Fixes "arch" and "checkdepends" never having been unset, fixes b2sums
(but not ${!b2sums_@}) being recently left out.
The "build" function used to be unset as well, explicitly unset it as a
function and do the same for other official functions as well.
Signed-off-by: Eli Schwartz <eschwartz@archlinux.org>
Signed-off-by: Allan McRae <allan@archlinux.org>
In commit b5191ea140 we moved to using
shell globbing to print package files for a couple of reasons including
reproducible packaging of .METADATA files.
Unfortunately, this only works reliably when the glob pattern does not
resolve to a symlinked directory due to a change in the bash 5.0
release. Note that the previous, desired behavior was rather to merely
refuse to recurse into symlinked directories, but due to an unrelated
issue, the symlink handling for globstar was reworked in a way that had
this side effect.
See https://lists.gnu.org/archive/html/bug-bash/2019-04/msg00015.html
for discussion; this may be fixed at some point, but bash 5.0 is broken
either way.
The appropriate way of handling this seems to be to use **/* to match
instead; this produces the same results on both bash 4 and bash 5, as
the ** matches any leading directory component (or none), and the *
matches any file, directory, or symlink to either one.
Fixes FS#62278
Signed-off-by: Eli Schwartz <eschwartz@archlinux.org>
Signed-off-by: Allan McRae <allan@archlinux.org>
The "tip" ref actually signifies the most recently updated branch. hg
does not support a default branch named anything other than "default",
except by creating a "@" bookmark. The correct way to explicitly update
to the default clone ref, is therefore to use one of these, rather than
"tip".
Fixes FS#62092
Signed-off-by: Eli Schwartz <eschwartz@archlinux.org>
Signed-off-by: Allan McRae <allan@archlinux.org>
One of the callers was changed to use known_hash_algos, one was not.
Signed-off-by: Eli Schwartz <eschwartz@archlinux.org>
Signed-off-by: Allan McRae <allan@archlinux.org>
The current completions don't properly handle redirection operators, and
attempt to complete command completions rather than completing filenames
to redirect to.
bash-completion provides both _get_comp_words_by_ref and a higher-level
wrapper _init_completion, but the latter provides handling of redirection
operators, so switch to using that.
Signed-off-by: Eli Schwartz <eschwartz@archlinux.org>
Signed-off-by: Allan McRae <allan@archlinux.org>
In commit 1825bd6716 this was split out
from makepkg, but the warning was not properly migrated; $ext did not
ever exist.
As a result, no matter what you did, the only possible warning was:
==> WARNING: '' is not a valid archive extension.
Fix to filter based on the presence of .tar in the argument, and
building the $ext variable for all checking and messaging purposes
within the function.
Signed-off-by: Eli Schwartz <eschwartz@archlinux.org>
Signed-off-by: Allan McRae <allan@archlinux.org>
In order to cache sources offline, makepkg creates *two* copies of every
git repo. This is a useful tradeoff for network time, but comes at the
cost of increased disk space.
Normally, git can smooth this over automagically. Whenever possible, git
objects are hardlinked to save space, but this does not work when
SRCDEST and BUILDDIR are on separate filesystems.
When the repo in question is both very large (linux.git for example is
2.2 GB) and crosses filesystem boundaries, this results in a lot of
extra disk space being used; the most likely scenario is where BUILDDIR
is a tmpfs for bonus ouch.
git(1) has a builtin feature which serves this case handily: the
--shared flag will create the info/alternates file instructing git to
not copy or hardlink or create objects/packs at all, but merely look for
them in an external location (that being the source of the clone).
The downside of using shared clones, is that if you modify and drop
commits from the original repo, or simply delete the whole repo
altogether, you break the copy. But we don't care about that here,
because
1) the BUILDDIR copy is meant to be a temporary copy strictly derived
via PKGBUILD syntax from the SRCDEST, and must be able to be
recreated at any time,
2) if the SRCDEST disappears, makepkg will redownload it, thus restoring
the objects needed by the BUILDDIR clone,
3) if the user does non-default things like hacking on the BUILDDIR copy
then deleting and re-cloning the SRCDEST may result in momentary
breakage, but ultimately should be fine -- the unique objects they
created will be stored in the BUILDDIR copy.
While it's theoretically possible that upstream will force-push to
overwrite the base tree from which makepkg is building (which they
should not do), *and* the user deleted their SRCDEST which they should
not do, *and* they saved work in makepkg's working directory which they
should not do either...
... this is an unlikely chain of events for which we should not care.
Using --shared is therefore helpful in immediately useful ways and IMHO
has no actual downsides; we should use it.
