| Age | Commit message (Collapse) | Author |
|---|
|
The --profile flag parsed the option but never actually enabled the
timing/memory output, because ConsoleIO::enableDebugging existed only as an
inherent method, unreachable through the `dyn IOInterface` handle held by
Application. Promote enable_debugging to an IOInterface trait method
(default panics, since only ConsoleIO/BufferIO ever legitimately receive
this call) so do_run can invoke it directly without downcasting.
|
|
Wires PartialComposer.as_full() to fetch the PluginManager, replacing
the todo!() stub. Mirrors PHP's assertion that $this->composer must be
a fully-loaded Composer instance.
|
|
ArgvInput stores variadic arguments as PhpMixed::List, not
PhpMixed::Array, so as_array() always returned None and the search
query was silently empty, causing packagist to reject every request
with a 400 Bad Request.
|
|
PHP's addOverriddenPackage calls parent::addPackage() (a non-virtual
call straight to ArrayRepository), but the port called self.add_package,
re-entering PlatformRepository::add_package. Since the newly created
override package's name also starts with "php-", this recursed
forever and blew the stack.
|
|
Avoid colliding with an existing Composer installation on the same
machine by defaulting COMPOSER_HOME/CACHE_DIR/DATA_DIR paths to
shirabe/Shirabe instead of composer/Composer, while keeping the env
var names, composer.json/lock, and vendor/composer/ unchanged for
ecosystem compatibility.
|
|
AnyPackage::dup() (PHP's `clone $package`) copied the `repository` and
`id` fields verbatim instead of resetting them like PHP's
BasePackage::__clone() does. LibraryInstaller::install() relies on the
duplicate being unbound so it can register the package with the local
repository; without the reset, add_package() silently failed with "A
package can only be added to one repository", leaving installed.json
empty after every install/create-project run.
|
|
Two related "already borrowed" panics reachable from
AutoloadGenerator::dump() (which holds the local-repository,
installation-manager, and config RefCells for the duration of its own
statement, per the temporary-lifetime-extension pattern fixed
separately in create_project_command.rs):
- ensure_bin_dir_is_in_path called config.borrow_mut() to read
"bin-dir", but Config::get only needs &self; use borrow() so it can
coexist with an outer borrow instead of conflicting with it.
- make_autoloader's real body needed composer_handle.borrow_mut() plus
the same local-repository/installation-manager RefCells the caller
already holds mutably, which cannot be made reentrant-safe without a
larger restructuring. Since all 3 call sites already discard its
return value, and its only effect (registering a Composer-generated
ClassLoader for autoloading during event-listener PHP execution) is
unobservable in this port — there's no embedded PHP interpreter to
register it into, and class_exists for user-defined classes is a
hardcoded-false shim so the caller's very next check always treats
the class as unavailable regardless — make it a genuine no-op.
This unblocks the post-autoload-dump event for any script listener
naming a PHP class (e.g. Illuminate\Foundation\ComposerScripts), which
every create-project/install run reaches once real packages get
installed.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
|
|
PHP's createRepositorySet does `$this->fixedRootPackage = clone
$this->package;` once and then passes that same object both to `new
RootPackageRepository($this->fixedRootPackage)` and, later, to
createRequest($this->fixedRootPackage) — object identity matters
because the solver assigns a pool id by mutating the package object
itself.
The port instead called RootPackageInterfaceHandle::dup() a second
time when registering the RootPackageRepository, producing a second
object that never went through the pool and so never got an id.
create_request's `request.fix_package(root_package_handle)` then
referenced a package with id -1, which add_rules_for_request treats as
a real bug: "Fixed package ... was not added to solver pool." This
surfaced whenever a create-project run reached the second-stage
install (i.e. every dist-installed project with real dependencies).
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
|
|
install_project's fluent builder chain called config.borrow_mut() four
times inline as method arguments. Rust extends every argument
temporary's lifetime to the end of the enclosing statement, so the
first borrow_mut() was still alive when the second one ran, panicking
with "already borrowed". Compute each value into a local before the
chain instead, matching the pattern already used in
install/update/require/remove_command.rs.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
|
|
Route the platform-repository seams that need real PHP introspection
through php-rpc instead of todo!()/hardcoded shims:
- Runtime::invoke now handles the two dynamic callables
PlatformRepository actually reaches (inet_pton, curl_version) via new
php-rpc calls; other callables remain unsupported.
- Runtime::get_extension_info uses a new `extension_info` php-rpc call
(ReflectionExtension::info() + output buffering) instead of todo!().
