Colin Walters [Wed, 21 Aug 2013 12:26:24 +0000 (08:26 -0400)]
doc: Add repo docs
Colin Walters [Tue, 20 Aug 2013 23:10:09 +0000 (19:10 -0400)]
libostree: Check out directories depth-first in serial, switch to sync API
The way we recurse into subdirectories in parallel makes it far too
easy to hit up against the arbitrary Linux fd limit of 1024.
Since the fix here is about dropping parallelism, let's just go all
the way for now and make a plain old synchronous API =(
This does simplify both internal callers which wanted a sync API
anyways.
https://bugzilla.gnome.org/show_bug.cgi?id=706380
Javier Martinez Canillas [Thu, 15 Aug 2013 10:33:25 +0000 (12:33 +0200)]
main: Add U-Boot bootlader backend support
This patch adds support to generate files that
can be used by Universal Bootloader (U-Boot).
U-Boot allows to modify boards default boot commands by
reading and executing a bootscript file or importing a
plain text file that contains environment variables that
could parameterize the boot command or a bootscript.
OSTree generates a uEnv.txt file that contains booting
information that is taken from Boot Loader Specification
snippets files as defined in the new OSTree deployment model:
https://wiki.gnome.org/OSTree/DeploymentModel2
On deploy or upgrade an uEnv.txt env var file is created
in the path /boot/loader.${bootversion}/uEnv.txt. Also, a
/boot/uEnv.txt symbolic link to loader/uEnv.txt is created
so U-Boot can always import the file from a fixed path.
Since U-Boot does not support a menu to list a set of
Operative Systems, the most recent bootloader configuration
from the list is used.
To boot an OSTree using the generated uEnv.txt file, a
board has to parameterize its default boot command using the
following variables defined by OSTree:
${kernel_image}: path to the Linux kernel image
${ramdisk_image}: path to the initial ramdisk image
${bootargs}: parameters passed to the kernel command line
Alternatively, for boards that don't support this scheme,
a bootscript that overrides the default boot command can be used.
An example of such a bootscript could be:
setenv scriptaddr
40008000
setenv kernel_addr 0x40007000
setenv ramdisk_addr 0x42000000
ext2load mmc 0:1 ${scriptaddr} uEnv.txt
env import -t ${scriptaddr} ${filesize}
ext2load mmc 0:1 ${kernel_addr} ${kernel_image}
ext2load mmc 0:1 ${ramdisk_addr} ${ramdisk_image}
bootm ${kernel_addr} ${ramdisk_addr}
Signed-off-by: Javier Martinez Canillas <javier.martinez@collabora.co.uk>
https://bugzilla.gnome.org/show_bug.cgi?id=706370
Javier Martinez Canillas [Tue, 20 Aug 2013 15:54:59 +0000 (17:54 +0200)]
admin: Extract ot_admin_join_config_lines() helper function
ot-bootloader-syslinux.c has a join_lines() function that is rather
generic and can be used in other places. Let's add it as a helper
function.
Signed-off-by: Javier Martinez Canillas <javier.martinez@collabora.co.uk>
https://bugzilla.gnome.org/show_bug.cgi?id=706370
Javier Martinez Canillas [Thu, 15 Aug 2013 22:59:01 +0000 (00:59 +0200)]
test: fix a trivial typo in libtests.sh
When running the test-admin-deploy-1.sh unit test,
cat shows the following error:
cat: boot/vmlinuz-3-6.0: No such file or directory
due a trivial typo in the kernel image file name.
Signed-off-by: Javier Martinez Canillas <javier.martinez@collabora.co.uk>
https://bugzilla.gnome.org/show_bug.cgi?id=706371
Colin Walters [Mon, 19 Aug 2013 14:16:05 +0000 (10:16 -0400)]
libostree: Extend gtk-doc coverage for refs and prune APIs
Colin Walters [Mon, 19 Aug 2013 13:01:20 +0000 (09:01 -0400)]
ostree.doap: Update description based on docs.
