Lots and lots of preparation led to this moment - when nothing
apparent changes for users! Woo!
But seriously, having the extra dependency is a minor annoyance, and
in the big picture I think the libgsystem idea was wrong - we need to
land things in GLib, and use git submodules for API-unstable or
Linux-specific sharing. For a lot of OSTree, the libgsystem `GFile*`
orientation was also wrong, we really want fd-relative.
Closes: #444
Approved by: jlebon
I hit an error with [CAHC](https://wiki.centos.org/SpecialInterestGroup/Atomic/Devel)
where we were doing time-based pruning. `ostree summary -u` started failing,
and it took me a bit to realize it was because we were pruning
even the tip of old branches, which I was not at all expecting,
and I don't think users will too.
Perhaps in the future we could add some sort of --prune-ref-tips or
something if people wanted it, but I doubt it.
Closes: #445
Approved by: jlebon
Since we already had a "recursive copy" implementation here, let's
reuse it rather than the libgsystem `gs_shutil_cp_a()`. Part of the
libglnx porting.
Closes: #428
Approved by: jlebon
It handles ownership of the `DIR*` for us more cleanly, and
is just a better API.
This is in preparation for further changes to this code to do SELinux
labeling while copying.
Closes: #428
Approved by: jlebon
I want to add `-Werror=int-conversion`, but it's only available in
newer GCC versions. So let's start autodetecting available compiler
flags.
Closes: #431
Approved by: giuseppe
I forgot to actually remove `config_file` in the previous
commit, the txn lock hasn't been used in a long time, and
for the uncompressed cache, everything uses the fd already.
Closes: #433
Approved by: giuseppe
I think we'd end up closing stdin...yuck. Need to investigate
having this be fatal, but CentOS 7 `gcc-4.8.5` doesn't understand
`-Werror=int-conversion`.
Closes: #430
Approved by: giuseppe
We have a better API now, drop use of the internal helper, which also
depended on libgsystem.
This required bumping libglnx to pull in a fix.
Closes: #429
Approved by: giuseppe
Since this is a new API, and adding booleans is the most likely thing
we'll do, let's stick some explicit padding for them in here now.
We could use the `unused_ints[]` but it'd be out of order, and this
will more clearly remind people about the padding. The efficiency hit
versus bitfields is annoying, but oh well, not a real world problem.
Closes: #427
Approved by: giuseppe
Since we're adding a new API, we have the opportunity to fix
the defaults. We expect clients to do a `syncfs()` or equivalent
on their own now, since it's way more efficient.
Flip the checkout fsync default to off.
Closes: #425
Approved by: giuseppe
This was the last use of libgsystem tmpfile APIs. The change here is
a bit uglier than it needs to be because we support creating the delta
in place as well as in the repo, so we needed to abstract over
locations.
Closes: #424
Approved by: giuseppe
Inlining this is uglier, but this code kind of sucks anyways =( We're
going to need to do some fd-relative porting in here at some point.
Closes: #424
Approved by: giuseppe
In one case, we already had relative fds and hence this was
nicer. Unfortunately the other areas got uglier. More fd-relative
porting to do later.
Closes: #424
Approved by: giuseppe
In general we want to support "idempotentcy" or "state
synchronization" across interruption. If a repo is only partially
created due to a crash or whatever, it's hard for a user to know that.
Let's just make `ostree_repo_create()` idempotent. Since all we're
doing is a set of `mkdirat()` invocations, it's quite simple.
This also involved porting to fd-relative, which IMO makes the
code a lot clearer.
Closes: #422
Approved by: 14rcole
In general this is even cleaner now, though it was better after I
extracted a helper function for the "write tempfile with contents"
bits that were shared between metadata and regular file codepaths.
Closes: #369
Approved by: jlebon
When reworking the ostree core [to use O_TMPFILE](https://github.com/ostreedev/ostree/pull/369),
I hit an issue in the way the untrusted delta codepath ends up trying
to re-open the file to checksum it. That's not possible with
`O_TMPFILE` since the fd (which we opened `O_WRONLY`) is the only
accessible reference to the content.
Fix this by changing the delta processing code to update a checksum as
we're doing writes, which is also faster, and ends up simplifying the
code as well.
What would be an even larger simplification here is if we e.g. used a
separate thread calling `write_object()` or something like that; the
main issue I see there is somehow bridging the fact that function
wants a `GInputStream*` but the delta code is generating stream of
writes.
Closes: #392
Approved by: jlebon
Otherwise we generate a static delta from the *previous* and get
confused. To make doubly sure, add `--require-static-deltas` to pull.
Closes: #416
Approved by: jlebon
This one is a bit subtle; we're generating a hash that contains
pointers to the strings we parsed, so we need to carefully track
ownership.
Closes: #410
Approved by: giuseppe
There are variants of this, but right now because `GTask` has its
own pool which isn't necessarily cleaned up on exit, we need
to suppress all of this.
