74274b5dc3 which added the real_line_size to the buffer struct
didn't properly account for what happens when archive_fgets has to loop
more than once to find the end of a line. In most cases, this isn't a
problem, but could potentially cause a longer line such as PGP signature
to be improperly read.
This patch fixes the oversight and focuses on only calculating the line
length when we hit the end of line marker. The effective length is then
calculated via pointer arithmetic as:
(start_of_last_read + read_length) - start_of_line
Signed-off-by: Dave Reisner <dreisner@archlinux.org>
Take advantage of the fact that our filelists are arrays sorted by
filename with a known length and use a binary search. This should speed
up file conflict checking, particularly when larger packages are
involved.
Signed-off-by: Dave Reisner <dreisner@archlinux.org>
Signed-off-by: Dan McGee <dan@archlinux.org>
This may very well be a no-op, but better safe than sorry.
Signed-off-by: Dave Reisner <dreisner@archlinux.org>
Signed-off-by: Dan McGee <dan@archlinux.org>
On the assumption that these arrays are already mostly sorted, use the
standard quicksort method to sort the files arrays. The files_msort
function name is tweaked to give it a more general name to reflect this
change.
Signed-off-by: Dave Reisner <dreisner@archlinux.org>
Signed-off-by: Dan McGee <dan@archlinux.org>
When checking if a package owns a directory, it is important to check
not only that all the files in the directory are part of the package,
but also if the directory is part of a package. This catches empty
subdirectories during conflict checking for directory to file/symlink
replacements.
Signed-off-by: Allan McRae <allan@archlinux.org>
Signed-off-by: Dan McGee <dan@archlinux.org>
When two packages own an empty directory, pacman finds no conflict when
one of those packages wants to replace the directory with a file or a
symlink. When it comes to actually extracting the new file/symlink,
pacman sees the directory is still there (we do not remove empty
directories if they are owned by a package) and refuses to extract.
Detect this potential conflict early and bail. Note that it is a
_potential_ conflict and not a guaranteed one as the other package owning
the directory could be updated or removed first which would remove
the conflict. However, pacman currently can not sort package installation
order to ensure this, so this conflict requires manual upgrade ordering.
Signed-off-by: Allan McRae <allan@archlinux.org>
Signed-off-by: Dan McGee <dan@archlinux.org>
Only load filesystem details for the mount points that we're actually
going to write to. This reduces our syscall count considerably. In the
case of installation, we would actually stat every mountpoint twice (an
extra round for download diskspace) which means (on my system) a total
of 60 syscalls to write to 3 partitions when installing the kernel
package. This change reduces the 60 syscalls down to the expected 3.
A slight debug output change is added here to discern between a
mountpoint added to our linked list versus when we actually load the fs
info.
Signed-off-by: Dave Reisner <dreisner@archlinux.org>
Signed-off-by: Dan McGee <dan@archlinux.org>
add mount_point_load_fsinfo() for platforms using getmntent().
Dan: move the #ifdef slightly so we don't have unused functions on
certain platforms (e.g., OS X).
Signed-off-by: Dave Reisner <dreisner@archlinux.org>
Signed-off-by: Dan McGee <dan@archlinux.org>
Similar to the case for makedepends, it is useful to be able to
access this information without parsing a PKGBUILD.
Signed-off-by: Allan McRae <allan@archlinux.org>
If known, callers can pass the line size to this function in order to
avoid an strlen call. Otherwise, they simply pass 0 and
_alpm_strip_newline will do the call instead.
Signed-off-by: Dave Reisner <dreisner@archlinux.org>
Signed-off-by: Dan McGee <dan@archlinux.org>
We inevitably call strlen() or similar on the line returned from
_alpm_archive_fgets(), so include the line size of the interesting line
in the struct.
Signed-off-by: Dave Reisner <dreisner@archlinux.org>
Signed-off-by: Dan McGee <dan@archlinux.org>
With lazy loading in place, it's now quite obvious that we aren't
necessarily checking the right mountpoint for necessary download space.
Signed-off-by: Dave Reisner <dreisner@archlinux.org>
This is useful for tools that automatically rebuild packages and
thus require to generate a build order. These entries are skipped
by pacman.
Signed-off-by: Allan McRae <allan@archlinux.org>
Signed-off-by: Dan McGee <dan@archlinux.org>
Apparently gcc 4.7 has decided that -Wshadow warnings aren't worth
reporting anymore even with the flag enabled. These were found on
an Ubuntu 10.04 install.
Signed-off-by: Dave Reisner <dreisner@archlinux.org>
Signed-off-by: Dan McGee <dan@archlinux.org>
This path is rarely (read: never) taken in any normal run of the code,
so injecting the fprintf() call everywhere with the macro is a bit
overkill. Instead, add a lightweight _alpm_alloc_fail() function that
gets called instead.
