In commit 75d23eec94 we moved to include
commitpkg arguments as the first line of the svn commit message, but we
simply dumped the result after the version number without separating the
two, increasing the cognitive burden of parsing the rationale. Since the
whole point of the change was to make it easier to see what happened
when using git log --oneline (reducing the cognitive burden of parsing
'pretty' output with author/date info), it makes sense to also delineate
the reason correctly.
Signed-off-by: Eli Schwartz <eschwartz@archlinux.org>
Lets prefer the explicit variant of gpg --verify by providing both, the
signature and the data file as parameters.
For the unlikely case there is a matching signature file already present
that was created outside of the toolchain and has an embedded signature
with data, we at least could detect it early with this check.
Signed-off-by: Levente Polyak <anthraxx@archlinux.org>
Several cases showed that we release packages that were built with
different PKGBUILDs than the one commited to the source tree. This is
bad for obvious reasons plus sploils reproducible builds.
We, under no circumstances, want to allow using commitpkg to publish and
release a packages whose PKGBUILD doesn't match the one to be commited.
Signed-off-by: Levente Polyak <anthraxx@archlinux.org>
The unknown packager check didn't worked so far as the wrongly ordered
call to find_cached_package lead to the enclosing block never being
executed.
Signed-off-by: Levente Polyak <anthraxx@archlinux.org>
- Use `read -r` instead of other forms of read or looping
- Use arrays instead of strings with whitespaces.
- In one instance, use ${var%%.*} instead of $(echo $var|cut -f. -d1)
This involves extending the signature of lib/common.sh's `stat_busy()`,
`lock()`, and `slock()`. The `mesg=$1; shift` in stat_busy even suggests
that this is what was originally intended from it.
In cases where there is no license specified, the file is tagged as
"License: Unspecified". Obviously, that is not ideal, but it
highlights the fact, and I hope that it encourages whoever has the
authority to specify the license to do so.
On that note, to anyone who may have the authority to specify the
license of files in devtools: the current licence of many files is
GPLv2 with no option for later versions; I impore you to re-license
them to have the "or any later version" option.
This function (currently) searches through $PWD and $PKGDEST looking
for a tarball matching the requested package name, architecture, and
pkgver. If found, it writes the full path to the located package to
stdout and returns 0, else 1. If more than 1 match is found, it's
treated as an error and the user will need to figure out what to do.
Use this in checkpkg and commitpkg, which previously implemented their
own less complete logic, to locate the build artifacts they rely on.
Signed-off-by: Dave Reisner <dreisner@archlinux.org>
Signed-off-by: Pierre Schmitz <pierre@archlinux.de>
Instead of dying at the first sight of an unversioned file, this lets
commitpkg dump all known unversioned files at once.
Signed-off-by: Dave Reisner <dreisner@archlinux.org>
Signed-off-by: Pierre Schmitz <pierre@archlinux.de>
Now that die() properly forwards arguments to error(), we can expect
that the first arg is a format string and not the entirety of the
output.
Signed-off-by: Dave Reisner <dreisner@archlinux.org>
Signed-off-by: Pierre Schmitz <pierre@archlinux.de>
The usage messages now begins with a "Usage:", i.e. capitalized and with a colon. Fixes FS#26956.
Signed-off-by: Eric Bélanger <snowmaniscool@gmail.com>
Signed-off-by: Pierre Schmitz <pierre@archlinux.de>
The "@" sign in file names in SVN marks the beginning of a pegged
version number -- from the Subversion book:
Peg revisions are specified to the Subversion command-line client
using at syntax, so called because the syntax involves appending an
“at sign” (@) and the peg revision to the end of the path with which
the revision is associated.
The trivial workaround is to always append an at sign to the end of the
path in the version control checks.
Before:
$ community-stagingpkg 'Add systemd units.'
==> ERROR: exim-submission@.service is not under version control
$ svn status -v | grep 'exim-submission@.service'
A - ? ? exim-submission@.service
After:
$ community-stagingpkg 'Add systemd units.'
==> Committing changes to trunk...done
==> Signing package exim-4.80-2-x86_64.pkg.tar.xz...
[...]
