Colin Walters [Tue, 3 Oct 2017 01:36:10 +0000 (21:36 -0400)]
lib/pull: Change fetcher to return O_TMPFILE
A lot of the libostree code is honestly too complex for its
own good (this is mostly my fault). The way we do HTTP writes
is still one of those. The way the fetcher writes tempfiles,
then reads them back in is definitely one of those.
Now that we've dropped the "partial object" bits in:
https://github.com/ostreedev/ostree/pull/1176 i.e. commit
https://github.com/ostreedev/ostree/commit/0488b4870e80ef575d8b0edf6f2a9e5ad54bf4df
we can simplify things a lot more by having the fetcher
return an `O_TMPFILE` rather than a filename.
For trusted archive mirroring, we need to enable linking
in the tmpfiles directly.
Otherwise for at least content objects they're compressed, so we couldn't link
them in. For metadata, we need to do similar logic to what we have around
`mmap()` to only grab a tmpfile if the size is large enough.
Colin Walters [Wed, 4 Oct 2017 19:06:31 +0000 (15:06 -0400)]
Deduplicate and fix up our use of mmap()
Buried in this large patch is a logical fix:
```
- if (!map)
- return glnx_throw_errno_prefix (error, "mmap");
+ if (map == (void*)-1)
+ return glnx_null_throw_errno_prefix (error, "mmap");
```
Which would have helped me debug another patch I was working
on. But it turns out that actually correctly checking for
errors from `mmap()` triggers lots of other bugs - basically
because we sometimes handle zero-length variants (in detached
metadata). When we start actually returning errors due to
this, things break. (It wasn't a problem in practice before
because most things looked at the zero size, not the data).
Anyways there's a bigger picture issue here - a while ago
we made a fix to only use `mmap()` for reading metadata from disk
only if it was large enough (i.e. `>16k`). But that didn't
help various other paths in the pull code and others that were
directly doing the `mmap()`.
Fix this by having a proper low level fs helper that does "read all data from
fd+offset into GBytes", which handles the size check. Then the `GVariant` bits
are just a clean layer on top of this. (At the small cost of an additional
allocation)
Side note: I had to remind myself, but the reason we can't just use
`GMappedFile` here is it doesn't support passing an offset into `mmap()`.
Jonathan Lebon [Wed, 4 Oct 2017 14:41:15 +0000 (14:41 +0000)]
lib/commit: minor coverity fix
Appease Coverity by using the same condition for both the ternary check
and the if-condition later on. It should be smart enough to figure out
that `dir_enum == NULL` implies that `dfd_iter != NULL` from the
assertion at the top of the function.
Jonathan Lebon [Tue, 3 Oct 2017 21:34:43 +0000 (21:34 +0000)]
ci: really turn on warnings
We didn't have `-Wall` in our `CFLAGS`. It's normally injected by
`configure.ac`, but because we *did* have `-Werror`, it was skipped.
Now, we just turn it on unconditionally directly in `build.sh`.
Jonathan Lebon [Tue, 3 Oct 2017 19:00:32 +0000 (19:00 +0000)]
app/main: use HAVE_LIBCURL_OR_LIBSOUP for pull CLI
We want `pull` to be included as long as we have at least either
`libcurl` or `libsoup` to back it. Of course, this is a moot point for
now since `libsoup` is currently a build requirement.
Philip Withnall [Tue, 3 Oct 2017 14:45:34 +0000 (15:45 +0100)]
lib/repo-commit: Import detached metadata even if hardlink exists
Spotted while reading through the code, it looks like the
copy_detached_metadata() call is accidentally omitted if a hardlink
already exists for the .commit object.
Signed-off-by: Philip Withnall <withnall@endlessm.com> Closes: #1242
Approved by: cgwalters
Matthew Leeds [Tue, 3 Oct 2017 05:35:17 +0000 (22:35 -0700)]
lib/gpg: Print debug info when reading GPG keys
This commit adds debug output whenever libostree reads GPG keys, which
can come from different locations in the file system. This is especially
helpful in debugging "GPG signatures found, but none are in trusted
keyring" errors, which in my case was caused by OSTree looking in
/usr/local/share/ostree/trusted.gpg.d/ rather than
/usr/share/ostree/trusted.gpg.d/.
