frequently used in paths), but regardless, there are cases where
this can be a handy thing.
Thanks
-Ben
--
Ben Walton
Systems Programmer - CHASS
University of T
s are available, testing can be done without committing every
little tweak
If nobody sees any obvious problems with this patch, I'll check it in.
Thanks
-Ben
--
Ben Walton
Systems Programmer - CHASS
University of Toronto
C:416.407.5610 | W:416.978.4302
GPG Key Id: 8E89F6D2; Key Server: pgp.m
tter to introduce
> a "releasepkg" target which makes sure eveything is checked in,
> builds a package and makes a copy of trunk/ to tags/.
Let me look at this tonight.
Thanks for the feedback.
-Ben
--
Ben Walton
Systems Programmer - CHASS
University of Toronto
C:416.407.5610 |
that checkpkg wouldn't like, which I felt was good.
The other points Dago raised (directory based packages and a release
target) will be addressed separately.
Feedback welcome.
Thanks
-Ben
--
Ben Walton
Systems Programmer - CHASS
University of Toronto
C:416.407.5610 | W:416.978.4302
GPG K
spec, then this modification won't be used. I'm
ok with that personally. What do you think?
[I added the UNCOMMITED at the end of the filename as replacing
REV=... was triggering a checkpkg rule.]
Thanks
-Ben
--
Ben Walton
Systems Programmer - CHASS
University of Toronto
C:416.407.5610 | W
Hi Dago,
Here's one more stab at this. It addresses the points your raised
with my last attempt.
Thoughts?
Thanks
-Ben
--
Ben Walton
Systems Programmer - CHASS
University of Toronto
C:416.407.5610 | W:416.978.4302
GPG Key Id: 8E89F6D2; Key Server: pgp.mit.edu
Contact me to arrange
gging the
package file name was appropriate. If someone builds packages with
GAR, but outside of the svn tree, would we release a package with the
NOTVERSIONED or NOSVN tags? If so, I'd say that CSW is still the best
choice. I'm not fussy either way though.
Thanks for putting up with my p
a
maintainer account and commit the source in that case, but since we
don't even force the use of GAR/svn within the existing maintainer
base, this may seem heavy handed.]
Thanks
-Ben
--
Ben Walton
Systems Programmer - CHASS
University of Toronto
C:416.407.5610 | W:416.978.4302
GPG Ke
rce it on long-time maintainers who already have
> potentially complex build processed on their own. For new
> maintainers GAR is, umh, strongly recommended. Releasable
> packages must be build on the buildfarm anyway.
Ok. This works for me too.
Thanks
-Ben
--
Ben Walton
Systems Prog
e; or as close to it,
> as possible. This would work better than just a commit message.
Agreed. It's the kind of thing that just begs to be deleted. :)
Thanks
-Ben
--
Ben Walton
Systems Programmer - CHASS
University of Toronto
C:416.407.5610 | W:416.978.4302
___
ks as though code review is starting to gain more
traction here... :)
Thanks
-Ben
--
Ben Walton
Systems Programmer - CHASS
University of Toronto
C:416.407.5610 | W:416.978.4302
___
devel mailing list
devel@lists.opencsw.org
https://lists.opencsw.org/mailman/listinfo/devel
Hi Guys,
Thoughts on changing:
Go to Ben Walton QA page
to
Go to the QA page for Ben Walton
The attached patch should implement this change if you like it. It's
consistent with the wording for jumping to the bug tracker info and is
easier than handling the pluralization with 's
Excerpts from skayser's message of Fri Mar 11 13:11:25 -0500 2011:
> Log Message:
> ---
> mgar: add bash autocompletion
Now you're talkin'! :) Nice feature.
Thanks
-Ben
--
Ben Walton
Systems Programmer - CHASS
University of Toronto
C:416.4
Excerpts from wahwah's message of Sun Mar 13 04:18:38 -0400 2011:
> checkpkg: An idea for ugfiles check
Looks like a reasonable check to me.
-Ben
--
Ben Walton
Systems Programmer - CHASS
University of Toronto
C:416.407.5610 | W:416.978.4302
_
Excerpts from Maciej Bliziński's message of Wed Mar 16 04:33:22 -0400 2011:
> This is caused by incorrect pkgmap entries in libgmp. I've added a
> workaround to checkpkg in r13833.
