This patch changes a variety of small things related to our pkghash
implementation with an eye toward performance, especially on native
32-bit systems.
* Use `unsigned int` rather than `size_t` for hash sizes. We already
return ERANGE for any attempted creation of a hash greater than 1
million elements, so unsigned int is more than large enough for our
purposes. Switching to this type allows 32 bit systems to do native
math without helper functions from libgcc.
* _alpm_pkghash_create() now internally adds extra padding for
additional array elements, rather than that being the responsibility of
the caller.
* #define values are moved into static const values in pkghash.c; a new
`stride` value is also extracted (but remains set at 1).
* Division and modulus operators are removed from the normal find and
add paths if possible. We store the upper limit of the number of
elements in the hash so we no longer need to calculate this every
element addition. When doing wraparound position calculations, we only
apply the modulus operator if the value is greater than the number of
buckets.
Signed-off-by: Dan McGee <dan@archlinux.org>
be_package.c: In function 'parse_descfile':
be_package.c:181:28: error: comparison between signed and unsigned
integer expressions [-Werror=sign-compare]
ptr - key + 2 is guaranteed to be > 0 so we can cast to size_t
Signed-off-by: Allan McRae <allan@archlinux.org>
Signed-off-by: Dan McGee <dan@archlinux.org>
Now that filelist arrays know their own size, we don't need to do the
bookkeeping we used to do when they were linked lists. Remove some of
the counter variables and use math instead.
Signed-off-by: Dan McGee <dan@archlinux.org>
This reduces the number of functions we call by log(n) in this function,
and the inlined version is trivial and barely increases the size of the
function.
Signed-off-by: Dan McGee <dan@archlinux.org>
This reduces the number of regcomp() calls when parsing delta entries in
the database from once per entry to once for the entire context handle
by storing the compiled regex data on the handle itself. Just as we do
with the cURL handle, we initialize it the first time it is needed and
free it when releasing the handle.
A few other small tweaks to the parsing function also take place,
including using the stack to store the transient and short file size
string while parsing it.
When parsing a sync database with 1378 delta entries, this reduces the
time of a `pacman -Sl deltas` operation by 50% from 0.22s to 0.12s.
Signed-off-by: Dan McGee <dan@archlinux.org>
In commit 4c5e7af32f, we changed this code to use the regex gathered
substrings. However, we failed to correctly store the delta file name
(leaking memory), as well as freeing the temporary string used to hold
the file size string.
Signed-off-by: Dan McGee <dan@archlinux.org>
More than likely the compiler will do the three operation breakdown we
had here before (2 shifts + subtraction), but let the compiler do the
optimizations and make the actual operation more obvious. This actually
slightly shrinks the function binary size, likely due to instruction
reordering or something.
Signed-off-by: Dan McGee <dan@archlinux.org>
This eliminates the need for strtrim() usage completely, instead relying
on the fact that the only allowed delimiter between key and value is the
" = " string.
Signed-off-by: Dan McGee <dan@archlinux.org>
Extract the actual unlinking of files into a new method, which
eliminates a goto used for flow control. Also fix up a few small issues
in the code:
* Unnecessary (unsigned long) cast, use '%zd' instead
* Total up errors returned from unlink_file calls and return to caller
* Be consistent with scriptlets- we run pre_remove on dbonly, so we
should also run post_remove. Both can be disabled by way of the
--noscriptlet argument.
* Don't pass an invalid pointer to oldpkg to the event callbacks;
instead call the callback before we free the object.
Signed-off-by: Dan McGee <dan@archlinux.org>
Used in alpm_compute_md5sum() and alpm_compute_sha256sum().
Signed-off-by: Diogo Sousa <diogogsousa@gmail.com>
Signed-off-by: Dan McGee <dan@archlinux.org>
Ensures that config.h is always ordered correctly (first) in the
includes. Also means that new source files get this for free without
having to remember to add it.
We opt for -imacros over -include as its more portable, and the
added constraint by -imacros doesn't bother us for config.h.
This also touches the HACKING file to remove the explicit mention of
config.h as part of the includes.
Signed-off-by: Dave Reisner <dreisner@archlinux.org>
Signed-off-by: Dan McGee <dan@archlinux.org>
Mostly a waste of time. Sure, we no longer make sure your pacman
database partition has enough space, but if you are using this option
you better know what you are doing anyway.
Signed-off-by: Dan McGee <dan@archlinux.org>
It is quite easy to hoist this potentially repeated computation out of
the loop; even if we don't end up using it, it is super cheap to do it
only once.