Colin Walters [Mon, 2 Oct 2017 19:36:47 +0000 (15:36 -0400)]
lib/pull: Minor cleanup to metadata scanning function, add docs
I'm regretting a bit having the `guint8*csum` variant of checksums
except for the serialized form. Once we start doing processing
it's easier to just have it remain hex.
Do an on-stack conversion for the metadata scanning function; this
drops a malloc and also just looks nicer.
Also add some long-awaited function comments to the two.
Philip Withnall [Mon, 2 Oct 2017 17:04:37 +0000 (18:04 +0100)]
lib/bloom: Add some missing preconditions on n_bytes
These shouldn’t change the bloom filter’s behaviour at all, but make it
a bit more obvious what the programmatical limitations are on the sizes
it can deal with.
In reality, those sizes should never be reached because they won’t fit
in a DNS-SD record.
Signed-off-by: Philip Withnall <withnall@endlessm.com> Closes: #1239
Approved by: cgwalters
Colin Walters [Mon, 2 Oct 2017 15:24:05 +0000 (11:24 -0400)]
lib/pull: Fix regression with pull-local for nonexistent refs
I was reading the pull code for the last release, and spotted
a bug in commit f923c2e1eaebe0c781f07d34ae1a03f94357bccd - in
the case where the ref doesn't exist, we don't set an error,
tripping an assertion in the main code.
The previous code wanted the ref to always exist, so just flip back the boolean
for "ignore noent". I moved the `g_strchomp()` just into the HTTP path - if a
local repo is corrupted in this way it's something to fix in that repo.
Philip Withnall [Mon, 11 Sep 2017 13:42:52 +0000 (14:42 +0100)]
ostree/summary: Generate an ostree-metadata ref when updating summary
This is the new way of publishing repository metadata, rather than as
additional-metadata in the summary file. The use of an ostree-metadata
ref means that the metadata from multiple upstream collections is not
conflated when doing P2P mirroring of many repositories.
The new ref is only generated if the repository has a collection ID set.
The old summary file continues to be generated for backwards
compatibility (and because it continues to be the canonical ref →
checksum map for the repository).
The new code is only used if configured with --enable-experimental-api.
Includes unit tests.
Signed-off-by: Philip Withnall <withnall@endlessm.com> Closes: #1158
Approved by: cgwalters
Philip Withnall [Sat, 30 Sep 2017 23:13:17 +0000 (00:13 +0100)]
lib/bloom: Fix a -Wconversion warning in OstreeBloom
Compiling with -Wconversion warns on this line, as the conversion from
guint64 to guint8 is implicit (but safe: there is no bug here, since the
implicit cast is applied after the modulus arithmetic).
Make the cast explicit to silence -Wconversion.
Signed-off-by: Philip Withnall <withnall@endlessm.com> Closes: #1231
Approved by: cgwalters
Jonathan Lebon [Thu, 28 Sep 2017 18:57:33 +0000 (18:57 +0000)]
tests: check for relabeling rather than overlay
Instead of checking for overlayfs, let's explicitly check for our
ability to relabel files since we now have a `libtest` function to do
this. Also port that logic to `libostreetest`.
Note that overlayfs *does* allow manipulating user xattrs. So ideally,
we should break down `OSTREE_NO_XATTRS` further to distinguish between
tests that use bare repos from other modes.
We check the current directory instead of `/` so that developers can
just point `TEST_TMPDIR` to a non-overlayfs mount point when hacking
from a container.
Jonathan Lebon [Thu, 28 Sep 2017 19:08:06 +0000 (19:08 +0000)]
lib/commit: don't query devino cache for modified files
We can't use the cache if the file we want to commit has been modified
by the client through the file info or xattr modifiers. We would
prematurely look into the cache in `write_dfd_iter_to_mtree_internal`,
regardless of whether any filtering applied.
We remove that path there, and make sure that we only use the cache if
there were no modifications. We rename the `get_modified_xattrs` to
`get_final_xattrs` to reflect the fact that the xattrs may not be
modified.
One tricky bit that took me some time was that we now need to store the
st_dev & st_ino values in the GFileInfo because the cache lookup relies
on it. I'm guessing we regressed on this at some point.
This patch does slightly change the semantics of the xattr callback.
Previously, returning NULL from the cb meant no xattrs at all. Now, it
means to default to the on-disk state. We might want to consider putting
that behind a flag instead. Though it seems like a more useful behaviour
so that callers can only override the files they want to without losing
original on-disk state (and if they don't want that, just return an
empty GVariant).