>
> http://sourceforge.net/apps/trac/gar/changeset/13833
Cool! Thanks.
-Ben
--
Ben Wal
Excerpts from phipsy's message of Fri Mar 18 11:05:31 -0400 2011:
Hi Mark,
Mind holding off your release of this until the new CSWlibruby18-1
hits the streets?
Thanks
-Ben
--
Ben Walton
Systems Programmer - CHASS
University of Toronto
C:416.407.5610 | W:416.978
e-isa-sparcv8] Error 2
> current9s%
> current9s%
This is important as WORKSRC is meant to be the directory with the
configuration stuff
in there. I don't have a smart idea on how to fix this. Maybe GIT patching
should start
at WORKDIR/PATCHDIR which defaults to DISTNAME and which is
comparator from the
NotificationFormatter too, but I didn't look too deeply at that.
Maciej, is this roughly accurate?
Thanks
-Ben
[1]
http://sourceforge.net/apps/trac/gar/browser/csw/mgar/gar/v2/lib/python/catalog_notifier.py#L251
--
Ben Walton
Systems Programmer - CHASS
University of Toronto
C:416.4
Excerpts from Dagobert Michelsen's message of Mon Mar 21 14:56:34 -0400 2011:
> WORKSRC ?= $(WORKDIR)/$(DISTNAME)
> PATCHDIR ?= $(WORKDIR)/$(DISTNAME)
Yes, making them independent is wise.
> IMHO the patch can be applied.
Done.
Thanks
-Ben
--
Ben Walton
Systems Prog
Y_/g? If this is being changed, how about
OBSOLETED_BY_?
If Dago is ok with it, it's fine for commit. Don't forget to update
the docs after the change and let maintainers@ know too.
Thanks
-Ben
--
Ben Walton
Systems Programmer - CHASS
Univers
Excerpts from Philip Brown's message of Sat Mar 26 13:28:56 -0400 2011:
> MULTI_WORD_CSWxxx =
Generally it's multi-word. RUNTIME_DEP_PKGS_CSWfoo, SPKG_DESC_...
Thanks
-Ben
--
Ben Walton
Systems Programmer - CHASS
University of Toronto
C:416.407.5610 | W
y the existence of the items listed, ensure the package delivers
nothing except a license file, has a catalog name ending in _stub if
the list of packages obsoleting the current package isn't empty, etc.
Sound ok to you?
Thanks
-Ben
--
Ben Walton
Systems Programmer - CHASS
University of
Excerpts from Ben Walton's message of Sat Mar 26 21:05:26 -0400 2011:
> Sound ok to you?
With the correction of s/if no os.path/if not os.path/.
Thanks
-Ben
--
Ben Walton
Systems Programmer - CHASS
University of Toronto
C:416.407.5610 | W:416.
#x27;t empty. There are many other checks
that can focus on this new data. A list that isn't empty but fails
the syntax check would be one such test.
Updated patch attached.
Thanks
-Ben
--
Ben Walton
Systems Programmer - CHASS
University of Toronto
C:416.407.5610 | W:416.978.4302
collect_obsoletion_info.patch
Description: Binary data
___
devel mailing list
devel@lists.opencsw.org
https://lists.opencsw.org/mailman/listinfo/devel
Excerpts from Ben Walton's message of Sun Mar 27 09:19:15 -0400 2011:
> We only care about the first two white space separated fields.
>
> > if 2 != len(fields):
Hmm. This needs to be a <= comparison since there will actually by
many fields when split by whitespace.
Thank
ading,
> and allow commenting inline.
Yes, good point.
Thanks
-Ben
--
Ben Walton
Systems Programmer - CHASS
University of Toronto
C:416.407.5610 | W:416.978.4302
___
devel mailing list
devel@lists.opencsw.org
https://lists.opencsw.org/mailman/listinfo/devel
return { "syntax_ok": obsoleted_syntax_ok, "by": obsoleted_by }
+
def CheckPkgpathExists(self):
if not os.path.isdir(self.directory):
raise PackageError("%s does not exist or is not a directory"
--
Ben Walton
Systems Programmer - CHASS
University of Toronto
C:
"obsoleted_by": obsoleted_by }
+
> Incrementing of the data structure version doesn't need to be done
> in this change, if it only adds this function. It will need to be
> in the change in the code creating pkg_stats in package_stats.py.
