When the output is going to a file, glibc seems to buffer way too much
making it hard to monitor progress while tailing a file.
Signed-off-by: Simo Leone <simo@archlinux.org>
Signed-off-by: Dan McGee <dan@archlinux.org>
If the delta line doesn't match our regex, we won't go and process it,
possibly walking off the end of the string.
Signed-off-by: Dan McGee <dan@archlinux.org>
This patch slightly modifies pacman.c/_parseconfig():
See FS#12148. Now pacman prints the following error message in that case:
"error: could not register 'unstable' database (could not open database)"
I also added an error message for alpm_db_setserver() error.
I changed the "return(1);" scheme to "ret = 1; goto cleanup;" to make
sure that we free allocated memory and close open files.
Signed-off-by: Nagy Gabor <ngaba@bibl.u-szeged.hu>
Signed-off-by: Dan McGee <dan@archlinux.org>
Not only does this require less sed-magic, it also fixes FS#12286 where
fetching the revision number fails if mercurial is in compact mode.
Signed-off-by: Dan McGee <dan@archlinux.org>
This patch addresses quite a few lingering issues in the pacman-optimize
script. FS#11767 provoked this look-over and the following issues were
noticed and fixed:
* If an alternate dbroot was specified, then the lockfile location was never
updated to reflect it. The lockfile location is now set after all dbpath
initialization.
* The inclusion of a trailing slash on dbroot was problematic and led to the
following command being executed:
bsdtar -xpf /tmp/pacman-optimize.p12Q4vAUWY/pacman-db.tar.gz \
-C /var/lib/pacman/.new/
It is doubtful we meant to create a hidden directory like this below our
database root, only to go and delete it a second later and then
re-extract. Fix the whole thing by ensuring our dbpath has its trailing
slash stripped and then appending it when necessary.
* The DB extraction was performed twice for no real apparent reason. This
opens the door for extraction problems the second time around, leaving you
with no original database to fall back to. Change the behavior so we only
extract once, and then perform a directory shuffle once we verify the
checksums are correct.
* Perform an explicit sync after we drop the new database on the disk. It
should work better this way.
* Tighten up our check for a pacman lockfile and the time we create one.
There is still a possible race condition but the window is shorter.
Signed-off-by: Dan McGee <dan@archlinux.org>
This is similar to the change we made in makepkg so it is cross-platform
compatible and doesn't require coreutils.
Signed-off-by: Dan McGee <dan@archlinux.org>
We never had a call to pclose() in here before, leaving our file descriptor
in some sort of limbo state. In addition, clean up some of the other logic
such as directly calling exit(1) on a popen() failure rather than going to
our cleanup block, and handling and respecting the exit status of the
subprocess correctly.
Signed-off-by: Dan McGee <dan@archlinux.org>
Calling printf() in a signal handler can be dangerous, so avoid it by
writing directly which is guaranteed to be safe according to signal(7).
Signed-off-by: Dan McGee <dan@archlinux.org>
I mess this up more often than not, and maybe this will do the trick. Remove
the --enable-asciidoc option as it has been superseded by the --disable-doc
option in usefulness. If you want to skip building docs, you skip building
all docs which is much easier when it comes to ensuring the make 'dist' and
'distcheck' targets will always build the manpages and always build the most
up to date manpages.
Developers shouldn't be affected in their normal builds, nor should end
users of the source tarball.
Signed-off-by: Dan McGee <dan@archlinux.org>
Do the checks in the tests that need it, and get rid of some of the
cluttered output when it is not available (one line per test run).
Signed-off-by: Dan McGee <dan@archlinux.org>
Every call to getcols() results in two ioctl() calls, which we really didn't
need as changing the number of columns in mid-print would be pretty crazy.
Signed-off-by: Dan McGee <dan@archlinux.org>
It is possible to throw EINTR from a system call such as open(), close(), or
waitpid() if custom signal handlers are set up and they are not initialized
with the SA_RESTART flag. This was noticed by Andreas Radke when ^C (SIGINT)
was given during the call to waitpid(), causing it to throw the EINTR error
and we could not accommodate it.
Simply wrap these calls in a simple loop that allows us to retry the call if
interrupted.
Signed-off-by: Dan McGee <dan@archlinux.org>
Because libalpm always returns a root path with a trailing slash, when we
use it to create our unspecified paths we get double slashes in the result.
Use the fix suggested by Jürgen Hötzel to remedy this.
Signed-off-by: Dan McGee <dan@archlinux.org>
This fixes FS#11339, which is a regression of commit 89c2c5196:
When totaldownload is enabled, the database downloading percent (-Sy) is
always at 0. That is because we have no guarantee that the totaldownload
callback was called by libalpm. In particular, it is not called (and it
would not make sense to) when a single file is downloaded, like it is the
case with databases.
So the correct way to detect if totaldownload should be used is checking
both config->totaldownload and list_total, like it was already done in
several places in the cb_dl_progress function.
Signed-off-by: Xavier Chantry <shiningxc@gmail.com>
Signed-off-by: Dan McGee <dan@archlinux.org>
A source entry can now have the following form, to specify a different
filename :
"filename::http://path/to/file"
Of course, the old syntax is still supported :
"http://path/to/file"
And as before, in the second case, the filename used is simply "file".
This fixes FS#11292, because handling multiple source files with the same
name is now possible (just choose a different filename).