Colin Walters [Thu, 28 Sep 2017 19:41:02 +0000 (15:41 -0400)]
libarchive: Fix assertion error if used with CANONICAL_PERMISSIONS
I was trying to do a change for rpm-ostree to use
`OSTREE_REPO_COMMIT_MODIFIER_FLAGS_CANONICAL_PERMISSIONS`
for container builds with `bare-user-only,` but hit an assertion here
ultimtely because we weren't setting `standard::type`.
Rather than hand-rolling `GFileInfo` creation, use the stat buffer conversion
code which is more robust and used in multiple places already.
Philip Withnall [Mon, 18 Sep 2017 14:43:41 +0000 (15:43 +0100)]
tests: Add a manual integration test for OstreeRepoFinderMount
Test an end-to-end flow of pulling refs from an online
repository → local OS repository → create a USB stick of
them → pull to a local OS repository on another machine.
This is a manual test, as it requires a throwaway USB stick which the
test can format as ext4 or vfat to test the flow works with both file
systems.
Run it as:
MOUNT_INTEGRATION_DEV=/dev/sdb1 make check \
TESTS=tests/test-repo-finder-mount-integration.sh
Signed-off-by: Philip Withnall <withnall@endlessm.com> Closes: #1203
Approved by: cgwalters
Colin Walters [Tue, 26 Sep 2017 21:21:59 +0000 (17:21 -0400)]
lib/pull: Don't request deltas for unchanged commits
I noticed this while debugging why I was seeing "2 metadata objects" fetched for
a different PR. I knew 1 was detached meta, but the other turned out to be this.
There's no reason to request a delta if the ref is unchanged.
Colin Walters [Wed, 27 Sep 2017 00:51:24 +0000 (20:51 -0400)]
tests/commit-sign: Update a bit to more modern style
I was trying to debug this while working on another PR. Add `echo ok` lines and
also use `assert_file_has_content` more directly; there's no reason to use an
intermediate `grep` since if it fails it won't print the original file.
Philip Withnall [Thu, 21 Sep 2017 15:08:26 +0000 (16:08 +0100)]
lib/repo-pull: Fix remote names in refspecs from non-mirror P2P pulls
Propagate the refspec_name from the OstreeRemote returned by an
OstreeRepoFinder through to the set_ref() call.
This changes ostree_repo_pull_with_options() to accept the
previously-disallowed combination of passing override-remote-name in
options and also setting a remote name in remote_name_or_baseurl.
ostree_repo_pull_with_options() will continue to pull using the remote
config named in remote_name_or_baseurl as before; but will now use the
remote name from override-remote-name when it’s setting the refs at the
end of the pull. This is consistent with the documentation for
override-remote-name.
Signed-off-by: Philip Withnall <withnall@endlessm.com> Closes: #1202
Approved by: cgwalters
Philip Withnall [Thu, 21 Sep 2017 15:04:11 +0000 (16:04 +0100)]
lib/remote: Store name of remote providing keyring for dynamic remotes
When pulling from a dynamic (peer to peer) remote, the remote’s name is
set to a unique, generated string which doesn’t exist in repo/config. If
doing a non-mirror pull, however, we don’t want to use this name in the
refspecs for newly created or updated refs — we want to use the name of
the remote which provided the keyring for the pull (this will be a
remote from repo/config whose collection ID matches that being used for
the peer to peer pull).
Store both names in OstreeRemote. The name to use for refspecs is stored
as refspec_name, and is typically NULL unless it differs from name.
Signed-off-by: Philip Withnall <withnall@endlessm.com> Closes: #1202
Approved by: cgwalters
Philip Withnall [Thu, 21 Sep 2017 15:00:05 +0000 (16:00 +0100)]
lib/repo: Change resolve_keyring_for_collection() to return a remote
Instead of returning just the keyring filename, return the entire
OstreeRemote, which has the keyring filename as one of its members. This
will simplify some upcoming changes, and allows slightly improved debug
logging.
Signed-off-by: Philip Withnall <withnall@endlessm.com> Closes: #1202
Approved by: cgwalters
Philip Withnall [Thu, 21 Sep 2017 15:01:44 +0000 (16:01 +0100)]
lib/repo-pull: Fix a potential minor leak
If override-remote-name is specified in the options to
ostree_repo_pull_with_options(), but the remote_name_or_baseurl argument
is also set to a remote name, the override-remote-name would be leaked.
