In some cases packages are built outside of a directory which contains
pkgname-pkgver, this results in source listing in debug packages having
a conflicting path like `/usr/src/debug/build/` which is not ideal.
This patch ensures we always include the pkgbase to ensure the paths are
unique.
Signed-off-by: Morten Linderud <morten@linderud.pw>
Signed-off-by: Allan McRae <allan@archlinux.org>
We use a multi step process during stripping to ensure permissions do
not get changed. However, if the initial objcopy fails, the subsequent cat
results in a blank file. Abandon early if objcopy fails.
Fixes FS#74486
Signed-off-by: Allan McRae <allan@archlinux.org>
We want to use -flto=auto in Arch Linux to speed up building, but we
can't hardcode it in buildenv/lto.sh because other downstreams might
have clang < 13.0.0 which did not recognize -flto=auto as equivalent
to -flto=full.
Introducing an LTOFLAGS variable to makepkg.conf seems the way to go.
Signed-off-by: Allan McRae <allan@archlinux.org>
This moves us from the fairly ugly AWK parsing line to debugedit which
originally comes out of the rpm project.
The original code has issues parsing anything that was not straight
C/C++ and languages like Rust or Go would return invalid source code
files. debugedit handles all these cases better.
Fixes FS#66755
Fixes FS#66888
Fixes FS#65677
Signed-off-by: Morten Linderud <morten@linderud.pw>
Signed-off-by: Allan McRae <allan@archlinux.org>
>From gcc(1):
-ffile-prefix-map=old=new
[...] Specifying this option is equivalent to specifying all the
individual -f*-prefix-map options. This can be used to make reproducible
builds that are location independent.
Specifically, this additionally enables -fmacro-prefix-map=, which causes
prefix mapping to be applied to expansions of __FILE__ and similar macros.
Without this option, if source files are compiled by passing the
absolute file path to the compiler (as done by e.g. cmake), any
expansions of __FILE__ (e.g. from uses of assert()) will contain
$srcdir.
Signed-off-by: Allan McRae <allan@archlinux.org>
Add linked libraries to a packages dependency list. This is the partner
to automatically generated library provides, and thus depends take the
same format. To help with bootstrapping, library dependencies are only
added if the relevant provide exists.
Signed-off-by: Allan McRae <allan@archlinux.org>
When the option "autodeps" is enabled, makepkg will add provides
entries for libraries found in the directories specified in LIB_DIRS
in makepkg.conf. The entries LIB_DIRS array have the format
"prefix:directory". For example, the entry "lib:usr/lib" will search
$pkgdir/usr/lib for library sonames and add "lib:libfoo.so.1" to the
provides array.
Signed-off-by: Allan McRae <allan@archlinux.org>
GCC automatically detects when it is linking LTO objects, but clang does
not. Add -flto to LDFLAGS to make this work for clang too.
Signed-off-by: Allan McRae <allan@archlinux.org>
Arch Linux has been setting PYTHONHASHSEED=0 to create deterministic
.pyc files. After a thorough review by the Arch Security Team, setting
this variable was determined not to generated vulnerable .pyc files, as
when the loader loads the .pyc file and unmarshalls it, the internal
runtime will just populate the unordered data structures and use a new
runtime hash for them.
Signed-off-by: Allan McRae <allan@archlinux.org>
We usually set this up to default to the build time configured install
location, but a couple of files crept in without this.
Signed-off-by: Eli Schwartz <eschwartz@archlinux.org>
Signed-off-by: Allan McRae <allan@archlinux.org>
Binutils commit 93df3340fd5ad32f784214fc125de71811da72ff enabled readelf
to report "Position-Independent Executable" files. Fix stripping to
account for this change.
Signed-off-by: Allan McRae <allan@archlinux.org>
readelf --debug-dump sometimes reports inscrutable warnings which don't
actually affect our extraction of source filenames. For example:
readelf: Warning: There is a hole [0xd3d - 0xd89] in .debug_loc section.