An alternative implementation would be to use worktrees. I've rejected
this since it is essentially the same as shared clones, except adding
additional restrictions on the branch namespace, and could potentially
break existing use cases such as manually handling the SRCDEST in order
to share repositories with normal working copies.
Signed-off-by: Eli Schwartz <eschwartz@archlinux.org>
Signed-off-by: Allan McRae <allan@archlinux.org>
Since DUFLAGS and DUPATH are not needed anymore remove them from the
source
Signed-off-by: Santiago Torres <santiago@archlinux.org>
Signed-off-by: Allan McRae <allan@archlinux.org>
MODECMD and OWNERCMD are not used by pacman itself, so we don't need to
check for and replace them now that pacman-optimize is removed.
Signed-off-by: Eli Schwartz <eschwartz@archlinux.org>
Signed-off-by: Allan McRae <allan@archlinux.org>
Makepkg used to use du --apparent-size to compute the size of the
package. Unfortunately, this would result in different sizes depending
on the filesystem used (e.g., btrfs vs ext4), which would affect
reproducible builds. Use a wc-based approach to compute sizes
Signed-off-by: Santiago Torres <santiago@archlinux.org>
Signed-off-by: Allan McRae <allan@archlinux.org>
coreutils 8.26 in December 2016 added this new hashing method which is
compatible with the existing md5sum and sha*sum tool usage, while using
the blake2 hash algorithm.
makepkg uses coreutils to provide source file integrity checks via
${integ}sum binaries and it makes sense to offer this as an additional
option.
Signed-off-by: Eli Schwartz <eschwartz@archlinux.org>
Signed-off-by: Allan McRae <allan@archlinux.org>
We cannot use most of the arrays defined in schema.sh as srcinfo is dependent
on the order, but migrate the hashes for now.
Signed-off-by: Eli Schwartz <eschwartz@archlinux.org>
Signed-off-by: Allan McRae <allan@archlinux.org>
We don't need to check the options twice, since it is the same check
both times. Instead, merge the conditionals.
As far as I can tell, the only reason the checks for:
- PACMAN_OPTS and
- whether to use sudo
were ever separated is due to the historic existence of --asroot, since
the second check included a check for (( ! ASROOT )) until it was
cleaned up in commit 61ba5c961e.
Signed-off-by: Eli Schwartz <eschwartz@archlinux.org>
Signed-off-by: Allan McRae <allan@archlinux.org>
When pacman is run as root to do -S, -U, or -R, it would immediately
abort if pacman is not ready for use. Instead, poll the lockfile and
wait until it becomes available.
Implements FS#28840
Original-patch-by: Georges Dubus <georges.dubus@compiletoi.net>
Signed-off-by: Eli Schwartz <eschwartz@archlinux.org>
Signed-off-by: Allan McRae <allan@archlinux.org>
Variables such as 'pkgdesc_x86_64' are invalid, instead of ignoring them
raise an error.
This also disallows using 'any' as an architecture specific variable
Signed-off-by: morganamilo <morganamilo@gmail.com>
Signed-off-by: Allan McRae <allan@archlinux.org>
makepkg will now error if disallowed variables are set inside of
the package function.
Disallowed variables are variables that do exist, like 'makedepends'
and 'pkgver' but can not be set inside of a package function.
Signed-off-by: morganamilo <morganamilo@gmail.com>
Signed-off-by: Allan McRae <allan@archlinux.org>
Refactor many of the different arrays of pkgbuild variables
into scripts/libmakepkg/util/schema.sh.in.
Signed-off-by: morganamilo <morganamilo@gmail.com>
Signed-off-by: Allan McRae <allan@archlinux.org>
This patch was inspired by FS#32723 which asks makepkg to install makedepends
before depends. The use case is to build a package depending on a virtual
package that is only provided by other packages (e.g. java-runtime in Arch
Linux), but wanting to build against a specific version. Installing makedepends
first (but not at the same time as depends) would allow specifying the version
to build against, instead of pacman resolving to the default version when
installing depends.
It turns out, we can already achieve installing makedepends first by specifying
dependencies only in the package function (and making sure makedepends includes
everything needed). The only issue is that if we use makepkg to install the
built package with the --install flag and along with the --rmdeps flag, we will
try to remove any installed dependencies that are specified in the depends
array in the package function. To counter this, we need to use the --unneeded
flag for the pacman call.
Signed-off-by: Allan McRae <allan@archlinux.org>
lzip is a lossless data compressor designed to replace gzip and bzip2 as
the standard general-purpose compressed format.