Note that this is currently an invalid configuration, so this leak is
basically never hit.
Signed-off-by: Philip Withnall <withnall@endlessm.com> Closes: #1202
Approved by: cgwalters
Keep track of how many objects we imported, and print that for `ostree
pull-local` (also do this even if noninteractive, like we did for `pull`).
In implementing this at first I used separate variables for import
from repo vs import from localcache, but that broke some of the
tests that checked those values.
It's easier to just merge them; we know from looking at whether or not
`remote_repo_local` is set whether or not we were doing a "HTTP pull with
localcache" versus a true `pull-local` and can use that when rendering status.
Philip Withnall [Tue, 26 Sep 2017 14:17:47 +0000 (15:17 +0100)]
lib/repo-pull: Use resolve() instead of list() in fetch_ref_contents()
This is more efficient in the non-collection case; in the collection
case, the implementation of ostree_repo_resolve_collection_ref() needs
to be rewritten to improve efficiency.
Signed-off-by: Philip Withnall <withnall@endlessm.com> Closes: #1182
Approved by: cgwalters
Philip Withnall [Tue, 26 Sep 2017 14:15:57 +0000 (15:15 +0100)]
lib/repo-refs: Add first version of ostree_repo_resolve_collection_ref()
This is a parallel for ostree_repo_resolve_rev_ext() which works on
collection–refs. At the moment, the implementation is simple and uses
ostree_repo_list_collection_refs(). In future, it could be rewritten to
check the checksum directly rather than enumerating all
potentially-relevant checksums.
Signed-off-by: Philip Withnall <withnall@endlessm.com> Closes: #1182
Approved by: cgwalters
Philip Withnall [Mon, 18 Sep 2017 13:01:21 +0000 (14:01 +0100)]
src/pull: Support local pulls for collection–refs
Previously, collection–refs could only be pulled from a repository if it
had a summary file (which listed them). There was no way to pull from a
local repository which doesn’t have a summary file, and where the refs
were stored as refs/remotes/$remote/$ref, with a config section linking
that $remote to the queried collection ID.
Fix that by explicitly supporting pull_data->remote_repo_local in
fetch_ref_contents().
Signed-off-by: Philip Withnall <withnall@endlessm.com> Closes: #1182
Approved by: cgwalters
Colin Walters [Tue, 26 Sep 2017 16:41:07 +0000 (12:41 -0400)]
tests,ci: Move "test-basic" (bare mode) to installed test
Our CI uses default Docker, which has SELinux labeling but is rather
evil in returning `EOPNOTSUPP` to any attempts to set `security.selinux`,
even if to the same value.
The previous fire 🔥 for this was: https://github.com/ostreedev/ostree/pull/759
The `bare` repo mode really only makes sense as uid 0, so our installed
test framework is a good match for this. However, the unit tests *do*
work in a privileged container even as non-root, and *also* should
work on SELinux-disabled systems. So let's teach the test framework
how to skip in those situations.
I tested this both in a priv container (my default builder) and an unpriv
container (like our CI).
At the same time, start executing the `test-basic.sh` from an installed test,
so we get better coverage than before.
This is just the start - all of the sysroot tests really need the
same treatment.
Colin Walters [Sat, 23 Sep 2017 14:23:47 +0000 (10:23 -0400)]
lib/pull: Default checksum for archive mirror, add TRUSTED_HTTP flag
I now think commit fab1e113db558cb7d6754e243919558df92d4864 was a mistake;
because it breaks the mental model that at least I'd built up that "local repos
don't have checksums verified, HTTP does".
For example, a problem with this is (with that mental model in place) it's easy
for people who set up mirrors like this to then do local pulls, and at that
point we've done a deployment with no checksum verification.
Further, since then we did PR #671 AKA commit 3d38f03 which is really most of
the speed hit.
So let's switch the default even for this case to doing checksum verification,
and add `ostree pull --http-trusted`. People who are in situations where they
know they want this can find it and turn it on.
Colin Walters [Sat, 23 Sep 2017 13:59:53 +0000 (09:59 -0400)]
lib/pull: Refactor to use OstreeRepoImportFlags internally
Rather than carrying two booleans, just convert `OstreeRepoPullFlags`
into `OstreeRepoImportFlags`. This allows us to drop an internal
wrapper function and just directly call `_ostree_repo_import_object()`.