Closes: #410
Approved by: giuseppe
This shows up as a leak in valgrind; the callee isn't sinking. In
general through the power of cleanup attributes we can do explicit
cleanup rather than relying on floating refs.
Closes: #410
Approved by: giuseppe
Right now our valgrind runs also end up valgrinding the
`trivial-httpd` code, so while it doesn't matter, let's fix this leak
anyways. We need to avoid calling `_exit()` since that won't run the
cleanup functions.
Closes: #410
Approved by: giuseppe
We don't presently support this, since the static delta code assumes
it can just `mmap()` file objects. We could at some point implement
this, but for now just skip executing deltas when doing
`archive -> archive` mirroring.
I noticed this when trying to mirror a repo in Jenkins in
[CAHC](https://wiki.centos.org/SpecialInterestGroup/Atomic/Devel).
Closes: #408
Approved by: giuseppe
Where we're referencing local files. It's what
ostree_repo_write_directory_to_mtree() is doing internally anyways,
and I consider the `GFile*` based APIs to be generally deprecated for
the most part.
Closes: #394
Approved by: giuseppe
When trying to switch ostree to `O_TMPFILE`, I hit the fact that
by default it uses mode `000`. It still works to write to the
open fd of course, but it *doesn't* work to set xattrs because
that code path for some reason in the kernel checks the mode bits.
This only broke for bare-user repos where we tried to set the xattr
before calling `fchmod()`, so just invert those two operations.
Closes: #391
Approved by: jlebon
This reverts commit 71301d1824.
I was confused by the fact that the non-matching content was
empty and forgot we already did print it.
Reported-by: smcv
Closes: #390
Approved by: jlebon
https://github.com/ostreedev/ostree/pull/372 caused these tests to
start failing when the host system is managed using ostree - since the
tests *do* replace the `ostree=` kernel argument.
Closes: #384
Approved by: cgwalters
Conceptually the session thread owns the session, so let's clear out
everything predictably there, rather than sometimes having it happen
on the main thread.
Also, this moves up clearing the pending/outstanding queues *before*
we unreference the session, since conceptually they need to reference
it as well.
Based on a patch from: Matthew Barnes <mbarnes@redhat.com>
Closes: #383
Approved by: mbarnes
I don't think this fixes the bug I was seeing, but it makes me more
comfortable to know we have a strong ref to the main context across
the thread lifetime, and we only unset the default right before
we go away.
If something in `thread_closure_unref()` used
`g_main_context_get_thread_default()` for example it'd be wrong
before.
Closes: #383
Approved by: mbarnes
Systems like pulp may want to keep retrying in a loop if the server
throws a (hopefully transient) 500, and we need test coverage of
handling these errors versus our existing 404 and 206 coverage.
Closes: #383
Approved by: mbarnes
We clean up the temporary directory on failure, which means it's hard
to know *why* a regex didn't match. Print it when we hit an error.
Closes: #383
Approved by: mbarnes
I often want to have "idempotent" systems that iterate to a known
state. If after generating a commit, the system is interrupted, I'd
like the next run to still generate a delta. But we don't want to
regenerate if one exists, hence this option.
Closes: #375
Approved by: jlebon
I've seen a few people hit this and wonder why checkouts are slow/take
space. Really, ensuring this happens is the *point* of OSTree.
Physical copies should be a last resort fallback for very unusual
situations (one of those is rpm-ostree checking out the db since
librpm doesn't know how to read from libostree).
Even I hit the fact that `/var` is a mountpoint disallowing hardlinks
with `/ostree` once and was confused. =)
Add this to the rofiles-fuse test case because it creates a mount
point.
Closes: #368
Approved by: jlebon
The program is called ro*files* and ostree creates physical
copies of directories, so changing them is fine.
I hit this when trying to do a copy checkout onto an rofiles-fuse
mount.
Closes: #368
Approved by: jlebon
Import `gs_file_enumerator_iterate()` for the next six months or
so...after RHEL 7.3 is released I'm strongly considering hard
requiring 2.46 or so.
Likely at some point we should figure out how to share more "glib
backport" code with NetworkManager at least.
Closes: #341
Approved by: jlebon
This was the last caller of libgsystem that isn't
`gs_file_get_path_cached()`. I think the use case ostree has where
the same code can be called via command line and via a shared library
*and* via a daemon is rather unusual, so let's just copy the code for
logging from libgsystem into here.
For example rpm-ostree hard depends on a daemon mode, so it'll just
use `sd_journal` directly.
Closes: #341
Approved by: jlebon
I was easily reproducing a hang on pulls with thousands of requests on
current git master. The initial symptom seemed to be that there are
multiple code paths where we don't invoke
`session_thread_process_pending_queue()`. We really need to do
that any time we remove something from the outstanding queue,
to ensure it gets filled again.