Now gcc 11 seems to have dramatically increased the number of warnings:
readelf: Warning: Corrupt offset (0x0000008e) in range entry 9
[...]
readelf: Warning: Corrupt offset (0x000010f0) in range entry 250
The resulting debuginfo created by the very same toolchain works fine,
as does the list of source filenames. But the warnings are quite
noisy... send them to /dev/null since they are not actionable in the
context of getting source files
Signed-off-by: Eli Schwartz <eschwartz@archlinux.org>
Signed-off-by: Allan McRae <allan@archlinux.org>
If specified, this will be used no matter what. If not, then we check if
sudo exists and use that, or else fall back on su.
Implements FS#32621
Signed-off-by: Eli Schwartz <eschwartz@archlinux.org>
Signed-off-by: Allan McRae <allan@archlinux.org>
Broken via refactoring in commit aa6fe1160b
but for obvious reasons only one person in the last 9 years has ever
actually tried to do this. Still, it's technically correct to allow it.
Fixes FS#70254
Signed-off-by: Eli Schwartz <eschwartz@archlinux.org>
Signed-off-by: Allan McRae <allan@archlinux.org>
The current gcc build from git master give different output from
readelf:
gcc-10.2.0
$ readelf "hello" --debug-dump | grep hello
<11> DW_AT_name : (indirect string, offset: 0xbfc): hello.cpp
gcc-git
$ readelf "hello" --debug-dump | grep hello
<12> DW_AT_name : (indirect line string, offset: 0x0): hello.cpp
This causes the awk statement extracting the file name to fail as it
relied on the information being in the 8th field. Instead, extract
the information from the final field.
Fixes FS#70168
Signed-off-by: Allan McRae <allan@archlinux.org>
GCC's LTO implementation emits bytecodes into .o files it generates.
These bytecodes are _not_ considered stable from one release of GCC
to the next. There we need to strip the LTO bytecode out of any .o
(and .a) file that gets installed into the package.
Signed-off-by: Allan McRae <allan@archlinux.org>
Add the 'lto' option to enable building with link time optimization
by adding '-flto' to both CFLAGS and CXXFLAGS. The 'lto' option can
be specificed both in the PKGBUILD or by setting the default in
makepkg.conf.
Signed-off-by: Allan McRae <allan@archlinux.org>
This enables us to extract files in the source array and ensures that we
can decompress files if the uncompressed signature is served.
Signed-off-by: Morten Linderud <morten@linderud.pw>
Signed-off-by: Allan McRae <allan@archlinux.org>
buildenv is set once for build() and a second time for package(). When
using both distcc and ccache, this lead to CCACHE_PREFIX="distcc distcc"
in package(), which breaks PKGBUILDs that execute the compiler in
package() because distcc complains:
distcc[383041] (main) CRITICAL! distcc seems to have invoked itself
recursively!
Avoid causing this error by only adding "distcc" to CCACHE_PREFIX if
it's not yet there.
Signed-off-by: Matti Niemenmaa <matti.niemenmaa+git@iki.fi>
Signed-off-by: Allan McRae <allan@archlinux.org>
It updates the stripped/objcopied file by creating a temp file,
chown/chmodding it, and replacing the original file. But upstream
binutils has CVE-worthy issues with this if running strip as root, and
some recent versions of strip don't play nicely with fakeroot.
Also, this has always destroyed xattrs. :/
Sidestep the issue by telling strip/objcopy to write to a temporary
file, and manually dump the contents of that back into the original
binary. Since the original binary is intact, albeit with different
contents, it retains its correct attributes in fakeroot.
Signed-off-by: Eli Schwartz <eschwartz@archlinux.org>
Signed-off-by: Allan McRae <allan@archlinux.org>
With commit 74aacf4495 creating uncompressed .tar
packages fails.
-> Compressing package...
/usr/share/makepkg/util/compress.sh: line 70: COMPRESS.TAR[@]: invalid variable name
bsdtar: Write error
Empty the '$ext' variable for the '.tar' extension in get_compress_command() to
fix this. We would fallback to cat for 'tar' anyways.
Signed-off-by: Michael Straube <michael.straubej@gmail.com>
Signed-off-by: Allan McRae <allan@archlinux.org>
Fix typo in a comment in tidy_emptydirs().
Signed-off-by: Michael Straube <michael.straubej@gmail.com>
Signed-off-by: Allan McRae <allan@archlinux.org>
get_compression_command() can now be used to do upfront checks for
whether a given extension is known to do something successfully. This is
useful when writing tools in which an unknown compression type is a
fatal error.
