* guix/ui.scm (show-what-to-build): When VERBOSITY is 1, add a newline
before the "would/will be downloaded" line, and wrap that message in
'highlight'.
This makes the test slightly less expensive.
* tests/store.scm ("add-text-to-store vs. delete-paths")
("add-to-store vs. delete-paths"): Delete and merge into...
("add-text-to-store/add-to-store vs. delete-paths"): ... this test.
In practice we're always using the same key pair,
/etc/guix/signing-key.{pub,sec}. Keeping them in cache allows us to
avoid redundant I/O and parsing when signing multiple store items in a
row.
* guix/scripts/authenticate.scm (load-key-pair): New procedure.
(sign-with-key): Remove 'key-file' parameter and add 'public-key' and
'secret-key'. Adjust accordingly.
(validate-signature): Add 'acl' parameter and pass it to
'authorized-key?'.
(guix-authenticate)[call-with-reply]: New procedure.
[with-reply]: New macro.
Call 'current-acl' upfront and cache its result. Add 'key-pairs' as an
argument to 'loop' and use it as a cache of key pairs.
Previously, we'd spawn 'guix authenticate' once for each item that has
to be signed (when exporting) or authenticated (when importing). Now,
we spawn it once for all and then follow a request/reply protocol. This
reduces the wall-clock time of:
guix archive --export -r $(guix build coreutils -d)
from 30s to 2s.
* guix/scripts/authenticate.scm (sign-with-key): Return the signature
instead of displaying it. Raise a &formatted-message instead of calling
'leave'.
(validate-signature): Likewise.
(read-command): New procedure.
(define-enumerate-type, reply-code): New macros.
(guix-authenticate)[send-reply]: New procedure.
Change to read commands from current-input-port.
* nix/libstore/local-store.cc (runAuthenticationProgram): Remove.
(authenticationAgent, readInteger, readAuthenticateReply): New
functions.
(signHash, verifySignature): Rewrite in terms of the agent.
* tests/store.scm ("import not signed"): Remove 'pk' call.
("import signed by unauthorized key"): Check the error message of C.
* tests/guix-authenticate.sh: Rewrite using the new protocol.
fixlet
* nix/libstore/build.cc (DerivationGoal::tryBuildHook): Add "offload" to
'args' and pass settings.guixProgram as the first argument to
Agent::Agent.
(pathNullDevice, commonChildInit, Agent, Agent::Agent)
(Agent::~Agent): Move to...
* nix/libutil/util.cc: ... here.
* nix/libutil/util.hh (struct Agent, commonChildInit): New
declarations.
* nix/libstore/local-store.cc (signHash, verifySignature): New
functions.
(LocalStore::exportPath): Use 'signHash' instead of inline code.
(LocalStore::importPath): Use 'verifySignature' instead of inline code.
* nix/libstore/build.cc (HookInstance): Rename to...
(Agent): ... this. Rename 'toHook' and 'fromHook' similarly and update
users. Change constructor to require a command and an argument list.
(DerivationGoal::tryBuildHook): Pass arguments to the 'Agent'
constructor.
* guix/svn-download.scm (download-svn-to-store): Add a subdirectory to
the download path. The subdirectory is used as the target for the 'svn
export' command, avoiding a 'directory exists' error when attempting to
use the parent directory directly.
It was made transactional in a4678c6ba1, with
the reasoning to prevent broken intermediate states from being visible. I
think this means something like an entry being in ValidPaths, but the Refs not
being inserted.
Using a transaction for this makes sense, but I think using one single
transaction for the whole register-items call is unnecessary to avoid broken
states from being visible, and could block other writes to the store database
while register-items is running. Because the deduplication and resetting
timestamps happens within the transaction as well, even though these things
don't involve the database, writes to the database will still be blocked while
this is happening.
To reduce the potential for register-items to block other writers to the
database for extended periods, this commit moves the transaction to just wrap
the call to sqlite-register. This is the one place where writes occur, so that
should prevent the broken intermediate states issue above. The one difference
this will make is some of the registered items will be visible to other
connections while others may be still being added. I think this is OK, as it's
equivalent to just registering different items.
* guix/store/database.scm (register-items): Reduce transaction scope.
Signed-off-by: Ludovic Courtès <ludo@gnu.org>
It's necessary that store items be locked and protected from garbage
collection while they are being registered. This documents that.
* guix/store/database.scm (register-path, register-items): document GC
protection and locking requirements.
Signed-off-by: Ludovic Courtès <ludo@gnu.org>
This causes with-writable-file to take into consideration the actual store
being used, as passed to 'deduplicate', rather than
whatever (%store-directory) may return.
* guix/store/deduplication.scm (replace-with-link): new keyword argument
'store'. Pass to with-writable-file.
(with-writable-file, call-with-writable-file): new store argument.
(deduplicate): pass store to replace-with-link.
Signed-off-by: Ludovic Courtès <ludo@gnu.org>