This though reveals that our mirroring import path doesn't check the
`OSTREE_REPO_PULL_FLAGS_UNTRUSTED` flag...it probably should.
Colin Walters [Sat, 23 Sep 2017 13:44:49 +0000 (09:44 -0400)]
lib/pull: Minor refactoring of metadata scanning function
Make the "local repo" processing conditional the same as the "localcache" bits;
this is really just a de-indent. Also add some comments. Prep for further work.
Jonathan Lebon [Wed, 13 Sep 2017 16:17:17 +0000 (16:17 +0000)]
lib/commit: fix using uninitialized var
Noticed this while reading the code. The `child` var hasn't been
initialized yet at the time we throw this error (and even then, it's
only conditionally initialized). To be nice, let's just always calculate
the child path and pass that along.
Also do some minor style porting to decl near use.
Jonathan Lebon [Wed, 13 Sep 2017 16:05:47 +0000 (16:05 +0000)]
lib/commit: add comments to explain dir commit path
Add a few comments for each of the central functions used for committing
data from a directory. Took me a bit to understand the relationship
between those functions.
Colin Walters [Wed, 20 Sep 2017 03:09:11 +0000 (23:09 -0400)]
lib/commit: Add a copy fastpath for imports
This fixes up the last of the embarassing bits I saw from
the stack trace in:
https://github.com/ostreedev/ostree/issues/1184
We had a hardlink fast path, but that doesn't apply across
devices, which occurs in two notable cases:
- Installer ISO with local repo
- Tools like pungi that copy the repo to a local snapshot
Obviously there are a lot of subtleties here around things like the
bare-user-only conversions as well as exactly what data we copy. I think to get
better test coverage we may want to add `pull-local --no-hardlink` or so.
Jussi Laako [Tue, 13 Jun 2017 13:56:21 +0000 (16:56 +0300)]
Add --with-crypto=gnutls
Introduce support for GnuTLS for computing cryptograpic
hashes, similar to the OpenSSL backend. A reason to do
this is some distributors want to avoid GPLv3, and GPG
pulls that in.
A possible extension of using GnuTLS would be replacing the GPG signing
with `PKCS#7` signatures and `X.509` keys.
We also support `--with-crypto=openssl`, which has the same effect
as `--with-openssl`, and continues to be supported.
Changes by Colin Walters <walters@verbum.org>:
- Drop libgcrypt option for now
- Unify buildsystem on --with-crypto
Jonathan Lebon [Wed, 20 Sep 2017 17:07:54 +0000 (17:07 +0000)]
tests/libtest: check that we have setfattr
We use `setfattr` to determine whether the filesystem we're on supports
xattrs, but we need to check that `setfattr` itself is available. We
just make it a hard requirement but only if trying to run tests that ask
about xattr support.
Colin Walters [Thu, 21 Sep 2017 19:26:06 +0000 (15:26 -0400)]
tree-wide: Remove Emacs modelines
We added a `.dir-locals.el` in commit: 9a77017d87b74c5e2895cdd64ad098018929403f
There's no need to have it per-file, with that people might think
to add other editors, which is the wrong direction.
Philip Withnall [Thu, 21 Sep 2017 19:22:09 +0000 (20:22 +0100)]
lib/repo: Add ostree_repo_hash() and tests
Add a hash function for OstreeRepo instances, which relies on the repo
being open, and hence being able to hash the device and inode of its
root directory.
Add unit tests for this and ostree_repo_equal().
Signed-off-by: Philip Withnall <withnall@endlessm.com>
https://github.com/ostreedev/ostree/issues/1191
Colin Walters [Thu, 21 Sep 2017 18:50:54 +0000 (14:50 -0400)]
lib/repo: Use correct name for tmpdir lockfile
Such an evil bug 🙈. I was just reading an strace trying to figure out what was
going on, and noticed we had the `XXXXXX` in the lockfile name. It was only
after that I realized that that this might *be* the cause of the skopeo issue.
This is another case where we definitely need more test coverage of things that
actually use the API multiple times in process; might look at dusting off the
work for the rpm-ostree test.
Colin Walters [Tue, 19 Sep 2017 19:19:05 +0000 (15:19 -0400)]
ci: Add an "all options disabled" context
We keep occasionally regressing this so let's start covering it now. I'm
intentionally not running the tests since that would likely entail a lot more
conditionalizing pain.