Right. And I'll do my be
Excerpts from Maciej Bliziński's message of Sun Mar 27 16:09:06 -0400 2011:
> The return statement could be de-duplicated in the following manner:
Yes, I agree. Still getting my feet under me so I'm not spotting
silly things like that. Change committed.
Thanks
-Ben
--
Ben W
stBinaries(),
"binaries_dump_info": self.GetBinaryDumpInfo(),
"depends": dir_pkg.GetDependencies(),
+"obsolete_info": dir_pkg.GetObsoletedBy(),
"isalist": sharedlib_utils.GetIsalist(arch),
"overrides": override_dicts,
&quo
Excerpts from Maciej Bliziński's message of Sun Mar 27 19:13:04 -0400 2011:
> The "obsolete_info" bit looks like if the information in question were
> obsolete. How about "obsoleteness_info"?
Sounds good. Will push.
Thanks
-Ben
--
Ben Walton
Systems Programmer
isting item adds a new key?
Thanks
-Ben
--
Ben Walton
Systems Programmer - CHASS
University of Toronto
C:416.407.5610 | W:416.978.4302
___
devel mailing list
devel@lists.opencsw.org
https://lists.opencsw.org/mailman/listinfo/devel
should
> bump the value.
In the case of adding a new feature, it's only required to get the db
updated then.
> Also, if you're breaking backward compatibility.
Ok, so dropping or renaming an existing key in the dict or any
sub-key, etc...
Cool. Thanks!
-Ben
--
Ben Wal
Wrb18-$(GEMPKGNAME)-doc ?= $(or
$(SPKG_DESC_CSWrb18-$(GEMPKGNAME)),$(SPKG_DESC)) documentation
# RUNTIME_DEP_PKGS_CSWgem-$(GEMPKGNAME) ?= $(shell gem specification
$(DOWNLOADDIR)/$(GEMFILE) | $(GARBIN)/gemdeps.rb)
[1] http://wiki.opencsw.org/ruby-dublin
--
Ben Walton
Systems Programmer - CHASS
Unive
-deps: GEM_DEPS?=$(addprefix CSWrb18-,$(shell gem
specification $(DOWNLOADDIR)/$(GEMFILE) | $(GARBIN)/gemdeps.rb))
configure-check-gem-deps:
@echo "=== Checking dependencies of GEM $(GEMFILE) ==="
@$(GARBIN)/check_for_deps $(GEM_DEPS)
--
Ben Walton
Systems Programmer - CHAS
Wruby/CSWruby18/ and that
would be valid too. What are your thoughts on this? The only
downside I see is if we ever make CSWruby pull in 1.9 instead of 1.8
by default, people might end up with the wrong dep. That's quite a
bit in the future though.
Thanks
-Ben
--
Ben Walton
Systems Programme
1.9
becomes a reality?
Thanks
-Ben
--
Ben Walton
Systems Programmer - CHASS
University of Toronto
C:416.407.5610 | W:416.978.4302
___
devel mailing list
devel@lists.opencsw.org
https://lists.opencsw.org/mailman/listinfo/devel
oves CSW.
It definitely does do it right by default now. Set the PACKAGE names
as required and the matching CATALOGNAME will be correct too. You
shouldn't need to explicitly set these unless you need to honour old
names. Being explicit doesn't hurt though. :)
Thanks
-Ben
--
Ben Walto
ter that I'd forgotten to include a
single subpackage or something. It was big and nasty, so small
manageable chunks was nicer to wrangle it.
...noisy though it may be. :) Done now.
Thanks
-Ben
--
Ben Walton
Systems Programmer - CHASS
University of
ster. *g*
I'll have to see if I can get a git-svn checkout to handle the recipe
part of the tree...