- Runtime::get_extensions/get_extension_version now query the real PHP
process (get_loaded_extensions, phpversion) instead of the
shirabe-php-shim's hardcoded "standard CLI environment" model, so
platform requirement checks see the extensions actually installed.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
|
|
Several Preg::*() call sites lost their PHP delimiter (and in one case
the `i` modifier) during porting, since preg_*() expects the delimiter
to be preserved in the caller's pattern literal and stripped internally.
This made compile_php_pattern panic or silently misparse the pattern.
Un-ignore the Version tests that were blocked by this bug.
|
|
Runtime::hasConstant/getConstant need a real PHP interpreter's defined()/
constant() to answer platform requirement checks (e.g. PHP_ZTS, PHP_INT_SIZE),
which the shim can't provide since Rust constants aren't queryable by string.
Extend shirabe-php-rpc's protocol to carry one string argument and return the
full PHP scalar range, add defined/constant dispatch entries to the worker,
and wire Runtime and get_php_version/get_php_binary onto them.
|
|
Extends no_banned_use to cover std::any::Any, std::io::Read/Write, and
std::process::Command, and teaches the linter to allow `as _` imports
so trait methods can still be brought into scope without binding the
banned name. Fully qualifies all existing usages across the codebase.
|
|
PlatformRepository::initialize() (php-version/extension detection) was
never invoked: the RepositoryInterface impl delegated straight to the
inner ArrayRepository without the ensure_initialized lazy-init guard
that sibling repositories (FilesystemRepository, PathRepository) use.
As a result pool.what_provides("php") was always empty, and any
package requiring php failed platform resolution. Also fixes the
trait search() bypassing PlatformRepository's own SEARCH_VENDOR
override.
|
|
Port testNoPluginsDisablesPluginsWhenScriptCommandsExist and
testScriptCommandTakesPriorityOverAbbreviatedBuiltinCommand. Both stay
#[ignore]d because do_run panics at the script-registration todo!()
(application.rs:2461) when composer.json has scripts. Add a test-only
ApplicationHandle::__get_composer accessor so the first test's
getPluginManager assertions can be expressed.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
|
|
Add a minimal shirabe-php-rpc crate that spawns the system PHP as a
child process and asks it for runtime information over a Unix domain
socket, then use it to fill the `--version` PHP line with the real
\PHP_VERSION and \PHP_BINARY instead of fixed placeholder values.
See docs/dev/php-rpc.md for the design and scope.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
|
|
Model Git::runCommand's mixed $commandOutput parameter with the
RunCommandOutput trait (one impl per PHP mode: discard, by-ref capture,
callable handler), mirroring ProcessExecutor's IntoExecOutput. This moves
the is_callable($commandOutput) value-inspection into the type system and
drops the dependency on the shim's incomplete is_callable.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
|
|
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
|
|
Resolve the two todo!() placeholders by passing self.inner.io.clone()
to VcsDriverBase::new and Forgejo::new, matching the PHP source's
$this->io arguments and the GitHub/GitLab driver idiom.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
|
|
StringInput inherits __toString from ArgvInput in PHP; mirror that with a
Display impl delegating to its inner ArgvInput, and use it in
GlobalCommand::input_to_string.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
|
|
Resolve the remaining todo! in update_requirements_after_resolution by
passing the stability-flags rewriter closure to Locker::update_hash,
porting the static closure from RequireCommand.php.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
|
|
Port AuditCommand::getPackages's non-locked path: build an
InstalledRepository from the local repository and return its packages,
filtered by RootPackage requires when --no-dev is set. The prior
TODO(phase-c) assumption (InstalledRepository::new vs get_local_repository
type mismatch) no longer holds since both sides use RepositoryInterfaceHandle.