Signed-off-by: Eli Schwartz <eschwartz@archlinux.org>
Signed-off-by: Allan McRae <allan@archlinux.org>
When a .SRCINFO file is generated via `makepkg --printsrcinfo`, each
section is concluded with an empty line. This means that at the end of
the file, an empty line remains. This is considered a trailing
whitespace error. In fact, `git diff --check` will warn about this,
saying "new blank line at EOF."
Instead of closing each section off with an empty line, use the empty
line to separate sections, omitting the empty line at the end of the
file.
Signed-off-by: Denton Liu <liu.denton@gmail.com>
Signed-off-by: Allan McRae <allan@archlinux.org>
We pass this to gpg -u and this gpg option can accept a number of
different formats, not just the historical hexadecimal fingerprint we
assumed. We should not barf hard if a format is used which happens to
contain spaces.
This also fixes a validation bug. When we initially check if the desired
key is available, we don't quote spaces, so gpg goes ahead and treats
each space-separated string as a *different key* to search for,
returning partial matches, and returning success if at least one key is
found. But gpg --detach-sign -u will certainly not accept multiple keys!
Fixes FS#66949
Signed-off-by: Eli Schwartz <eschwartz@archlinux.org>
Signed-off-by: Allan McRae <allan@archlinux.org>
In commit 882e707e40 we changed message
output to go to stdout by default, unless it was an error. The plain()
function doesn't *look* like an error function, but in practice it was
-- it's used to continue multiline messages, and all in-tree uses were
for warning/error.
This is a problem both because we're sending output to the wrong place,
and because in some cases, we were performing error logging from a
function which would otherwise return a value to be captured in a
variable using command substution.
Fix this and straighten out the API by providing two functions: one for
continuing msg output, and one which wraps this by sending output to
stderr, for continuing error output.
Change all callers to use the second function.
This was broken in commit 882e707e40,
which changed 'plain()' messages to go to stdout, which was then
captured as the download client in question: cmdline=("Aborting...").
The result was a very confusing error message e.g.
/usr/share/makepkg/source/file.sh: line 72: $'\E[1m': command not found
or with makepkg --nocolor:
/usr/share/makepkg/source/file.sh: line 72: Aborting...: command not found
The problem here is that we checked to see if an asynchronous subshell,
in our case <(...), failed, by checking if its captured stdout is
non-empty. Which is terrible, and also a limitation of old bash. But
bash 4.4 can use wait $! to retrieve the return value of an asynchronous
subshell. Now we target that as our minimum, we can sanely handle errors
in such functions.
Losing error messages on stdout by capturing them in a variable instead
of printing them, continues to be a problem, but this will be fixed
systematically in a later commit.
Signed-off-by: Eli Schwartz <eschwartz@archlinux.org>
Signed-off-by: Allan McRae <allan@archlinux.org>
If something like source=(..."#commit=") is used, e.g. due to failed
variable expansion, we try to check out an empty refspec as nothing at
all, and end up just running "git checkout". This happens because we
fail at variable expansion too -- so let's quote our variables properly
and make sure git sees this as an empty refspec, so it can error out.
Also make sure it is interpreted as a ref instead of a path.
Signed-off-by: Eli Schwartz <eschwartz@archlinux.org>
Signed-off-by: Allan McRae <allan@archlinux.org>
It's either a waste of work, or triggers edge cases in some packages
(like coreutils-8.31) where the source file is readonly and cp gets a
permission denied error trying to overwrite it with an identical copy of
itself.
Also while we are at it, make the variable names be something readable,
because I could barely tell what this was doing while editing it.
Signed-off-by: Eli Schwartz <eschwartz@archlinux.org>
Signed-off-by: Allan McRae <allan@archlinux.org>
Checksums arrays should be filled with values provided by upstream. We
currently have md5 set as an unsecure default, and are constantly asked to
change it to sha2. However, just changing the default to a stronger checksum
gives the user the impression that "makepkg -g" checksums are perfect.
Instead, change the default checksum to a CRC, to make it clear that any
checksum generated purely by "makepkg -g" is not ideal.