Colin Walters [Mon, 19 Aug 2013 13:00:02 +0000 (09:00 -0400)]
main: Use macro to reduce duplication among builtin prototypes
Just less code.
Colin Walters [Sun, 18 Aug 2013 11:51:25 +0000 (07:51 -0400)]
Revert "libotutil: Make use of GBytes in ot_variant_read()"
This reverts commit
c77908bf514d61e75798932f61b5b414d9e36a3c; we can't
do this since g_variant_get_data_as_bytes() is a GLib 2.36 API.
Colin Walters [Sun, 18 Aug 2013 11:50:50 +0000 (07:50 -0400)]
libotutil: Drop accidental use of GLib 2.36 API
Sticking with 2.34 for a while longer.
Colin Walters [Sat, 17 Aug 2013 11:59:23 +0000 (07:59 -0400)]
Use { 0, } for structure initialization rather than memset()
It's cleaner, safer, and I had a totally wrong idea stuck in my head
about why memset() should be used.
https://bugzilla.gnome.org/show_bug.cgi?id=705968
Colin Walters [Sat, 17 Aug 2013 21:05:17 +0000 (17:05 -0400)]
Release 2013.5
Colin Walters [Sat, 17 Aug 2013 19:21:31 +0000 (15:21 -0400)]
build: Fix and prettify build option output
The documentation one was broken by the gtk-doc changes, and let's
align things.
Sjoerd Simons [Mon, 12 Aug 2013 08:53:03 +0000 (10:53 +0200)]
ostree init: Fix mode list help string
It's archive-z2, not archive-z.
https://bugzilla.gnome.org/show_bug.cgi?id=705849
Sjoerd Simons [Mon, 12 Aug 2013 14:27:42 +0000 (16:27 +0200)]
Fix make distcheck
make distcheck was unhappy for various reasons:
* headers aren't data, so use _HEADERS otherwise compilation fails
* Mark the gir & typelib data as cleanfiles so they aren't left around
after make clean
* Don't nuke the .la file. This breaks make uninstall, leave it up to
distributions to not install .la files if they don't want them.
https://bugzilla.gnome.org/show_bug.cgi?id=705850
Colin Walters [Sat, 17 Aug 2013 18:08:04 +0000 (14:08 -0400)]
libostree: Delete some leftover remnant API for archive files
Colin Walters [Sat, 17 Aug 2013 17:57:22 +0000 (13:57 -0400)]
libostree: Increase gtk-doc coverage of ostree-repo.[ch]
Colin Walters [Sat, 17 Aug 2013 17:22:58 +0000 (13:22 -0400)]
libostree: Add gtk-doc section for ostree-repo.c
Colin Walters [Sat, 17 Aug 2013 14:54:00 +0000 (10:54 -0400)]
libostree: More gtk-doc updates for ostree-core
Colin Walters [Sat, 17 Aug 2013 13:22:08 +0000 (09:22 -0400)]
docs: Add an overview section, migrate some content from the wiki
Colin Walters [Sat, 17 Aug 2013 12:41:31 +0000 (08:41 -0400)]
libostree: Fix many gtk-doc warnings
Colin Walters [Sat, 17 Aug 2013 12:23:28 +0000 (08:23 -0400)]
core: Drop duplicated type declarations
3 fewer gtk-doc warnings, 99 still left on the wall...
Colin Walters [Sat, 17 Aug 2013 12:21:04 +0000 (08:21 -0400)]
libostree: Document more core macros
Stef Walter [Wed, 14 Aug 2013 08:20:41 +0000 (10:20 +0200)]
Fix use of uninitialized memory in ostree_builtin_checksum()
https://bugzilla.gnome.org/show_bug.cgi?id=705968
Colin Walters [Sat, 17 Aug 2013 03:19:30 +0000 (23:19 -0400)]
TODO: Tweak
Colin Walters [Sat, 17 Aug 2013 03:14:30 +0000 (23:14 -0400)]
build: Fix the build without documentation
Just key everything of gtk-doc.