-Ben
--
Ben Walton
Systems Programmer - CHASS
University of Toronto
C:416.407.5610 | W:416.978.4302
___
devel mailing list
devel@lists.open
Excerpts from aigoshin's message of Mon Apr 18 08:40:20 -0400 2011:
> +CHECKPKG_OVERRIDES_CSWnginx += bad-location-of-file|file=/var/run/nginx
> +
I thought that /var/opt/csw/run/($foo)* was the preferred location for
the pid files?
(Sorry for double post Andy...)
Thanks
-Ben
--
+
This is what I planned to tackle last night but I forgot about a
parent council meeting for my daughter's school. It looks as though
the file read bit still needs to happen though, right? (eg:
~/.csw-upload-pkg storing u/p)
Thanks
-Ben
--
Ben Walton
Systems Programmer - CHASS
Universi
Excerpts from Maciej Bliziński's message of Wed Apr 20 12:24:17 -0400 2011:
> getpass if unavailable. Passwords will be stored in
> /etc/opt/csw/releases/auth/.secrets (or similar of your
> choice, under /etc/opt/csw).
Is that so the backend can leverage the same file?
Than
Excerpts from Maciej Bliziński's message of Wed Apr 20 13:56:03 -0400 2011:
> It cannot. But both can be populated from the same source,
> e.g. using a script.
Ok then wouldn't this data file be better placed in $HOME?
Thanks
-Ben
--
Ben Walton
Systems Programmer - CHASS
Unive
Excerpts from Sebastian Kayser's message of Wed Apr 20 16:16:59 -0400 2011:
Hi Sebastian,
> you shouldn't need to explicitly install README.CSW. As long as it's
> listed in DISTFILES, GAR will merge it automatically. Same goes for
> changelog.CSW.
Yes, I forgot about th
Before asking the user for a password to submit their package, look
for it in /etc/opt/csw/releases/auth/$username. Fall back to getpass
on any exception.
Signed-off-by: Ben Walton
---
gar/v2/lib/python/csw_upload_pkg.py | 10 +-
1 files changed, 9 insertions(+), 1 deletions(-)
diff
Excerpts from Maciej Bliziński's message of Fri Apr 22 02:03:34 -0400 2011:
> 2011/4/22 Ben Walton :
> > + try:
> > + af = open(authfile, 'r')
> > + password = af.readline().rstrip()
>
> Try using the 'with' syntax:
>
> try:
>
Before asking the user for a password to submit their package, look
for it in /etc/opt/csw/releases/auth/$username. Fall back to getpass
on any exception.
Signed-off-by: Ben Walton
---
gar/v2/lib/python/csw_upload_pkg.py | 11 ++-
1 files changed, 10 insertions(+), 1 deletions
Excerpts from Ben Walton's message of Fri Apr 22 08:10:23 -0400 2011:
> +password = getpass.getpass("{0}'s pkg release password>
> ".format(username))
FYI, this bug in python is annoying. :(
http://bugs.python.org/issue11236
Thanks
-Ben
--
Ben Walton
Systems
Before asking the user for a password to submit their package, look
for it in /etc/opt/csw/releases/auth/$username. Fall back to getpass
on if there is any IOError while reading from the file.
Signed-off-by: Ben Walton
---
gar/v2/lib/python/csw_upload_pkg.py | 11 ++-
1 files changed
Excerpts from Maciej Bliziński's message of Fri Apr 22 19:02:36 -0400 2011:
> LGTM
Ok, here's a pull request then since this branch isn't yet public.
The http-auth branch in ~bwalton/opencsw-git contains the commit.
Thanks
-Ben
--
Ben Walton
Systems Programmer - CHASS
Unive
ISA=i386)
#-> _modulate2(,,isa-i386-static-yes,ISA STATIC,ISA=i386
STATIC=yes)
--snip--
Thanks
-Ben
--
Ben Walton
Systems Programmer - CHASS
University of Toronto
C:416.407.5610 | W:416.978.4302
___
devel mailing list
devel@lists.opencs
Excerpts from Dagobert Michelsen's message of Sun Apr 24 15:30:49 -0400 2011:
Hi Dago,
> I see no problem with that, please proceed.
I just commited this change as r14394.
