Files smaller than 8 KiB typically represent ~70% of the entries in
/gnu/store/.links but only contribute to ~4% of the space savings
afforded by deduplication.
Not considering these files for deduplication speeds up file insertion
in the store and, more importantly, leaves 'removeUnusedLinks' with
fewer entries to traverse, thereby speeding it up proportionally.
Partly fixes <https://issues.guix.gnu.org/24937>.
* config-daemon.ac: Remove symlink hard link check and CAN_LINK_SYMLINK
definition.
* guix/store/deduplication.scm (%deduplication-minimum-size): New
variable.
(deduplicate)[loop]: Do not recurse when FILE's size is below
%DEDUPLICATION-MINIMUM-SIZE.
(dump-port): New procedure.
(dump-file/deduplicate)[hash]: Turn into...
[dump-and-compute-hash]: ... this thunk.
Call 'deduplicate' only when SIZE is greater than
%DEDUPLICATION-MINIMUM-SIZE; otherwise call 'dump-port'.
* nix/libstore/gc.cc (LocalStore::removeUnusedLinks): Drop files where
st.st_size < deduplicationMinSize.
* nix/libstore/local-store.hh (deduplicationMinSize): New declaration.
* nix/libstore/optimise-store.cc (deduplicationMinSize): New variable.
(LocalStore::optimisePath_): Return when PATH is a symlink or smaller
than 'deduplicationMinSize'.
* tests/derivations.scm ("identical files are deduplicated"): Produce
files bigger than %DEDUPLICATION-MINIMUM-SIZE.
* tests/nar.scm ("restore-file-set with directories (signed, valid)"):
Likewise.
* tests/store-deduplication.scm ("deduplicate, below %deduplication-minimum-size"):
New test.
("deduplicate", "deduplicate, ENOSPC"): Produce files bigger than
%DEDUPLICATION-MINIMUM-SIZE.
* tests/store.scm ("substitute, deduplication"): Likewise.
This avoids having to traverse and re-read the files that we have just
restored, thereby reducing I/O.
* guix/serialization.scm (dump-file): New procedure.
(restore-file): Add #:dump-file parameter and honor it.
* guix/store/deduplication.scm (tee, dump-file/deduplicate): New
procedures.
* guix/nar.scm (restore-one-item): Pass #:dump-file to 'restore-file'.
(finalize-store-file): Pass #:deduplicate? #f to 'register-items'.
* tests/nar.scm <top level>: Call 'setenv' to set "NIX_STORE".
* guix/serialization.scm (restore-file): Set the permissions and mtime
of FILE.
* guix/nar.scm (finalize-store-file): Pass #:reset-timestamps? #f to
'register-items'.
* tests/nar.scm (rm-rf): Add 'chmod' calls to ensure files are writable.
("write-file + restore-file with symlinks"): Ensure every file in OUTPUT
passes 'canonical-file?'.
* tests/guix-archive.sh: Run "chmod -R +w" before "rm -rf".
* guix/serialization.scm (fold-archive): Call PROC with a
'directory-complete tag when done with a directory.
(restore-file): Handle it.
* guix/scripts/archive.scm (list-contents): Likewise.
* guix/scripts/challenge.scm (archive-contents): Likewise.
* tests/nar.scm ("write-file-tree + fold-archive"): Adjust accordingly.
Fixes <https://bugs.gnu.org/33603>.
Reported by Maxim Cournoyer <maxim.cournoyer@gmail.com>.
* guix/serialization.scm (port-conversion-strategy): New variable.
(restore-file): Parameterize it.
* tests/nar.scm ("restore-file with non-UTF8 locale"): New test.
Fixes <https://bugs.gnu.org/33361>.
* guix/store/deduplication.scm (replace-with-link): Call 'set-file-time'
and 'chmod' after 'rename-file'.
* tests/nar.scm ("restore-file-set with directories (signed, valid)"):
New test.
* guix/tests.scm (canonical-file?): New procedure.
* tests/nar.scm ("restore-file-set (signed, valid)"): Check that every
item of FILES matches 'canonical-file?'.
* guix/serialization.scm (write-contents-from-port): New procedure.
(write-contents): Write in terms of 'write-contents-from-port'.
(filter/sort-directory-entries, write-file-tree): New procedures.
(write-file): Rewrite in terms of 'write-file-tree'.
* tests/nar.scm ("write-file-tree + restore-file"): New test.