Colin Walters [Thu, 15 Aug 2013 10:34:48 +0000 (06:34 -0400)]
core: Remove old "archive" mode
We'll always have "bare" mode for keeping files-as-hardlinks as root.
But "archive" was my second attempt at a format for non-root file
storage, used by the gnome-ostree buildsystem which runs as non-root.
It was really handy to have a "tar" like mode where I can create
tarballs as a user, that contain files owned by root for example.
The "archive" mode stored content files as two pieces in the
filesystem; ".file" contained metadata, and ".filecontent" was the
actual content, uncompressed. The nice thing about this was that to
check out a tree as non-root, you could just hardlink into the repo.
However, archive was fairly bad for serving via HTTP; it required
*two* HTTP requests per content object, greatly magnifing the already
inefficient fetch process. So "archive-z2" was introduced.
To allow gnome-ostree to still check out trees as a user, the
"uncompressed-object-cache" was introduced, and that's how things have
been working for a while.
So we should just be able to kill this code. Specifically note just
how much better the stage_object() function became.
https://bugzilla.gnome.org/show_bug.cgi?id=706057
Colin Walters [Sat, 17 Aug 2013 02:54:35 +0000 (22:54 -0400)]
Add gtk-doc support
Yes, it's really me. Colin Walters. Writing documentation. You
don't need to do a DNA test.
Stef Walter [Wed, 14 Aug 2013 10:42:13 +0000 (12:42 +0200)]
Add 'ostree reset' command to undo a bad commit
Accepts the following arguments: ref checksum
Checks that the checksum is a parent of the ref before rewriting
the ref.
https://bugzilla.gnome.org/show_bug.cgi?id=705979
Colin Walters [Thu, 15 Aug 2013 22:33:19 +0000 (18:33 -0400)]
configure: Use AS_IF() consistently
See https://bugzilla.gnome.org/show_bug.cgi?id=681413
Sjoerd Simons [Mon, 12 Aug 2013 17:27:27 +0000 (19:27 +0200)]
Install systemd units in the right location
Debian uses /lib/systemd/system for system unit files, while i'm
putting ostree under the /usr prefix which means the hardcoded path
fails. Leave it to configure to work out the right location for systemd
units (method copied from pollkit).
Furthermore instead of installing the unit in local-fs.target.wants by
hand add a [Install] section so systemctl enable does the right thing
https://bugzilla.gnome.org/show_bug.cgi?id=705864
Colin Walters [Thu, 15 Aug 2013 18:21:26 +0000 (14:21 -0400)]
core: Use gs_stream_fstat()
Just a cleanup.
Colin Walters [Thu, 15 Aug 2013 15:17:16 +0000 (11:17 -0400)]
core: Add some more gtk-doc
Colin Walters [Thu, 15 Aug 2013 11:04:29 +0000 (07:04 -0400)]
core: Add some gtk-doc
Just documenting a few functions to get in the habit of things.
Stef Walter [Wed, 14 Aug 2013 09:47:30 +0000 (11:47 +0200)]
Add 'ostree log' command
Follows the parent of commits showing each in turn until it reaches
the top of the commit tree.
https://bugzilla.gnome.org/show_bug.cgi?id=705973
Stef Walter [Wed, 14 Aug 2013 09:46:50 +0000 (11:46 +0200)]
Add ostree_commit_get_parent() to get parent from variant
https://bugzilla.gnome.org/show_bug.cgi?id=705973
Stef Walter [Wed, 14 Aug 2013 09:25:32 +0000 (11:25 +0200)]
Intelligible display for 'ostree show'
Show something similar to git metadata display. Show raw variant
data when --raw is specified
https://bugzilla.gnome.org/show_bug.cgi?id=705973
Colin Walters [Wed, 14 Aug 2013 22:18:10 +0000 (18:18 -0400)]
repo: Add API to load any object as a stream
We have APIs to load metadata as variants, and files as parsed
content/info/xattrs, but for some cases such as static deltas, all we
want is to operate on all objects in their canonical representation.