Thanks
-Ben
--
Ben Walton
Systems Programmer - CHASS
University of Toronto
C:416.407.5610 | W:416.
group (revision 14394)
+++ files/CSWcswclassutils.i.cswusergroup (working copy)
@@ -5,8 +5,10 @@
# $Id$
#
# Written by Peter Bonivart
+# Modified by Ben Walton
#
# 2009-02-10 First release
+# 2011-04-29 Add support for auxgroups
#
# Documentation: http://wiki.opencsw.org/cswclassutils-pa
near future. At that time, I'll update the docs to match.
Thanks
-Ben
--
Ben Walton
Systems Programmer - CHASS
University of Toronto
C:416.407.5610 | W:416.978.4302
___
devel mailing list
devel@lists.opencsw.org
https://lists.opencsw.org/mailman/listinfo/devel
atch and store a commit message with it (part of the patch file).
The basics are:
mgar patch # extract source, apply existing patches
cd work/solaris.../build-isa.../graphviz-$version/
edit source files
cd -
mgar makepatch
(add commit message)
(follow instructions for adding it to your recipe)
This seems like something we could correct without an override...what
is the RPATH value in these files?
Thanks
-Ben
--
Ben Walton
Systems Programmer - CHASS
University of Toronto
C:416.407.5610 | W:416.978.4302
___
devel mailing list
devel@lists.
ask for other eyeballs to help.
Thanks
-Ben
--
Ben Walton
Systems Programmer - CHASS
University of Toronto
C:416.407.5610 | W:416.978.4302
___
devel mailing list
devel@lists.opencsw.org
https://lists.opencsw.org/mailman/listinfo/devel
f re.search(gzipman_re, entry["path"]):
+error_mgr.ReportError("gzipped-manpage-in-pkgmap",
entry["path"])
def CheckMultipleDepends(pkg_data, error_mgr, logger, messenger):
new_depends = set()
I tested it against CSWpuppet and I think it works as expecte
up front. I'll also expand the wiki page
detailing the error tags.
> - a .gz manpage outside of share/man (no error?)
Is this even worth checking then?
> Do you need any help with the unit test?
I think I can handle it.
Thanks
-Ben
--
Ben Walton
Systems Programmer - CHASS
Universi
erllocal-pod-in-pkgmap", entry["path"])
+def CheckGzippedManpages(pkg_data, error_mgr, logger, messenger):
+ gzipman_re = re.compile(r'share/man/man.*/.*\.gz$')
+ for entry in pkg_data["pkgmap"]:
+if entry["path"]:
+ if re.search(gzipman_re, entry[
eportError("perllocal-pod-in-pkgmap", entry["path"])
+def CheckGzippedManpages(pkg_data, error_mgr, logger, messenger):
+ gzipman_re = re.compile(r'share/man/man.*/.*\.gz$')
+ for entry in pkg_data["pkgmap"]:
+if entry["path"]:
+
My mail client picked the wrong address on this reply...(and then I
didn't notice).
Thanks
-Ben
--- Begin forwarded message from Ben Walton ---
From: Ben Walton
To: Igor Galić
Cc: devel
Date: Fri, 24 Jun 2011 19:00:08 -0400
Subject: Re: [csw-devel] Making checkpkg work outside the build
ou make use of cswetcshells.
> +SPKG_CLASSES = none cswetcshells
This is done automatically too. (For this CAS and all the rest.)
> +ETCSHELLS = /opt/csw/bin/zsh
This is the only change you need other than dropping the manual
scripts.
Thanks
-Ben
--
Ben Walton
Systems Programmer - CHASS
Uni
I wasn't aware of that function. I'll switch to using it. Is checkpkg sitting
out "recommend" lines with this function?
Thanks
-Ben
Dagobert Michelsen wrote:
Hi Ben,
Am 09.07.2011 um 14:35 schrieb bdwal...@users.sourceforge.net:
> Revision: 15023
> http://gar.svn.sourceforge.net/gar/?rev=150
> +CATALOGNAME_CSWscrn = screen_transitional
> +RUNTIME_DEP_PKGS_CSWscrn = CSWscreen
You can just use:
OBSOLETED_BY_CSWscreen = CSWscrn
This will generate the proper stub package for you, including any
required overrides.