Signed-off-by: Lukas Fleischer <archlinux@cryptocrack.de>
Signed-off-by: Pierre Schmitz <pierre@archlinux.de>
We do not allow packages to be uploaded without signatures so force
all unsigned packages to be signed. This has the bonus of not
breaking makepkg signing support by requiring you use an internal
makepkg variable.
Signed-off-by: Allan McRae <allan@archlinux.org>
Signed-off-by: Pierre Schmitz <pierre@archlinux.de>
This fixes a problem where rsync won't work if the pkgver contains a
colon (epoch). In this case rsync assumes that the colon is a
remote:path separator and having src and dest both being remote
arguments is not supported.
Signed-off-by: Florian Pritz <bluewind@xinu.at>
Signed-off-by: Pierre Schmitz <pierre@archlinux.de>
In the case of a .pkg.tar.xz and a .pkg.tar.gz existing in the same
directory, all commitpkg would say is:
==> WARNING: Could not find . Skipping x86_64
Upon digging into the logic, we did a few things poorly, mostly in
getpkgfile:
- getpkgfile tried to die in a subshell (within the command substituion
assignment to 'pkgfile'). This will never work.
- We assumed that proper glob expansion happened when we received
exactly 1 arg. This isn't necessarily true without nullglob in effect.
- We dumped the real error (spewed by getpkgfile) to /dev/null.
- We checked for the package twice in both $PWD and $DESTDIR/.
- We checked for file existance multiple times.
Address this by:
- not hiding errors. revamp the wording a little bit to make it more
obvious why we failed, particularly in the case of a glob expanding to
more than 1 file. Logic here is simplified to pointing out the failure
cases of 0 and >1.
- setting nullglob so the number of arguments passed into getpkgfile is
meaningful from a 'did it decisively resolve' point of view.
- not trying to exit the entire script from a subshell. Just return a
value (and use it).
- avoiding the package file existance check afterwards. this is a
freebie from getpkgfile when the glob passed fails to expand.
Signed-off-by: Dave Reisner <dreisner@archlinux.org>
Signed-off-by: Pierre Schmitz <pierre@archlinux.de>
We already fixed a couple of these in previous patches - this one should
replace all remaining uses of single brackets ("[") by double brackets.
Also, use arithmetic evaluation instead of conditional expressions where
appropriate and make use of "-z" and "-n" instead of comparing variables
to empty strings.
Signed-off-by: Lukas Fleischer <archlinux@cryptocrack.de>
Move the message template before the if block. We moved this to the else
branch in commit aaa68e49e8 which lead to
"msgtemplate" being unset if one specifies a commit message on the
command line, thus stripping the "upgpkg:" part.
Signed-off-by: Lukas Fleischer <archlinux@cryptocrack.de>
Signed-off-by: Pierre Schmitz <pierre@archlinux.de>
Some of the output/error messages were capitalized, some were
not. This patch capitalize everything for consistency sake. Other
minor changes were done to the messages like removing the superfluous
"error:" from die messages and adding a final period to messages that
were complete sentences as appropriate.
Signed-off-by: Eric Bélanger <snowmaniscool@gmail.com>
Signed-off-by: Pierre Schmitz <pierre@archlinux.de>
makepkg(8) currently uses a smarter method to extract all changelog and
install files from a PKGBUILD. Sync commitpkg to use the same code (with
small modifications). This also adds support for changelog/install files
that contain a whitespace.
Signed-off-by: Lukas Fleischer <archlinux@cryptocrack.de>
Signed-off-by: Pierre Schmitz <pierre@archlinux.de>
This build system overhaul allows for adding (define-style) macros to
our scripts. All source files are now suffixed with ".in" to clarify
that they might contain unprocessed defines. The Makefile provides a new
rule to preprocess source files and generate proper output scripts.
Also, add a "@pkgdatadir@" define (as used in GNU Autotools) and use it
instead of hardcoded paths to "/usr/share/devtools" everywhere. We
missed this when adding PREFIX support to the build system in commit
35fc83ce7d.
Signed-off-by: Lukas Fleischer <archlinux@cryptocrack.de>
Eli Schwartz
Daniel M. Capella
Levente Polyak
Bartłomiej Piotrowski
Luke Shumaker
Sven-Hendrik Haase
Johannes Löthberg
Dave Reisner
Pierre Schmitz
Eric Bélanger
Lukas Fleischer
Allan McRae
Florian Pritz
Dave Reisner