https://bugzilla.gnome.org/show_bug.cgi?id=706031
Colin Walters [Wed, 14 Aug 2013 23:54:30 +0000 (19:54 -0400)]
libotutil: Make use of GBytes in ot_variant_read()
This is just cleaner; we avoid using GObject data, etc.
https://bugzilla.gnome.org/show_bug.cgi?id=706031
Colin Walters [Wed, 14 Aug 2013 22:17:09 +0000 (18:17 -0400)]
core: Add API to convert checksum -> csum in place
We already have the opposite, and this will be used in some
places to avoid a malloc.
https://bugzilla.gnome.org/show_bug.cgi?id=706031
Colin Walters [Wed, 14 Aug 2013 22:20:02 +0000 (18:20 -0400)]
libotutil: Add API to create an "ay" GVariant from GBytes
We used to have a version of this, but since I'm trying to use
GBytes more, this became a more common operation, and it's annoying
to type out the whole G_VARIANT_TYPE ("ay") each time, and pass
TRUE for trusted.
https://bugzilla.gnome.org/show_bug.cgi?id=706031
Jeremy Whiting [Wed, 14 Aug 2013 20:08:25 +0000 (14:08 -0600)]
trivial-httpd: Add --force-range-requests option to force range requests
This will be used to test resuming interrupted downloads for
ostree-pull.
With this option, if a whole file is asked for, only half of the file
is given. Then the client should retry with a range request, and
we'll give them the other half.
https://bugzilla.gnome.org/show_bug.cgi?id=705925
Colin Walters [Wed, 14 Aug 2013 18:46:46 +0000 (20:46 +0200)]
test-xattrs: Skip if current FS doesn't support user_xattr
tmpfs doesn't =(
Colin Walters [Wed, 14 Aug 2013 15:27:30 +0000 (17:27 +0200)]
core: Add a comment header about the format
This should probably be a design document or something, but this is
useful for now.
Stef Walter [Wed, 14 Aug 2013 08:10:08 +0000 (10:10 +0200)]
Don't pass non-checksums to ostree_repo_has_object()
The 'ostree show' command passed non-checksum arguments
to the ostree_repo_has_object() function which led to
an assertion.
https://bugzilla.gnome.org/show_bug.cgi?id=705967
Colin Walters [Wed, 14 Aug 2013 13:05:09 +0000 (15:05 +0200)]
tests: Add xattr test case
Since it was broken up until now for multiple attributes =(
Stef Walter [Wed, 14 Aug 2013 07:01:18 +0000 (09:01 +0200)]
Fix crasher when running 'ostree cat'
Cause was a buffer read overrun.
https://bugzilla.gnome.org/show_bug.cgi?id=705964
Stef Walter [Tue, 13 Aug 2013 13:42:06 +0000 (15:42 +0200)]
Add some verbose log output when xattr functions fail
If any of the system xattr functions fail, clearly write out the
reason in the verbose log output.
https://bugzilla.gnome.org/show_bug.cgi?id=705893
Stef Walter [Tue, 13 Aug 2013 13:41:12 +0000 (15:41 +0200)]
Don't use XATTR_REPLACE with lsetxattr()
If we pass XATTR_REPLACE then the attribute must already exist, which
is not our intent. Passing zero creates the attribute if necessary,
or replaces it when it already exists.
https://bugzilla.gnome.org/show_bug.cgi?id=705893
Stef Walter [Tue, 13 Aug 2013 13:40:06 +0000 (15:40 +0200)]
Properly separate sorted xattr names
We expect to be handling a string delimited by \0 characters, as
returned by llistxattr(). So stick to that behavior here.