- add .lz (lzip) support to libmakepkg/util/compress.sh:compress_as
- add COMPRESSLZ to makepkg.conf.in
- document COMPRESSLZ
- document PKGEXT with `.tar.lz`
Signed-off-by: Chloe Kudryavtsev <toast@toastin.space>
Signed-off-by: Allan McRae <allan@archlinux.org>
Lookup the existence of matching functions for each protocol, and
fallback on the generic file handler. New verification protocols can
then be added via thirdparty libmakepkg drop-ins without requiring
modifications to verify_signature.sh
Signed-off-by: Allan McRae <allan@archlinux.org>
Lookup the existence of matching functions for each protocol, and
fallback on the generic file handler. New source protocols can then be
added via thirdparty libmakepkg drop-ins without requiring modifications
to source.sh
Fixes FS#49076
Signed-off-by: Allan McRae <allan@archlinux.org>
e.g. git+https:// is commonly used for git repositories cloned over
HTTPS, but we assume a proto with a plus in it is actually a protocol
followed by some URI handler. So we might as well simplify the return
value and not have to always add glob matching everywhere when checking
the proto in use.
This is required in order to use the proto directly in function calls,
which will be used in a followup patch.
Signed-off-by: Allan McRae <allan@archlinux.org>
Making the undescore be translated is probably not something we need
translators to think about.
Additionally, a number of places which use the same text differ only by
the variable being referenced, so simplifying the string means we can
drop a redundant translation.
Bonus: we save a few bytes here and there. \o/
Signed-off-by: Eli Schwartz <eschwartz@archlinux.org>
Signed-off-by: Allan McRae <allan@archlinux.org>
Error if the arch array contains any and any other values. This also
fixes a bug where the check for `$arch == 'any'` which only evaluated
the first value in the array, meaning the rest of the values would not
be linted.
Signed-off-by: morganamilo <morganamilo@gmail.com>
Signed-off-by: Allan McRae <allan@archlinux.org>
In commit f7efa6a93d we added a new file,
and also wired it up to the build systems, but it got added under the
wrong name in meson.build
Signed-off-by: Eli Schwartz <eschwartz@archlinux.org>
Signed-off-by: Allan McRae <allan@archlinux.org>
This is not really an error with a "user function".
Signed-off-by: Eli Schwartz <eschwartz@archlinux.org>
Signed-off-by: Allan McRae <allan@archlinux.org>
Use mapfile instead of hacking around read -a with the $IFS.
Signed-off-by: Eli Schwartz <eschwartz@archlinux.org>
Signed-off-by: Allan McRae <allan@archlinux.org>
This can only ever be an int, and the specification states that a
malformed timestamp should be considered a fatal error.
https://reproducible-builds.org/specs/source-date-epoch/
Signed-off-by: Eli Schwartz <eschwartz@archlinux.org>
Signed-off-by: Allan McRae <allan@archlinux.org>
paths can contain printf-unsafe chars, and printf -v is not somehow immune to this
Signed-off-by: Eli Schwartz <eschwartz@archlinux.org>
Signed-off-by: Allan McRae <allan@archlinux.org>
If you have a malformed pkgrel, the error message says that it must be a
"decimal". That isn't quite true, as that would mean that `1.1 == 1.10`.
Signed-off-by: Allan McRae <allan@archlinux.org>
Given the depends
depends=('foo>=1.2-1.par2')
and the error message
==> ERROR: pkgver in depends is not allowed to contain colons, forward slashes, hyphens or whitespace.
One would be lead to believe that the problem is that they gave a pkgrel in
depends at all, not that the pkgrel contains letters.
Each of the (check,make,opt)depends, conflicts, and provides linters use a
glob to trim off properly formed epoch an rel from the full version string,
and pass the remainder to check_pkgver(). This does a good job of
accepting/rejecting full versions, but doesn't do a good job of generating
good error messages when rejecting if it's because of the epoch or rel.
1. Factor out check_epoch() and check_pkgrel() from lint_epoch() and
lint_pkgrel(), similarly to check_pkgver().
2. Add a check_fullpkgver() that takes a full [epoch:]ver[-rel] string and
splits it in to epoch/ver/rel, and calls the appropriate check_ function
on each.
3. Use check_fullpkgver() in the {,check,make,opt}depends, conflicts, and
provides linters.
Signed-off-by: Allan McRae <allan@archlinux.org>
Checking the length of the variable to be non-zero before considering it
an error is inconsistent; license=() and depends='' and `declare arch`
should be considered just as wrong.
In fact the current check detects depends='' as non-zero and returns an
error, but happily considers the others to be perfectly okay.