Colin Walters [Tue, 19 Sep 2017 19:07:30 +0000 (15:07 -0400)]
lib/commit: Don't try to call linkat() for import on distinct devices
While opening a repo we've recorded the device/inode for a while; use it to
avoid calling `linkat()` during object import if we know it's going to fail.
Colin Walters [Mon, 18 Sep 2017 19:46:03 +0000 (15:46 -0400)]
lib: Move bareuseronly verification into commit/core
Conceptually `ostree-repo-pull.c` should be be written using
just public APIs; we theoretically support building without HTTP
for people who just want to use the object store portion and
do their own fetching.
We have some nontrivial behaviors in the pull layer though; one
of those is the "bareuseronly" verification. Make a new internal
API that accepts flags, move it into `commit.c`. This
is prep for further work in changing object import to support
reflinks.
Jonathan Lebon [Wed, 20 Sep 2017 18:38:16 +0000 (18:38 +0000)]
rofiles-fuse: also pass mode for O_RDONLY
In the `O_RDONLY` case, we were calling `openat` without a mode
argument. However, it's perfectly legal (albeit unusual) to do
`open(O_RDONLY|O_CREAT)`. One such application that makes use of this is
`flock(1)`.
This was actually caught by `_FORTIFY_SOURCE=2`, and once we run
`rofiles-fuse` with `-f`, the message is clear:
Colin Walters [Mon, 18 Sep 2017 18:29:16 +0000 (14:29 -0400)]
lib/repo: Don't syncfs or fsync() dirs if fsync opt is disabled
There are use cases for not syncing at all; think build cache repos, etc. Let's
be consistent here and make sure if fsync is disabled we do no sync at all.
I chose this opportunity to add tests using the shiny new strace fault
injection. I can forsee using this for a lot more things, so I made
the support for detecting things generic.
Matthew Leeds [Tue, 19 Sep 2017 23:34:11 +0000 (16:34 -0700)]
lib/pull: Clarify use of unsigned summary support
Update the comments and remove an unneeded variable to make it clear
that the find_remotes_async() / pull_from_remotes_async() functions use
the unsigned summary support.
This is a follow-up of commit 8c148eb7e "lib/repo-finder: Emit
gpg-verify-summary=false in dynamic remote config".
Simon McVittie [Tue, 19 Sep 2017 17:37:58 +0000 (18:37 +0100)]
tests: Reset umask to 022 while creating test repository
In test-basic-root.sh we make assertions about the permissions
of files like baz/cow, which were created without an explicit chmod.
We can't do that unless we control the permissions.
For some reason the "debomatic" autobuilder used to do some Debian
archive rebuilds does the entire build including build-time tests
as uid 0 with umask 002, which broke those assertions. This seems
a weird thing to do, and I've opened a bug, but it also seems
reasonable to fix this test.
This also lets us remove a couple of existing workarounds for the
same issue.
(Why that library has a destructor I don't know offhand, can't find
it in the source in a quick look)
Anyways, global library destructors and worker threads continuing simply don't
mix. Let's wait for our outstanding operations before we exit. This is also a
good idea for projects using libostree as a shared library, as we don't want
worker threads outliving operations.
Colin Walters [Mon, 18 Sep 2017 19:46:03 +0000 (15:46 -0400)]
lib/repo: Minor cleanup to object import function
We have a lot of layers of abstraction here; let's fold in the `trusted`
conditional into the call, since that's all the public API we're using does
anyways.
Prep for a future patch around object copying during imports.
Simon McVittie [Tue, 19 Sep 2017 13:41:18 +0000 (14:41 +0100)]
tests: Explicitly unset LANGUAGE after setting LC_ALL
As a GNU extension, LANGUAGE takes precedence over LC_ALL for
gettext(3) whenever the locale is not C, causing tests that grep for
specific English strings to fail when run in non-English locales.
The upstream glibc proposal for C.UTF-8 would give C.UTF-8 the same
special case as C here, but the implementation in Debian does not
currently have this, so we have to unset LANGUAGE too.
Signed-off-by: Simon McVittie <smcv@collabora.com> Closes: #1188
Approved by: jlebon
Philip Withnall [Fri, 15 Sep 2017 14:59:32 +0000 (15:59 +0100)]
lib/repo-finder-mount: Change the schema for finding repos on volumes
See issue #1174 for the rationale behind this. In summary:
• It required two lists of collection–refs to be maintained: one in the
repository, and one pointing to the repository.