https://bugzilla.gnome.org/show_bug.cgi?id=705893
Stef Walter [Tue, 13 Aug 2013 13:37:59 +0000 (15:37 +0200)]
Add some verbose logging to 'ostree commit'
This is so we can see near which file we fail when 'ostree commit'
falls over.
https://bugzilla.gnome.org/show_bug.cgi?id=705905
Stef Walter [Tue, 13 Aug 2013 12:46:38 +0000 (14:46 +0200)]
Add --verbose and -v arguments to ostree for details
Use the GLib debug mechanism to handle verbose output.
https://bugzilla.gnome.org/show_bug.cgi?id=705905
Stef Walter [Tue, 13 Aug 2013 12:13:04 +0000 (14:13 +0200)]
Better parsing for global ostree options
* Specifying global options after the command for a more natural:
# ostree commit --repo=/path/to/repo ...
* Support asking for --help without --repo
# ostree commit --help
* Support short form of -h
* Support specifying --repo without equals sign
# ostree --repo /path/to/repo commit ...
* Support global --help and -h
# ostree --help
* Ditto for ostree admin sub commands
* Removed some leaky code
https://bugzilla.gnome.org/show_bug.cgi?id=705903
Stef Walter [Tue, 13 Aug 2013 12:10:25 +0000 (14:10 +0200)]
Ignore some built files
https://bugzilla.gnome.org/show_bug.cgi?id=705900
Colin Walters [Wed, 7 Aug 2013 08:36:52 +0000 (10:36 +0200)]
Update libgsystem
With fd leak fix.
Colin Walters [Tue, 6 Aug 2013 21:58:29 +0000 (23:58 +0200)]
core: Clean up temporary file creation
Use new libgsystem API, don't expose the repo's tmpdir.
Colin Walters [Wed, 31 Jul 2013 07:48:31 +0000 (03:48 -0400)]
core: Finish making object path API private
This allows us to more easily change the internals later.
Colin Walters [Wed, 31 Jul 2013 00:56:33 +0000 (20:56 -0400)]
core: Reduce API surface for file access
This function was only used by internals, make it private.
Colin Walters [Wed, 31 Jul 2013 00:51:43 +0000 (20:51 -0400)]
TODO: Update
Colin Walters [Mon, 29 Jul 2013 22:44:41 +0000 (18:44 -0400)]
admin: Clean up builtin prototype
Pass through GCancellable, and just use GFile *sysroot, since that's
all OtAdminBuiltinOpts was.
Colin Walters [Sun, 28 Jul 2013 00:12:07 +0000 (20:12 -0400)]
main: Assume / for ls if no path given
Since it just is kind of a sane default.
Colin Walters [Sat, 27 Jul 2013 14:13:30 +0000 (10:13 -0400)]
core: Fix all introspection warnings
Still lots more docs to write and API to cleanup, but this is better.
Colin Walters [Fri, 26 Jul 2013 23:25:07 +0000 (19:25 -0400)]
Install a shared library
This required a fair bit of surgery because previously ostree.h
included otutil.h, but that's supposed to be a private library.
Colin Walters [Fri, 26 Jul 2013 18:48:15 +0000 (14:48 -0400)]
core: Work around libguestfs/FUSE issue with setuid binaries
For some reason, the setuid bits are being stripped. This
workaround is enough for now.
Colin Walters [Thu, 25 Jul 2013 22:52:06 +0000 (18:52 -0400)]
Add API to retrieve statistics from transactions, use it in commit
It's just interesting to see =) This is also kind of prototyping out
some more "structured" output.
Colin Walters [Thu, 25 Jul 2013 21:09:00 +0000 (17:09 -0400)]
lib: Use -export-symbol-regex
Just a small step towards being a better library.
Colin Walters [Thu, 25 Jul 2013 16:06:21 +0000 (12:06 -0400)]
fsck: Fix crash on repositories with < 10 objects
Colin Walters [Wed, 24 Jul 2013 22:20:19 +0000 (18:20 -0400)]
pull: Tweak status line again, do what git does
Almost, we don't have bytes transferred. And we lie about percentages
since as metadata comes in the total number of objects to fetch goes
up.