But it will also allow to deal much more nicely with funny url like this by
using a sane filename (and unfortunately, there are quite a few) :
http://www.vim.org/scripts/download_script.php?src_id=6992
Signed-off-by: Xavier Chantry <shiningxc@gmail.com>
Signed-off-by: Dan McGee <dan@archlinux.org>
* Update all .po files because of the last "-q,--quiet" fix.
Also for some strange reason, en_GB was missing a few c-format tags.
* Finally, delete all unused translations.
Signed-off-by: Xavier Chantry <shiningxc@gmail.com>
Signed-off-by: Dan McGee <dan@archlinux.org>
We had one "None" and one "None\n" string; we can let the program do the
addition of the newline so we don't have to.
Signed-off-by: Dan McGee <dan@archlinux.org>
* Change the return values to be more informative.
It was previously boolean, only indicating if a sync package was newer than
a local package.
Now it is a simple wrapper to vercmp, handling the force flag.
* Remove the verbose output from _alpm_pkg_compare_versions.
The "force" message is not so useful.
The "package : local (v1) is newer than repo (v2)" message can be moved to
-Su operation.
For the -S operation, it is better to have something like :
"downgrading package from v1 to v2"
* Don't display the "up to date -- skipping" and "up to date -- reinstalling"
messages, when the local version is newer than the sync one.
* Fix the behavior of --needed option to not skip a target when the local
version is newer, and clarify its description.
* Add a new alpm_pkg_has_force function
This allows us to access the pkg->force field like any other package fields.
Signed-off-by: Xavier Chantry <shiningxc@gmail.com>
Signed-off-by: Dan McGee <dan@archlinux.org>
This is a work around for FS#8725.
There are some bad combination of proxies and mirrors where the Content
Length is not returned, and thus the progress bar can't be displayed
correctly.
Dan: Note that this patch also adds a "downloading" message when the
progress bar is disabled, which was formerly not indicated at all in the
output.
Signed-off-by: Xavier Chantry <shiningxc@gmail.com>
Signed-off-by: Dan McGee <dan@archlinux.org>
The HACKING file seemed to be broken :
http://archlinux.org/pacman/HACKING.html
And indeed, running asciidoc HACKING issued a number of warnings :
WARNING: HACKING: line 27: missing [paradef-default] C-style entry
type: numbered : expected 1 got 3
WARNING: HACKING: line 44: list item 3 out of sequence
WARNING: HACKING: line 49: missing [paradef-default] C-style entry
type: numbered : expected 2 got 4
WARNING: HACKING: line 62: list item 4 out of sequence
type: numbered : expected 3 got 5
WARNING: HACKING: line 69: list item 5 out of sequence
type: numbered : expected 4 got 6
WARNING: HACKING: line 75: list item 6 out of sequence
type: numbered : expected 5 got 7
WARNING: HACKING: line 83: list item 7 out of sequence
WARNING: HACKING: line 104: missing [paradef-default] C-style entry
WARNING: HACKING: line 116: missing [paradef-default] C-style entry
WARNING: HACKING: line 126: missing [paradef-default] C-style entry
I just followed the syntax example there :
http://www.methods.co.nz/asciidoc/userguide.html#X56
And all is fine now :)
Signed-off-by: Xavier Chantry <shiningxc@gmail.com>
Signed-off-by: Dan McGee <dan@archlinux.org>
pacman already localizes the yesno stuff, so doing the same in makepkg is
more consistent.
Signed-off-by: Xavier Chantry <shiningxc@gmail.com>
Signed-off-by: Dan McGee <dan@archlinux.org>
list_display puts several members on the same line, which is not appropriate
for optdepends:
Optdepends: foo: feature1 bar: feature2 baz: feature3
The new list_display_linebreak function puts every member on its own line,
which is much better with optdepends:
Optdepends: foo: feature1
bar: feature2
baz: feature3
Signed-off-by: Nagy Gabor <ngaba@bibl.u-szeged.hu>
[Xav: implement this new behavior as a new function rather than as a
parameter of list_display]
Signed-off-by: Xavier Chantry <shiningxc@gmail.com>
Signed-off-by: Dan McGee <dan@archlinux.org>
So dump_pkg_full will indent all strings correctly.
Signed-off-by: Nagy Gabor <ngaba@bibl.u-szeged.hu>
[Xav: add string_length function]
Signed-off-by: Xavier Chantry <shiningxc@gmail.com>
Signed-off-by: Dan McGee <dan@archlinux.org>
This fixes FS#11283 , which was originally reported on the forums :
http://bbs.archlinux.org/viewtopic.php?id=53794
Insensitive comparisons were implicitly made before since md5sum --status was
used for checking. Now that we use openssl and compare checksums manually in
bash, we lost that feature.
This can be easily reintroduced using tr '[A-F]' '[a-f]'
What convinced me to fix it is that the md5 command line tool generates md5sums
in upper case by default :
http://www.fourmilab.ch/md5/
And finally, A-F and a-f are the same in hex and both are used.
Signed-off-by: Xavier Chantry <shiningxc@gmail.com>
Signed-off-by: Dan McGee <dan@archlinux.org>
Commit 149839c539 introduced a small behavior regression as a drawback
for a better portability. repo-add now includes the approximate size (to the
nearest KB) rather than an exact size due to the switching of the du command
to a more portable form. Instead of sacrificing the exact size, use
configure to help us determine a valid command to acquire our filesize and
place it in the sync database.
Signed-off-by: Dan McGee <dan@archlinux.org>