Enables the two previously ignored audit command tests.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
|
|
It is faithful to PHP's by-value array/string semantics.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
|
|
The `!@json_decode($children)` guards in addSubNode/removeSubNode were
translated as `is_null() || as_bool() == Some(false)`, which drops PHP's
other falsy cases (empty array/string, "0", 0). At the removeSubNode site
(assoc=true) this diverged from Composer: an empty object node decodes to
an empty array, which PHP treats as falsy and aborts on, but the Rust
guard kept going. Replace both with php_truthy, the faithful rendering of
PHP `!`, preserving the deliberate empty-`{}` behavior difference between
the assoc=false (addSubNode) and assoc=true (removeSubNode) sites.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
|
|
Auditor::audit took io as &mut dyn IOInterface, forcing the audit and
installer post-audit call sites to hold a borrow_mut() on the shared IO
RefCell for the whole call. During advisory fetching the repositories
write to their own clones of the same handle, so the borrow_mut()
collided with their borrow() and panicked with 'RefCell already mutably
borrowed' on 'audit --locked'. Take the Rc<RefCell<dyn IOInterface>>
handle instead so writes borrow briefly and never overlap. Un-ignore the
locked-audit regression test that this unblocks.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
|
|
Validate JSON syntax with serde_json's parse errors in JsonFile, and detect
duplicate keys in ConfigValidator with a hand-written serde visitor, dropping
the now-unused JsonParser/Lexer/DuplicateKeyException ports. ParsingException
is kept as the thrown error type and downcast signal.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
|
|
Add a no_banned_use linter that forbids importing anyhow::Result, and
update all call sites to reference it via its fully-qualified path so
it is never confused with std::result::Result.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
|
|
flattenRepositories must recurse into InstalledRepository (which extends
CompositeRepository in PHP) and ShowCommand must unwrap FilterRepository
when categorizing repos. Without this, installed/locked/platform packages
all fell through to the "available" bucket, dropping the version column
and per-section grouping. Un-ignores 10 show_command tests.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
|
|
execute_batch only awaited prepare(), leaving the install/update/
uninstall + cleanup + repo.write promise chain as a todo!() stub, so
packages were downloaded to cache but never installed into the target
directory. Wire the operation step (mirroring PHP's promise chain),
propagate errors from the install/update/uninstall wrappers instead of
swallowing them with .ok()?, and write the repo after each op.
Un-ignore the create-project functional tests and the install/remove
command tests this unblocks. The remaining --no-install case still fails
on a separate install-path bug; its ignore reason is updated to match.
|
|
Port composer/tests/Composer/Test/InstallerTest.php. testInstaller (the
provideInstaller cases) is fully ported and passes; the three integration
tests port doTestIntegration in full (the .test fixture loader, FactoryMock,
the in-process console Application with install/update commands, and the
PHPUnit assertStringMatchesFormat matcher) and remain #[ignore]'d since the
install pipeline is not yet executable end-to-end.
Add test-only `__`-seams to the concrete types the test depends on, since
their consumers (e.g. Locker takes the concrete InstallationManager) and the
subclass-style mocks have no trait to mock: InstallationManager (recording
mock + as_any), Factory (__create_mock), VersionGuesser, and
InstalledFilesystemRepository. The production path (mock: false) is unchanged.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
|
|
The CurlDownloader owned a tokio runtime and block_on'd reqwest from its
sync tick(), while the repository/installer/downloader sync bridges each
created another Runtime and block_on'd async fns that reach that leaf.
Driving one Runtime::block_on from within another panics with "Cannot
start a runtime from within a runtime", hit by `require` when fetching
p2 metadata.
Switch CurlDownloader to a blocking reqwest client (its own internal
thread, never nested) and replace the per-call Runtime::new().block_on
bridges with a no-reactor sync_executor::block_on helper. No awaited
future parks on a reactor once the only async I/O is blocking, so the
helper can be nested freely.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
|
|
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
|
|
|
|
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
|
|
The structural methods are inherited from ArrayRepository in PHP, where
the lazy package load is driven by the overridden initialize(). Without
virtual dispatch, each repository now ensures that load via a &self
ensure_initialized() before delegating to the inner ArrayRepository.
This required moving the repositories' lazily-populated state behind
interior mutability (RefCell/Cell) and adding &self find_package_internal
/find_packages_internal helpers on ArrayRepository.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
|
|
Add __base_application accessor on ApplicationHandle so the test can
build an ApplicationDescription the way console commands do.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
|
|
LibraryInstaller and PluginInstaller upgrade the Composer back-reference
in their constructors, so they could not be built inside Rc::new_cyclic
where the weak handle is not yet upgradeable. Defer create_default_
installers until after the cyclic Rc is established, where the weak
handle resolves, and implement it to register Library -> Plugin ->
Metapackage with a single shared BinaryInstaller.
To share one BinaryInstaller (as Composer does), LibraryInstaller's
binary_installer becomes Rc<RefCell<dyn BinaryInstallerInterface>>
instead of an owned Box; PluginInstaller and the __set_binary_installer
test seam follow.
This clears "Unknown installer type: metapackage". Un-ignores the six
remove tests that now pass; the remaining install/remove tests are
re-labeled for the next blocker (InstallationManager::execute_batch
still leaves the install/cleanup/repo.write promise chain as a todo!()
stub, so package operations do not actually execute).
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
|
|
Replace the empty set_definition stub with the full InputArgument/
InputOption set from Composer's UpdateCommand. The symfony input
modeling was already complete; this was the last command still passing
an empty definition, which made it reject its own options.