Colin Walters [Wed, 24 Jul 2013 18:46:53 +0000 (14:46 -0400)]
main: Fix previous commit to find repo in cwd over /ostree/repo
The tests when run inside gnome-ostree broke, and this is better
behavior.
Colin Walters [Wed, 24 Jul 2013 18:05:12 +0000 (14:05 -0400)]
main: Support using repo from current directory
$ cd repo
$ ostree ls foo /
...
Can be a lot more convenient than typing --repo=repo a lot.
Colin Walters [Wed, 24 Jul 2013 17:10:28 +0000 (13:10 -0400)]
TODO: More bits about commit objects
Colin Walters [Wed, 24 Jul 2013 16:59:27 +0000 (12:59 -0400)]
Drop support for related objects and metadata in commit objects
While the actual commit object format is presently the same, for a
number of reasons we'd like to change it fairly radically. Among
other things, we need to drop our a{sv} types in objects, to protect
against GVariant changing format.
Since now gnome-ostree now longer uses related objects, and nothing
ever used metadata, just drop them both.
Colin Walters [Wed, 24 Jul 2013 13:36:51 +0000 (09:36 -0400)]
main: Fix an uninitialized variable
Colin Walters [Tue, 23 Jul 2013 23:36:15 +0000 (19:36 -0400)]
pull: Always scan for commit object, even if ref is unchanged
If the admin encounters corruption and does:
$ ostree admin fsck --delete
We want them to be able to recover the objects easily from the
network; with this patch, they do:
$ ln -s dummyvalue /ostree/repo/transaction
$ ostree refs --delete remotename:branchname
$ ostree pull remotename
This patch avoids the need for the refs --delete; we might as well
force scan the commit, and with this patch we still print that it
changed.
Colin Walters [Tue, 23 Jul 2013 22:41:03 +0000 (18:41 -0400)]
main: ls: Behave like standard Unix "ls" with directories
Default to listing directory contents, add -d/--dironly to suppress
this.
Colin Walters [Tue, 23 Jul 2013 22:34:34 +0000 (18:34 -0400)]
main: Squash a compiler warning
Colin Walters [Tue, 23 Jul 2013 22:33:29 +0000 (18:33 -0400)]
main: Consistently use "opt_" as a prefix for option local variables
This makes it easier to find where we're using global state, which is
useful when refactoring code into libostree.
Colin Walters [Tue, 23 Jul 2013 22:24:52 +0000 (18:24 -0400)]
main: commit: Drop broken --parent option
No idea why we had this, it didn't do anything in the current code.
Noticed while doing other refactoring.
Colin Walters [Tue, 23 Jul 2013 22:16:54 +0000 (18:16 -0400)]
main: Drop log builtin
We may revive this later, but commits in their current form aren't
very useful for humans to read, so it doesn't make sense to have a
tool to show a history of useless stuff.
More interesting things are diffs between commits, object statistics,
etc.
Colin Walters [Tue, 23 Jul 2013 21:29:46 +0000 (17:29 -0400)]
main: pull: Reduce the status line length a bit
Drop "objects" where it's obvious [ to me anyways =) ], and drop the
bytes transferred since our math was off, and the HTTP status kind of
shows that.
Colin Walters [Tue, 23 Jul 2013 18:42:07 +0000 (14:42 -0400)]
TODO: Add an item about commit objets
Colin Walters [Tue, 23 Jul 2013 13:19:24 +0000 (09:19 -0400)]
admin: Add an "undeploy" command
Otherwise it's really easy to keep accumulating deployments. Also, we
may want to run this after rebooting, so we're back down to one
operating system.
Colin Walters [Thu, 18 Jul 2013 18:23:55 +0000 (14:23 -0400)]
core: Don't strip setuid bits when creating files
This ugly regression occurred because I overlooked the fact that our
chown() invocation would strip off setuid.