A further issue is that we were tying the lifecycle of the pending
object to the `GTask`, but the task could be unref'd from the main
thread (via a `GSource` on the main thread), and that introduced
threadsafety issues, because the hash table and other data suddenly
could be concurrently modified.
Both of these need to be fixed together. First, we introduce
`Arc<Pending>`, and ensure that both the main and worker threads hold
references.
Second, we ensure that we re-process the queue *immediately* whenever
a task is done, inside the worker thread, rather than doing it
incidentally via an unref. This architecture is quite similar to what
the outside pull code is doing.
Closes: #350
Approved by: jlebon
For some reason I'm really not inclined to debug right now, `libtool`
is eating the `-fsanitize=address` argument when linking
`libreaddir-rand.so`, which causes failures since we're trying to
`LD_PRELOAD`.
Closes: #351
Approved by: jlebon
It's unfortunate that in automake one has to explicitly include the globa
`$(AM_CFLAGS)` if one sets `CFLAGS`, and similarly for other variables.
I'm trying to use `-fsanitize=address`, and not including it was
causing linker failures.
We also weren't inheriting the global warnings etc., so I had to fix a
decl-after-statement.
Closes: #351
Approved by: jlebon
The previous code was subject to a divide by zero if less than a
second had passed. Rework it so we only do the divide if more than a
second has passed.
Closes: #349
Approved by: Mathnerd314
I find the "-z2" is really a long ago relic of the past when I changed
the format. We no longer have anything to do with the original
`archive`, so let's start allowing people to type `--mode=archive`
which just looks saner.
At some point later I'll update the docs too, but it'll be an annoying
transition period as we'll have to say "On older OSTree, use -z2" etc.
Closes: #346
Approved by: giuseppe
Just noticed this while reading some code, we didn't have many manual
`out: close()` bits left, this pushes us over the edge to autocleanup
almost everywhere.
Closes: #332
Approved by: jlebon
The recent memleak fixes motivated me to look at the bitrotted code to
run invocations of `ostree` in the test suite underneath valgrind.
There are a few things here. First, update suppressions file from
libhif, since I recently worked on it.
When running *uninstalled* as we now support, we need
`libtool --mode=execute` in the mix so it expands out to
the uninstalled binary and we don't valgrind the intermediate shell.
However, it's harder than that because we chdir into a tmpdir,
which defeats the libtool logic. AFAICS, the only fix for this
is to determine the realbin path before we chdir, and then unfortunately
we need to change every use of `ostree` to `${OSTREE}` =(
Then this immediately breaks for me on RHEL7 because my ancient
copy of `valgrind-3.10.0-16.el7.x86_64` is unaware of syscall 306, i.e.
`syncfs`.
But let's do this first before I dive into that.
Closes: #292
Approved by: krnowak
As the docs say, `g_regex_match()` still allocates a match even if it
returns `FALSE`. Using `g_autoptr` is just plain better.
Closes: #292
Approved by: krnowak
This kills another GSystem consumer...I think down the line I'd like
to do something like "detect whether file is > 1k if so, mmap,
otherwise just readall()" so we can use this helper in more places.
Closes: #319
Approved by: jlebon
We have a lot of "allow_noent" type wrapper functions since
a common pattern is to allow files to not exist, but still
throw cleanly on other issues.
This is another instance of that, and cleans up duplicated error
handling code.
Part of this is prep for moving away from `GFile` consumers.
Closes: #319
Approved by: jlebon
We have had in the past issues with running `ostree_repo_pull()`
multiple times in the same process, embarassingly enough. Nothing in
the current test suite covers this, so let's start.
Closes: #322
Approved by: jlebon
I was extending the C-based test suite to cover more, and noticed that
we weren't aborting if a command failed. That made the tests somewhat
less useful, so let's fix it.
Closes: #321
Approved by: jlebon
The "no atime" thing was mostly useful only before "relative atime"
updates landed. Users who care about performance will turn it off
entirely anyways.
Closes: #316
Approved by: jlebon
In addition to generic fd relative porting,
this is a necessary preparatory step for libglnx porting, because
when I tried to use `g_mapped_file_new` I hit an issue with
it using a different error domain from GIO.
Thankfully libglnx consistently uses the GIO error domain, and here
we're now using it for the `open()` call.
Closes: #317
Approved by: jlebon
This is similar to changes Krzesimir has been doing recently - we
really don't need the ergonomics of floating refs since we have
autocleanups.
We should continue to change most of our code to sink refs.
Specifically here it was pretty broken that the `_map()` API was
sinking but the other two weren't, and this broke some refactoring I
was trying to do later.
Closes: #317
Approved by: jlebon
I happened to have the github page open with my commit after it was
merged, and then noticed a bug.
We still need to throw if we hit a non-`ENOENT` error.
Closes: #315
Approved by: giuseppe
Colin Walters