Thanks
-Ben
--
Ben Walton
Systems Programmer - CHASS
University
Excerpts from Maciej Bliziński's message of Wed Jul 20 04:40:21 -0400 2011:
> > when mgar will be available on the buildfarm?
I'm installing it now...
Thanks
-Ben
--
Ben Walton
Systems Programmer - CHASS
University of Toronto
C:416.407.5610
date, why not split these old libs out
into an openssl_legacy package, depended on by openssl. Then, use the
new library splitting for the current version. It would give you a
clean break from having to drag this baggage forward any further...?
Thanks
-Ben
--
Ben Walton
Systems Programm
my purposes, I can discern them from the file system...I
considered that but erred toward hard coding for now.
Thoughts?
Thanks
-Ben
--
Ben Walton
Systems Programmer - CHASS
University of Toronto
C:416.407.5610 | W:416.978.4302
___
devel mailing li
d to catch everything under either /clearsign or /detach-sign.
Now it gets W*X*Y*Z URL's explicitly defined, where W = 2 (sign type)
and X, Y and Z are the catalog, arch and os combinations.
Thanks
-Ben
--
Ben Walton
Systems Programmer - CHASS
University of Toronto
C:416.407.5610 | W:416.978.4302
n then have all of the tools converted
to require the language specific (if that's the route) file in the
defined location.
Sound good?
Thanks
-Ben
--
Ben Walton
Systems Programmer - CHASS
University of Toronto
C:416.407.5610 | W:416.978.4302
___
feels wrong to me. If I want to build two packages with the same
description, I should be able to do so...and then override checkpkg
which enforces the policy.
What do others think about moving the check?
Thanks
-Ben
--
Ben Walton
Systems Programmer - CHASS
University of T
criptions could be matched against all descriptions in the
> catalog, but the database doesn't have the metadata yet.
I would only look at the current set for a check like this.
Thanks
-Ben
--
Ben Walton
Systems Programmer - CHASS
Univers
ist. In opencsw-future it is only a
> symlink to dublin.
Done. As my comment notes, I'm leaning toward making this dynamic
(with limits on discovery) instead of using a preset array.
Thanks
-Ben
--
Ben Walton
Systems Programmer - CHASS
University
Excerpts from Maciej Bliziński's message of Sun Jul 24 17:26:35 -0400 2011:
> > unset DISPLAY
> > +PATH=/opt/csw/gnu:/opt/csw/bin:$PATH
>
> Perhaps we should ignore the original PATH as a security measure?
Done.
Thanks
-Ben
--
Ben Walton
Systems Programmer - CHASS
n't,
maybe we're not calling gem properly?
Thanks
-Ben
--
Ben Walton
Systems Programmer - CHASS
University of Toronto
C:416.407.5610 | W:416.978.4302
___
devel mailing list
devel@lists.opencsw.org
https://lists.opencsw.org/mailman/listinfo/devel
ct_user_table () ...
function update_pkg_in_webdb(p)
delete_pkg_in_webdb(p)
add_pkg_to_webdb(p)
function update_pkg_in_mantis()
update mantis_project_user_table () ...
Thanks
-Ben
--
Ben Walton
Systems Programmer - CHASS
University of Toronto
C:416.407.5610 | W:416.978.4302
_
d by an update operation. As things are keyed on
the catalog name, a change (based on version comparison) will trigger
a refresh of file list, maintainer name, etc. Mantis cares not about
package name so that won't be a problem.
Reasonable?
Thanks
-Ben
--
Ben Walton
Systems Programmer - C
SWxcbproto
Update pydocutils/CSWpydocutils
Update pm_html_parser/CSWpm-html-parser
Update cas_migrateconf/CSWcas-migrateconf
Update phpmyadmin/CSWphpmyadmin
Update pcre/CSWpcre
Update pm_htmlparser/CSWpmhtmlparser
Update libpcre0/CSWlibpcre0
Update unzoo/CSWunzoo
Removing drupal/CSWdrupal
Removing libclam
CSWfoodev in the old catalog, see that it's called
foo_dev and remove foo_dev from the catalog diff, storing it as a
rename.