Colin Walters [Thu, 18 Jul 2013 18:23:12 +0000 (14:23 -0400)]
main: show: Don't strip mode, let's see the whole thing
Just useful for debugging.
Colin Walters [Thu, 18 Jul 2013 16:09:44 +0000 (12:09 -0400)]
main: Add --delete option to fsck
This is useful for clearing out all corrupted objects locally.
Colin Walters [Thu, 18 Jul 2013 13:05:58 +0000 (09:05 -0400)]
main: Clean up fsck code: honor --quiet, warn (but continue) on missing objects
When we make fsck --delete work again, it will be convenient to
continue on missing objects.
Colin Walters [Thu, 18 Jul 2013 03:00:03 +0000 (23:00 -0400)]
Update libgsystem
Colin Walters [Thu, 18 Jul 2013 01:10:04 +0000 (21:10 -0400)]
main: Make "ostree show" actually useful for files
We already have "ostree cat" for files, here we want to see a dump of
the metadata.
Colin Walters [Thu, 18 Jul 2013 00:20:50 +0000 (20:20 -0400)]
TODO: Add more
Colin Walters [Wed, 17 Jul 2013 23:47:40 +0000 (19:47 -0400)]
tests: Add a test for link checkout speedup
Colin Walters [Wed, 17 Jul 2013 23:20:22 +0000 (19:20 -0400)]
diff: Add --stats option
This a neat way to see how many objects are shared between two
commits.
Colin Walters [Wed, 17 Jul 2013 20:16:45 +0000 (16:16 -0400)]
main: Add --link-checkout-speedup option to commit
And disable devino scan by default. For the gnome-ostree build case,
our commits are from "make install DESTDIR=", so they won't be
hardlinks into the repo. In that case, particularly as the repository
size grows, the cost of building up the devino -> checksum mapping
becomes a problem.
The compose step will use this option though.
Jiro Matsuzawa [Wed, 17 Jul 2013 03:20:31 +0000 (12:20 +0900)]
doc: Fix a typo
https://bugzilla.gnome.org/show_bug.cgi?id=704365
Colin Walters [Wed, 17 Jul 2013 00:41:47 +0000 (20:41 -0400)]
Release 2013.4
Colin Walters [Tue, 16 Jul 2013 22:40:11 +0000 (18:40 -0400)]
doc: A long overdue manpage update
Ok, just a little one, but still. I actually updated documentation!
Colin Walters [Tue, 16 Jul 2013 22:13:39 +0000 (18:13 -0400)]
main: Move GCancellable to toplevel
This is just cleaner, and makes the builtins slightly more of a
library, which in turn makes it easier to turn them into real API.
Colin Walters [Tue, 16 Jul 2013 14:01:58 +0000 (10:01 -0400)]
prune: Don't fail if a to-be-pruned object doesn't exist
Previously, if a prune was interrupted, further attempts would fail.
It's just better to silently continue here.
Colin Walters [Tue, 16 Jul 2013 13:35:44 +0000 (09:35 -0400)]
Drop support for fifos and devices
While the first was useful way back in the day when we were importing
Debian bits and /sbin/init was expecting to find /dev/.initctl as a
named pipe, that's no longer an issue with systemd since it uses
dynamic Unix sockets.
Likewise, character and block devices in /dev are now dynamically
created by the devtmpfs from the kernel.
Less complexity and code here if we just support directories, regular
files, and symbolic links.
Colin Walters [Tue, 16 Jul 2013 13:22:42 +0000 (09:22 -0400)]
core: Drop a leftover unnecessary #define _GNU_SOURCE
We now use AC_USE_SYSTEM_EXTENSIONS.
Colin Walters [Tue, 16 Jul 2013 12:50:31 +0000 (08:50 -0400)]
core: Use gsystem API for generating temporary names
Drops duplicated code.