Signed-off-by: Dan McGee <dan@archlinux.org>
The return should probably be checked to ensure its not longer than
PATH_MAX, but I have no idea what the correct behavior is when that
happens.
Signed-off-by: Dave Reisner <dreisner@archlinux.org>
Signed-off-by: Dan McGee <dan@archlinux.org>
As per HACKING file, we use 'CTRL(' rather than 'CTRL ('
Signed-off-by: Dave Reisner <dreisner@archlinux.org>
Signed-off-by: Dan McGee <dan@archlinux.org>
Since we know the length of the line, we can use this all the way
through and do a cheaper operation than strdup() by just invoking malloc
and memcpy directly.
Signed-off-by: Dan McGee <dan@archlinux.org>
We were using the size of a pointer, not the size of the whole
archive_read_buffer struct. Thanks to Clang/LLVM 3.0 and Allan/Dave in
IRC for finding this one.
Signed-off-by: Dan McGee <dan@archlinux.org>
We had a 16 KiB limit on database signatures, we should do the same here
too to have a slight sanity check, even if we can't do so for the
package itself yet.
Signed-off-by: Dan McGee <dan@archlinux.org>
We do this in several of the package duplication steps; add a helper
function for doing so to reduce some of the repetitive code.
Also add a free_deplist function for our repeated depend list free calls
of both the data and the list.
Signed-off-by: Dan McGee <dan@archlinux.org>
Made existing documentation more consistent and added
documentation where there was none. One function still
needs documentation and is marked with 'TODO'.
Signed-off-by: Andrew Gregory <andrew.gregory.8@gmail.com>
Signed-off-by: Dan McGee <dan@archlinux.org>
This moves the common setup code of about 5 different callers into one
method. Error messages will now be common and shared in all places;
several paths did not have any messages at all before.
In addition, we now pick an ideal block size for the archive read based
off the larger value of our default buffer size or the st.st_blksize
field. For a filesystem such as NFS, this is often much larger than the
default 8192- values such as 32768 and 131072 are common.
Signed-off-by: Dan McGee <dan@archlinux.org>
When doing a bare -U operation on a local package that doesn't pull in
any dependencies from the sync databases, we can get away with missing
database files. This makes the check conditional on no sync targets
found in the target list. This is not the prettiest code here so we have
a bit of hackish behavior required to straighten both the behavior and
the nonsensical error message out.
Addresses FS#26899.
Signed-off-by: Dan McGee <dan@archlinux.org>
This is consistent with the other enums and structs, and should be
slightly more readable.
Signed-off-by: Jonathan Conder <jonno.conder@gmail.com>
Signed-off-by: Dan McGee <dan@archlinux.org>
Bump the version, update the translation template files, and fill in
NEWS with relevant commits and changes since 4.0.0.
Signed-off-by: Dan McGee <dan@archlinux.org>
This is work originally provided by Sascha Kruse on FS#20360 with only
minor adjustments to the implementation. It's been expanded to cover:
NoUpgrade, NoExtract, IgnorePkg, IgnoreGroup.
Adds tests ignore008, sync139, sync502, and sync503.
Also satisfies FS#18988.
Original-work-by: Sascha Kruse <knopwob@googlemail.com>
Signed-off-by: Dave Reisner <dreisner@archlinux.org>
This is a simple change that allows comparions to be more in line with
how other checks are done. It will be necessary for ensuing patchwork
that implements fnmatch for comparing and assumes a specific argument
ordering.
Signed-off-by: Dave Reisner <dreisner@archlinux.org>
The point of this early compare to NULL byte check was so we could bail
early and skip the strcmp() call. Given we weren't doing the check
right, this never exited early. Fix it to work as intended.
Noticed-by: Pepe Juárez <trulustapa@gmail.com>
Signed-off-by: Dan McGee <dan@archlinux.org>
This is a trivial operation that doesn't require calling a function over
and over- just do some math and indexing into a character array.
Signed-off-by: Dan McGee <dan@archlinux.org>
This gives us a bit more control and over the archive reading process,
and a bit less is done behind the scenes. It also allows us to use
fstat() in preference to stat(), which should avoid some potential race
conditions.
Some reorganization is necessary to move the stat calls after the open()
calls. Error handling and cleanup in general is also improved, as we had
several potential memory and file handle leaks before in some error
paths.
Signed-off-by: Dan McGee <dan@archlinux.org>
This removes an unnecessary level of buffering. We are not doing
line-based I/O here, so we can read in blocks of 8K at a time directly
from the file.
Signed-off-by: Dan McGee <dan@archlinux.org>