A more reliable check is to simply see if the name has been declared
(whether it is set or not), and then enforce that it's been declared to
the right type.
As an added benefit, avoiding the creation of proxy-evaled variables to
count the number of indexes results in simpler code.
Signed-off-by: Eli Schwartz <eschwartz@archlinux.org>
Signed-off-by: Allan McRae <allan@archlinux.org>
In commit d8ee8d0c99 we made use of
fakeroot absolutely mandatory, and disabled a lot of the code which
checked to see if this now-defunct BUILDENV option was set, before
setting up the environment to use fakeroot. Unfortunately, we missed one
spot.
The check_software routine still checked to see if fakeroot was
enabled, but due to the option being removed, thought that it was in
fact disabled, and as a result this check would never run.
Fix by checking to see if we are trying to build either a package or a
source package, and if so, checking for fakeroot. These are the only two
situations where fakeroot is needed.
Signed-off-by: Eli Schwartz <eschwartz@archlinux.org>
Signed-off-by: Allan McRae <allan@archlinux.org>
There are state variables for everything else, and we use them to do
conditional checks on things, but it's currently a bit difficult to test
whether a package is being built, as it's the default action if *no*
options are specified.
Signed-off-by: Eli Schwartz <eschwartz@archlinux.org>
Signed-off-by: Allan McRae <allan@archlinux.org>
Merge the similar code handling unsplit PKGBUILDs and individual
packages in a split PKGBUILD and make it a new function.
Signed-off-by: Allan McRae <allan@archlinux.org>
Our sed parser for xdelta3 headers will greedily match on ":" which
coincidentally is also the character we use to define a version with an
epoch.
While we are at it, simply use sed for the whole pipeline, rather than
using both grep and sed.
Fixes FS#61195
Signed-off-by: Eli Schwartz <eschwartz@archlinux.org>
Signed-off-by: Allan McRae <allan@archlinux.org>
We have code in order to remove deltas when removing a package, but it
is never run, since we try to remove the wrong file.
This was broken in commit cb0f2bd038 which
modified the internal layout we use to modify the db, changing "tree" to
"db", but did not update all locations where it was used.
This worked swimmingly well as long as only repo-add updates were
handling the backup and restore of the delta file, as the delta file
therefore got backed up to the correct location (db) in the shared
db_remove_entry() function.
But later on in the repo-remove logic, we tried removing a different
file that will never exist (tree).
Fixes FS#53041
Signed-off-by: Eli Schwartz <eschwartz@archlinux.org>
Signed-off-by: Allan McRae <allan@archlinux.org>
directories are created by install_dir within the subdir custom_target
installation targets.
Signed-off-by: Eli Schwartz <eschwartz@archlinux.org>
Signed-off-by: Allan McRae <allan@archlinux.org>
Simply fix a typo: in written -> is written
Signed-off-by: Michael Straube <michael.straube@posteo.de>
Signed-off-by: Allan McRae <allan@archlinux.org>
Elsewhere, we return 1 if a library dropin fails, and when running
functions in a loop, we use `|| ret=1` to preserve scope. This ensures
the return value of the function remains useful in isolation. Do the
same thing here as well.
Drop trivial function which wraps a dropin that also uses $ret, since
it's no longer needed.
Signed-off-by: Eli Schwartz <eschwartz@archlinux.org>
Signed-off-by: Allan McRae <allan@archlinux.org>
Now that repo-add uses libmakepkg, it needs to have $LIBRARY set before
testing it in-tree.
[Allan: fix "make distcheck"]
Signed-off-by: Allan McRae <allan@archlinux.org>
Instead of assuming all scripts are .sh.in and leaving a comment to that
effect, just take the input file directly.
This depends on the first dependency for the target being the source of
the script.
Signed-off-by: Allan McRae <allan@archlinux.org>
All of our scripts depend on the same pattern .sh.in, and since commit
b5d62d2c91, they also all (not just
makepkg itself) depend on libmakepkg.
There's no real reason to include separate targets for them just to
establish dependency rules.
While we are at it, fix a longstanding bug where generated wrapper
scripts did not depend on wrapper.sh.in (which due to moving to .lib,
requires we regenerate the script too), by making the shared target
pattern depend on it. All our generated scripts now require the wrapper,
even repo-add which now uses libmakepkg.
Signed-off-by: Allan McRae <allan@archlinux.org>
repo-remove and repo-elephant don't care whether repo-add.sh.in is
updated... but they do require the repo-add target to be up to date, so
use that instead. As a bonus, use the same rule for both of them.
