Joey Hess [Wed, 9 Aug 2023 17:30:43 +0000 (13:30 -0400)]
Add git-annex remote refs that are not newer to the merged refs list
Significant startup speed increase by avoiding repeatedly checking if some
remote git-annex branch refs need to be merged when it is not newer.
One way this could happen is when there are 2 remotes that are themselves
connected. The git-annex branch on the first remote gets updated. Then the
second remote pulls from the first, and merges in its git-annex branch.
Then the local repo pulls from the second remote, and merges its git-annex
branch. At this point, a pull from the first remote will get a git-annex
branch that is not newer, but is not on the merged refs list.
In my big repo, git-annex startup time dropped from 4 seconds to 0.1 seconds.
There were 5 to 10 such remote refs out of 18 remotes.
Joey Hess [Wed, 9 Aug 2023 16:43:48 +0000 (12:43 -0400)]
info: Added --dead-repositories option
I considered a more wide-ranging config option to make other commands
also show dead repositories. But it would be difficult to implement that
because Remote.keyLocations is used to get locations, filtering out dead
repos, and commands like get then try to use those locations. So a config
setting would make dead repos sometimes be acted on by commands.
Joey Hess [Wed, 2 Aug 2023 16:48:12 +0000 (12:48 -0400)]
windows back to lts-18.13 temporarily
I can't seem to get stack to resolve dependencies with Win32-2.13.4.0,
no matter what I try. Why it blows up, I don't know.
And allow-newer: true actually causes it to downgrade Win32 to the one
version that won't build. Unbelivable that allows downgrades.
So just gonna have to wait for that to get into stackage nightly, and
then stack.yaml can be updated to use that, and the changes in this
commit reverted.
Joey Hess [Wed, 2 Aug 2023 15:50:12 +0000 (11:50 -0400)]
fix stack build on windows
For whatever reason, putting Win32-2.13.4.0 in stack.yaml results in
stack blowing up with many unrelated dependency problems.
But making git-annex depend on that version lets stack resolve deps.
Joey Hess [Wed, 2 Aug 2023 14:08:00 +0000 (10:08 -0400)]
avoid cpp failure on windows
Seems that while the module is not imported by anything on windows, it
still gets cpped, and MIN_VERSION_unix is not defined so it failed to
preprocess.
Joey Hess [Wed, 2 Aug 2023 00:17:29 +0000 (20:17 -0400)]
fix hang when built with unix-2.8
git-annex test hang when running git-annex add in an adjusted unlocked
branch. I couldn't seem to reproduce the hang outside the test suite.
Seems that the code added in 26a9ea12d11f85b43987d863449c790f7320d9d1
was buggy, and as that commit was made without testing it, building with
unix-2.8 exposed the bug.
I don't fully understand the bug, which involves fdToHandle
and then closing the fd, vs closing the handle. May somehow involve
laziness or forcing around the S.hGet? Using hClose solved it
in any case.
(Also eliminated checkcontentfollowssymlinks to fix a build warning
when it's not used.)
Joey Hess [Tue, 1 Aug 2023 22:53:26 +0000 (18:53 -0400)]
stack.yaml: Update to build with ghc-9.6.2 and aws-0.24
This enables some new features that need the new aws.
Use http-client-restricted-0.1.0 because it uses the crypton side of the
cryptonite/crypton fork, which seems to be needed for ghc-9.6.2.
Dependency on connection removed because of the cryptonite/crypton fork.
This avoids needing a build flag. It was only used to throw a typed
exception in Utility.Url, which nothing depended on.
Used a fork of bloomfilter because it's not being maintained and no longer
builds as-of this ghc version. (I have been trying to contact its
maintainer about it, and emailed him today suggesting I take over the
package.)
Joey Hess [Tue, 1 Aug 2023 21:47:30 +0000 (17:47 -0400)]
fix build with unix-2.8.0
It made UserInfo into a pattern to discourage manually constructing
them, so just to use UserInfo in a type signature of a function that
consumes them, have to import the new ByteString module.
Joey Hess [Tue, 1 Aug 2023 19:08:56 +0000 (15:08 -0400)]
Setup.hs: Stop installing man pages, desktop files, and the git-annex-shell and git-remote-tor-annex symlinks
Anything still relying on that, eg via cabal v1-install will need to
change to using make install-home. Which was added back in 2019 in 6491b626140a9e5cc8e99fff4a1434d7c0e9bc39 because cabal new-build
(now the default) already didn't use Setup in a way that let its
installation of those things work.
Notably this means Setup does not need to depend on unix-compat, which is
useful because in 0.7 it removed System.PosixCompat.User, which Setup
needed to determine where to install the desktop files. See
https://github.com/haskell-pkg-janitors/unix-compat/issues/3
Using == and != proved too hard to read, so went with [TRUE] and [FALSE]
after the term. I would kind of liked to have used emojis for a green
check and red X, but probably that is too fancy to be a good idea.
Make --explain output be inside [ ] with whitespace around them, to
avoid it ending with eg "[FALSE]]" and to make it easier to visually
find the start of it.
Also git-annex matchexpression --explain explains why its input
expression matches or fails to match.
When there is no limit, avoid explaining why the lack of limit
matches. This is also done when no preferred content expression is set,
although in a few cases it defaults to a non-empty matcher, which will
be explained.
Currently it only displays explanations of options like --in and --copies.
In the future, it should explain preferred content expression evaluation
and other decisions.
The explanations of a few things could be better. In particular,
"standard" will just appear as-is (or as "!standard" if it doesn't
match), rather than explaining why the standard preferred content expression
for the group matches or not.
Currently as implemented, it goes to stdout, and so commands like
git-annex find that have custom output will not display --explain
information. Perhaps that should change, dunno.
explain is a kind of debug message, but not formatted in the same way.
So it makes sense to reuse the debug machinery for it, since that is
already quite optimised.
This is to be used to explain why something did or didn't match.
Note that this reimplements match in terms of matchMrun.
Implementing match' as a Writer and matchMrun' as a MonadWriter
resulted in nearly identical implementations, which collapsed into the
same thing thanks to Writer being WriterT Identity.
MAnd and MOr implement short circuiting. So an expression
like "not (foo and bar)" will be explained as [MatchedNot, MatchOperation "foo"]
when foo does not match; whether bar matches is irrelevant. Similarly
"foo or bar" will be explained as [MatchedOperation "foo"] when foo
matches. It seems like that will keep the explanations more
understandable. But also, matchMrun already did short circuiting, and it
could be considerably more work to check if bar matches in these cases.
Note that the type signature of matchMrun changed, but it was
over-generic before.
Note that these changes are licensed under the AGPL. Changed module
license accordingly.
deal with Amazon S3 breaking change for public=yes
* S3: Amazon S3 buckets created after April 2023 do not support ACLs,
so public=yes cannot be used with them. Existing buckets configured
with public=yes will keep working.
* S3: Allow setting publicurl=yes without public=yes, to support
buckets that are configured with a Bucket Policy that allows public
access.
projects / git-annex.git / log