Slightly different buffering of custom binary output ports in Guile 2.1.7
would lead tests missing 'close-port' or 'force-output' to fail.
* tests/hash.scm ("open-sha256-port, hello"): Add call to 'force-output'
before call to 'get'.
* tests/nar.scm ("write-file puts file in C locale collation order"):
Call 'close-port' before calling 'get-hash'.
Before that '.log' files for scheme tests were fragmented and not
included in test-suite.log. This unifies the semantics of SRFI-64 API
with Automake test suite.
* build-aux/test-driver.scm: New file.
* Makefile.am (SCM_LOG_DRIVER, AM_SCM_LOG_DRIVER_FLAGS): New variables.
(SCM_LOG_COMPILER, AM_SCM_LOG_FLAGS): Delete variables.
(AM_TESTS_ENVIRONMENT): Set GUILE_AUTO_COMPILE to 0.
* test-env.in: Silence guix-daemon.
* doc/guix.texi (Running the Test Suite): Describe how to display the
detailed results. Bug reports require only 'test-suite.log' file.
* tests/base32.scm, tests/build-utils.scm, tests/builders.scm,
tests/challenge.scm, tests/cpan.scm, tests/cpio.scm, tests/cran.scm,
tests/cve.scm, tests/derivations.scm, tests/elpa.scm,
tests/file-systems.scm, tests/gem.scm, tests/gexp.scm,
tests/gnu-maintenance.scm, tests/grafts.scm, tests/graph.scm,
tests/gremlin.scm, tests/hackage.scm, tests/hash.scm,
tests/import-utils.scm, tests/lint.scm, tests/monads.scm, tests/nar.scm,
tests/packages.scm, tests/pk-crypto.scm, tests/pki.scm,
tests/profiles.scm, tests/publish.scm, tests/pypi.scm,
tests/records.scm, tests/scripts-build.scm, tests/scripts.scm,
tests/services.scm, tests/sets.scm, tests/size.scm, tests/snix.scm,
tests/store.scm, tests/substitute.scm, tests/syscalls.scm,
tests/system.scm, tests/ui.scm, tests/union.scm, tests/upstream.scm,
tests/utils.scm: Don't exit at the end of test groups.
* tests/containers.scm: Likewise. Use 'test-skip' instead of exiting
with error code 77.
Fixes <http://bugs.gnu.org/21280>.
Reported by Leo Famulari <leo@famulari.name>.
* tests/nar.scm (file-tree-equal?): Use #t as the initial result. Btrfs would
set nlink to 1 instead of 2 for an empty directory, thereby leading
'file-tree-equal?' to return #f.
Fixes <http://bugs.gnu.org/19756>.
Reported by <sleep_walker@suse.cz>.
* guix/serialization.scm (currently-restored-file): New variable.
(get-bytevector-n*): New procedure.
(read-int, read-long-long, read-string, read-latin1-string,
read-contents): Use it instead of 'get-bytevector-n'.
(restore-file): Parameterize 'currently-restored-file' and set it.
* tests/nar.scm ("restore-file with incomplete input"): New test.
Fixes <http://bugs.gnu.org/18935>.
Reported by Mark H Weaver <mhw@netris.org>.
Previously, if tests/nar.scm ran concurrently with tests/gexp.scm, the
store item "foo" containing the string "Hello, world!" could be
deleted (via 'delete-paths') from the tests/nar.scm daemon while the
tests/gexp.scm daemon would still consider it live.
* tests/nar.scm ("restore-file-set (missing signature)"):
Use (random-text) rather than "Hello, world!", to avoid concurrent
deletion of store item "foo" used in tests/gexp.scm.
* guix/nar.scm (write-file) <directory>: Pass 'string<?' as the second
argument to 'scandir'.
* tests/nar.scm ("write-file puts file in C locale collation order"):
New test.
* guix/nar.scm (write-file): Add case for type `symlink'.
(restore-file): Likewise.
* tests/nar.scm (random-file-size, make-file-tree, delete-file-tree,
with-file-tree, file-tree-equal?, make-random-bytevector,
populate-file): New procedures.
(%test-dir): New variable.
("write-file + restore-file"): Use `%test-dir' and `file-tree-equal?'.
("write-file + restore-file with symlinks"): New test.
* guix/nar.scm (&nar-error, &nar-read-error): New condition types.
(dump): New procedure.
(write-contents)[dump]: Remove. Use the one above instead.
(read-contents, write-file, restore-file): New procedures.
(%archive-version-1): New variable.