Does anyone think that case three should _not_ be an error?
Thanks
-Ben
--
Ben Walton
Systems Programmer - CHASS
Univ
harmful to
the users, a catalog name change should be banned unless it's an
explicit action (done with a manual script/tool instead of via
automation).
Agreed?
Thanks
-Ben
--
Ben Walton
Systems Programmer - CHASS
University of Toronto
C:416.407.5610 | W:416.978.4302
___
The plan is that this will halt the assembly line. Some (not sure who) will get
mail.
Thanks
-Ben
"Maciej Bliziński" wrote:
2011/8/30 Ben Walton :
> I've already written tests that catch the first one case as an error,
> but I think the second should be too. Although
Excerpts from igalic's message of Wed Aug 31 17:21:35 -0400 2011:
Hi Igor,
> Update of logrotate to 3.8.1:
> - remove now obsolete patches
> - Fix platforms (Sol10 only)
This doesn't build on solaris 9?
Thanks
-Ben
--
Ben Walton
Systems Programmer - CHASS
University of Tor
Ooops...replied to the @sourceforge.
Thanks
-Ben
--- Begin forwarded message from Ben Walton ---
From: Ben Walton
To: igalic
Cc: devel
Date: Wed, 31 Aug 2011 17:53:42 -0400
Subject: Re: [csw-devel] SF.net SVN: gar:[15474] csw/mgar/pkg/logrotate/trunk
Excerpts from igalic's message of We
_name is SunOS5.8. Maybe that is the best way to go but I'm not
positive.
Also, should we block adding a package to 5.8 without a specific
tool/interface?
Thanks
-Ben
--
Ben Walton
Systems Programmer - CHASS
University of Toronto
C:416.407.5610 | W:41
Excerpts from igalic's message of Tue Sep 06 20:31:12 -0400 2011:
Hi Igor,
> PACKAGES += CSWhwloc-dev
> +PACKAGES += CSWhwlocrt
I think this would likely be better as CSWlibhwloc3...? If it's only
delivering a single library, this is almost certainly the case.
HTH.
-Ben
--
Be
hings up to 2.2.20. They should be
available on the mirrors now. pkgutil -UC.
Thanks
-Ben
--
Ben Walton
Systems Programmer - CHASS
University of Toronto
C:416.407.5610 | W:416.978.4302
___
devel mailing list
devel@lists.opencsw.org
https://lists.
lude them by setting the
> EXTRA_MERGE_EXCLUDE_FILES to for example .*\.pyc
I looked at the python category file and it doesn't set this
vaule...I'd argue that it should.
I think that we should add:
_MERGE_EXCLUDE_CATEGORY += .*pyc
to v2/categories/python/category.mk
Any objection
nts, I guess! :)
I guess subversion can't really declare itself to be in the python
category though. The manual addition of the excludes in the local
recipe is the way to go.
Thanks
-Ben
--
Ben Walton
Systems Programmer - CHASS
University of Toronto
C:416.4
the log run...I just don't want to release updates
for the legacy things going forward as there is no point.
Time was sapped before I could continue on this last night.
Thanks
-Ben
--
Ben Walton
Systems Programmer - CHASS
University of Toronto
C:416.407.5610
Excerpts from Maciej Blizinski's message of Fri Feb 03 07:43:14 -0500 2012:
> Targets like 'spotless' can take time on large builds, let's
> background them and return the prompt.
+1
Thanks
-Ben
--
Ben Walton
Systems Programmer - CHASS
University of Toronto
C:416
h $(COOKIEDIR))/post-patch-gitsnap; ) \
fi; \
else \
echo "No extracted sources so we can't create patches..."; \
--
Ben Walton
Systems Programmer - CHASS
University of Toronto
C:416.407.5610 | W:416.978.4302
___
een pushed.
Thanks
-Ben
--
Ben Walton
Systems Programmer - CHASS
University of Toronto
C:416.407.5610 | W:416.978.4302
___
devel mailing list
devel@lists.opencsw.org
https://lists.opencsw.org/mailman/listinfo/devel
1 - 100 of 122 matches
Mail list logo