• It didn’t automatically work for live USBs of OSs based on OSTree
(where there’s always a repository at /ostree/repo).
• It was unnecessarily complex.
The new scheme allows a list of repositories to be searched, but without
needing a layer of indirection through their collection–refs. It adds
/ostree/repo and /.ostree/repo as well-known repository locations which
are always checked on a mounted volume (if they exist).
Update the unit tests accordingly.
Signed-off-by: Philip Withnall <withnall@endlessm.com>
https://github.com/ostreedev/ostree/issues/1174
Philip Withnall [Fri, 15 Sep 2017 14:26:22 +0000 (15:26 +0100)]
lib/repo: Add ostree_repo_equal() for comparing repos
This will compare their root directory inodes to see if they are the
same repository on disk. A convenience method for the users of the
public API who can’t access OstreeRepo.inode.
Signed-off-by: Philip Withnall <withnall@endlessm.com> Closes: #1179
Approved by: cgwalters
Colin Walters [Tue, 19 Sep 2017 13:46:17 +0000 (09:46 -0400)]
ci: Use distro-sync instead of upgrade
This should help ensure that if installing `glibc-devel` downgrades `glibc`
we'll run through our hackround for
<https://bugzilla.redhat.com/show_bug.cgi?id=1394862>.
Colin Walters [Fri, 1 Sep 2017 19:55:17 +0000 (15:55 -0400)]
bin/prune: Add --only-branch
In 5c940987e768523ef1411b65bcaad09fba6befef / #646 we
added `--retain-branch-depth`; this adds a symmetric
`--only-branch` for the case where a repo owner just
wants to prune a specific branch.
The implementation here is pretty straightforward; we
just walk all refs and inject the equivalent of
`--retain-branch-depth=$ref=-1` if they're *not* in
`--only-branch`.
Colin Walters [Tue, 12 Sep 2017 13:40:58 +0000 (09:40 -0400)]
Update libglnx
Update libglnx, which is mostly port the repo stagedir code
to the new tmpdir API. This turned out to require some
libglnx changes to support de-allocating the tmpdir ref while
still maintaining the on-disk dir.
Simon McVittie [Fri, 15 Sep 2017 15:05:48 +0000 (16:05 +0100)]
test-help.sh: Skip trivial-httpd if enabled
Because it runs a binary in ${libexecdir}, it will only work if
libostree was already installed, which makes the build-time test fail.
It also doesn't produce the output we expect: its usage mechanism
mentions "ostree-trivial-httpd", not "ostree trivial-httpd".
Signed-off-by: Simon McVittie <smcv@collabora.com> Closes: #1177
Approved by: jlebon
Colin Walters [Thu, 14 Sep 2017 19:10:54 +0000 (15:10 -0400)]
lib/pull: Drop partial fetch code from libsoup backend
Doing this in prep for libglnx tmpdir porting, but I think we should also do
this because the partial fetch code IMO was never fully baked; among other
things it was never integrated into the scheme we came up with for "boot id
sync" that we use for complete/staged objects.
There's a lot of complexity here that while we have some coverage for, I think
we need to refocus on the core functionality. The libcurl backend doesn't have
an equivalent to this today.
In particular for small objects, this is simply overly complex. The downside is
clearly for large objects like FAH's 61MB initramfs; not being able to resume
fetches of those is unfortunate.
In practice though, I think most people should be using deltas, and we need to
make sure deltas work for large objects anyways.
Further ultimately the peer-to-peer work should help a lot for people
with truly unreliable connections.
Colin Walters [Wed, 13 Sep 2017 20:22:18 +0000 (16:22 -0400)]
lib/checkout: Do UNION_FILES via atomic renameat()
I was looking at fixing an `rpm-ostree livefs` bug where we need to replace
`/usr/lib/passwd`. It's obviously bad if that temporarily disappears 😉. My plan
is to do a subpath checkout of just `/usr/lib/{passwd,group}`.
Make this atomic (i.e. file always exists) by changing the logic to create a
temporary link in repo/tmp, then rename() it into place.
A bonus here is we kill one of the very few (only?) non-error-cleanup i.e.
"non-linear" `goto`s in the ostree codebase.
projects / ostree.git / log