This does have a reasonable effect on the size of the generated code;
most places using the macros provided by util.c have their code size
reduced.
Signed-off-by: Dan McGee <dan@archlinux.org>
Increment the strlen() provided value by 1 for the NULL byte so we use
the right value in all three places we later reference it.
Signed-off-by: Dan McGee <dan@archlinux.org>
There is little reason here to grab 4K from the heap only to return it a
few lines later. Instead, just use the stack to hold the returned value
saving ourselves the malloc/free cycle.
Signed-off-by: Dan McGee <dan@archlinux.org>
No one seems to do this "correctly", but for the sake of having an easy
method of detecting the presence and version of libalpm on a given
system, we provide a straightforward .pc file.
Signed-off-by: Dave Reisner <dreisner@archlinux.org>
There isn't a whole lot of reason other than code clarity for this, but
it makes it a bit more obvious where multivalued attributes start.
Signed-off-by: Dan McGee <dan@archlinux.org>
Detected by clang scan-build static code analyzer.
* Don't attempt to free an uninitialized gpgme key variable
* Initialize answer variable before asking frontend a question
* Pass by reference instead of value if uninitialized fields are
possible in download signal handler code
* Ensure we never call strlen() on NULL payload->remote_name value
Signed-off-by: Dan McGee <dan@archlinux.org>
Not sure why this one wasn't showing up on x86_64, but this fixes the
compile on i686.
diskspace.c: In function 'calculate_removed_size':
diskspace.c:247:4: error: assuming signed overflow does not occur when negating a division [-Werror=strict-overflow]
cc1: all warnings being treated as errors
Signed-off-by: Dan McGee <dan@archlinux.org>
This fixes a bunch of small issues in order to enable a clean
successful build with a crazy number of GCC warning flags. A lot of
these changes are covered by -Wshadow, -Wformat-security, and
-Wstrict-overflow=5.
Signed-off-by: Dan McGee <dan@archlinux.org>
Continue the trend of not touching the environment CFLAGS, ensuring that
the user always has the final say.
Signed-off-by: Dave Reisner <dreisner@archlinux.org>
- handle gpgme libs and cflags separately rather than appending to
CFLAGS and LDFLAGS
- be consistent in AC_LINK_IFELSE check for gpgme 1.3.0 (though this is
irrelephant since we don't actually run)
- be consistent with usage of "have" and "with" variables (this
actually ends up reducing SLOC)
- when voluntary detection fails, unset GPGME_CFLAGS and GPGME_LIBS
- when requested support fails the version check, complain about the min
version.
Signed-off-by: Dave Reisner <dreisner@archlinux.org>
Failure isn't always due to the package file location not existing;
permission issues can also play a part on something like a FUSE-based
filesystem inaccessible to root.
Signed-off-by: Dan McGee <dan@archlinux.org>
The initial patch to implement this achieved nothing apart from
adding a configure option. This patch makes that configure option
do what it advertises.
Note that specifing any shell apart from /bin/sh causes testsuite
failures as /bin/sh is the only shell in the testing environment.
Bug-found-by: Dave Reisner <dreisner@archlinux.org>
Signed-off-by: Allan McRae <allan@archlinux.org>
Signed-off-by: Dan McGee <dan@archlinux.org>
Dan was right. This should have been FREE(), not free().
Signed-off-by: Dave Reisner <dreisner@archlinux.org>
Signed-off-by: Dan McGee <dan@archlinux.org>
Pull updates from transifex, run update-po on all files, fix a few
errors, and push them back to Transifex.
Signed-off-by: Dan McGee <dan@archlinux.org>
For key searches only, gpg2 will fail to lookup any and all keys that
are not prefixed with 0x.
Signed-off-by: Dave Reisner <dreisner@archlinux.org>
Signed-off-by: Dan McGee <dan@archlinux.org>
Using fputs should be faster as no format string parsing is required. It
also prevents silly errors related to unescaped '%' signs, and removes
the need to double them up in a lot of places.
Signed-off-by: Dan McGee <dan@archlinux.org>
This removes a call to _alpm_local_db_pkgpath() as well as an access()
call when reading the local database. This appears to be code from 2006
that has stuck around. We don't need it because:
1) We never use this path except to check it via access(); however, we
are already in a readdir() loop so it exists, or at least did at the
time of the call.
2) The fopen() and other calls will fail on accessing the database files
anyway, and we need to check those for errors.
Signed-off-by: Dan McGee <dan@archlinux.org>
In case we have a mirror failure, unlink_on_fail would remain set,
causing an interrupt in a successive download attempt to be wrongly
unlinked.
This also fixes a memory leak in the url member, as we would allocate
over the previous, unfreed URL.
Signed-off-by: Dave Reisner <dreisner@archlinux.org>