Signed-off-by: Allan McRae <allan@archlinux.org>
Extracting function variables containing arbitrarily scoped variables of
arbitrary nature is a disaster, but let's at least cover the common case
of using the actual '$pkgname' in an install/changelog file. It's the
odd case of actually being basically justified use of disambiguating
between the same variable used in multiple different split packages...
and also, --printsrcinfo already uses and overwrites the variable
'pkgname' in pkgbuild_extract_to_srcinfo, so this "works" in .SRCINFO
but doesn't work in .src.tar.gz
It doesn't work in lint_pkgbuild either, but in that case the problem is
being too permissive, not too restrictive -- we might end up checking
the same file twice, and printing that it is missing twice.
Fixes FS#64932
Signed-off-by: Eli Schwartz <eschwartz@archlinux.org>
Signed-off-by: Allan McRae <allan@archlinux.org>
In commit 9c817b6549 we made these sources
extendable, and heuristically determined the correct extraction
functions to use. But our fallback for protos that didn't have an exact
extract_* function didn't take into account that 'extract_file' matches
an actual proto... so we passed the netfile in while the function
expected a file.
Solution: the function should expect a netfile too, thereby allowing us
to delay an attempted resolution of netfile -> file, to the one case
where it is actually used. This makes us slightly more efficient in the
non-file case, makes our functions a bit more consistent, and makes
file:// extraction work again.
Fixes FS#64648
Signed-off-by: Eli Schwartz <eschwartz@archlinux.org>
Signed-off-by: Allan McRae <allan@archlinux.org>
Previously parseopts checked if there was an argument by checking
that the string was non-empty, resulting in empty arguments being
incorrectly considered non-existent. This change makes parseopts check
if arguments exist at all, rather than checking that they are non-empty
Signed-off-by: Ethan Sommer <e5ten.arch@gmail.com>
Signed-off-by: Allan McRae <allan@archlinux.org>
Adds a "?" suffix that can be used to indicate that an option's argument is
optional.
This allows options to have a default behaviour when the user doesn't
specify one, e.g.: --color=[when] being able to behave like --color=auto
when only --color is passed
Options with optional arguments given on the command line will be returned
in the form "--opt=optarg" and "-o=optarg". Despite that not being the
syntax for passing an argument with a shortopt (trying to pass -o=foo
would make -o's argument "=foo"), this is done to allow the caller to split
the option and its optarg easily
Signed-off-by: Ethan Sommer <e5ten.arch@gmail.com>
Reviewed-by: Dave Reisner <dreisner@archlinux.org>
Signed-off-by: Allan McRae <allan@archlinux.org>
Exclude files with hardlinks when cat'ing all the files, and do a second
run to look at each file with hardlinks, keep track of the ones we've
already operated on, and only cat each inode once. Then use "wc -c" to get
the size of all (deduplicated) files the same way we were already doing.
Original-patch-by: Ronan Pigott <rpigott@berkeley.edu>
Signed-off-by: Eli Schwartz <eschwartz@archlinux.org>
Signed-off-by: Allan McRae <allan@archlinux.org>
zipman:
read -r protects against those evil manpages whose filenames contain
backslash escapes, (muahahaha?)
IFS= read protects against filenames with:
- leading whitespace (but no one is actually stupid enough to configure
their MAN_DIRS=() in makepkg.conf with such silly directories, *right*?)
- trailing whitespace (but likewise, no one should be stupid enough to
write an uncompressed manpage for section '1 ' or something)
Also fix several other cases where we read filenames without protecting
against surrounding whitespace, or without using null-delimited
filenames when we could trivially do so.
Signed-off-by: Eli Schwartz <eschwartz@archlinux.org>
Signed-off-by: Allan McRae <allan@archlinux.org>
makepkg now complains when PACKAGER is not in the format
"name <email>".
Hide this warning when PACKAGER is unset but still warn if it is set to
something out of format.
Signed-off-by: Allan McRae <allan@archlinux.org>
Not all compression types can be detected in the seccomp sandbox, so we
need to disable it. This requires either configuring makepkg to know the
sandbox is available, or checking for file >= 5.38 in which the sandbox
option is a no-op even when seccomp is disabled.
- Requires autoconf-archive for autotools version compare macro.
- meson version comparison could be made a lot simpler using meson-git.
Fixes FS#58626
Signed-off-by: Eli Schwartz <eschwartz@archlinux.org>
Signed-off-by: Allan McRae <allan@archlinux.org>