Un-ignores test_no_security_blocking_allows_insecure_packages (now
passing) and re-labels the six remaining update tests with their actual
blockers (regex porting, resolver temporary-constraint, interactive
mode, bump-after-update solver pool) since the old "empty
InputDefinition" reason no longer applies.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
|
|
- StreamContextFactory: User-Agent reports the HTTP stack as "reqwest"
- RequestProxy::supports_secure_proxy: always true (reqwest+rustls can always
TLS to a proxy); drop the now-dead curl<7.52 guard in get_curl_options
- DiagnoseCommand::get_curl_version: phase-D TODO placeholder
Empirically verified: reqwest sends no default User-Agent (shirabe sets it
explicitly) and accepts https:// proxy URLs. init+install output stays
byte-identical to Composer.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
|
|
Empty PhpMixed::Array now serializes as [], so the several places that build a
PHP `new \stdClass` via an empty Array were emitting [] where Composer writes
{}. Use PhpMixed::Object for those empty-object cases:
- InitCommand: require / require-dev
- Locker::fixupJsonDataType: stability-flags / platform / platform-dev
- JsonConfigSource fallback: require/config keys that must stay objects
Align the affected test expectations with Composer: JsonFile::read() decodes
with assoc=true, collapsing the on-disk {} back to [], so reads expect [].
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
|
|
PHP heredocs drop the newline before the closing marker, so a two-blank-line
fragment must port to "\n\n", not "\n\n\n", and the static initializer needs a
"\n" after the $initializer placeholder. This makes autoload_real.php and
autoload_static.php byte-identical to Composer's output. Rewrite the templates
as raw strings mirroring the PHP heredocs.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
|
|
Port 11 categories of previously-ignored Composer tests now reachable with
the mockall crate: DownloadManager, VCS/Perforce/File downloaders,
VersionSelector, PlatformRepository, Auditor, installer/FilesystemRepository,
RootPackageLoader, util auth/http, commands, and Cache.
Extract test seams additively on concrete structs as *Interface traits
(Runtime, HhvmDetector, VersionGuesser, RepositorySet, Perforce,
BinaryInstaller) plus mock-field seams (Cache, Filesystem); consumers take
trait objects. Mocks are defined locally in the test crates via
mockall::mock!, since automock-generated mocks are cfg(test)-gated and
invisible across the integration-test boundary.
dataProviders are ported in full; tests blocked by unported shims stay
#[ignore] with documented reasons rather than reduced or weakened.
Fix product bugs surfaced by the ports:
- util/github: use the exception code, not the HTTP status, for 401/403
- advisory: serialize empty audit maps as [] to match PHP json_encode
- repository/filesystem and downloader/file: fix RefCell double-borrow panics
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
|
|
Three porting mismatches caused show/info output to diverge from PHP:
- "not found" hint appended " in /composer.json" whenever the
working-dir key existed; PHP uses isset(), which is false for the
null default. Now only appended when the value is non-null.
- printPackages and generatePackageTree inserted "" for a missing
description instead of null, so isset() rendered a spurious trailing
space (and JSON emitted "" instead of null). Both now preserve null.
Un-ignores the eight rendering-gap tests these fix (plus one already
passing), and rewrites the remaining ignore reasons to name the real
blocker (package categorization, --no-dev filtering, RefCell borrow,
installer resolution) instead of a stale generic message.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
|
|
The jsonschema crate reports a missing required property against its
parent object, leaving the instance path empty at the root. This skipped
the "PROPERTY : MESSAGE" formatting that Composer produces. Append the
missing property name to the path so the output shape matches.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
|
|
Composer/PartialComposer exposed its RepositoryManager, InstallationManager,
EventDispatcher, Locker, DownloadManager, AutoloadGenerator and ArchiveManager
as concrete types, but Composer's public setters (setDownloadManager() etc.)
let plugins swap in subclasses. Introduce a *Interface trait per manager and
store each as Rc<RefCell<dyn ...Interface>> so a replacement is honored.
Only Composer's slots and the sinks fed from its accessors become trait
objects; managers injected concretely at construction keep their concrete
references, matching PHP semantics. Fluent setters on the affected classes now
return () and Locker::update_hash is de-generified to a boxed FnOnce so the
traits stay object-safe.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
|
|
|
|
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
|
|
file_put_contents_if_modified held a filesystem borrow_mut() receiver
while its argument re-borrowed the same RefCell via dump_to_php_code(),
panicking with "RefCell already borrowed". Build the file contents
before taking the borrow.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
|
|
getMissingRequirementInfo created an empty InstalledRepository (both
repos were left commented out as a stub), so every locked requirement
was reported as "not present in the lock file". Pass [lockedRepo,
rootRepo] to the constructor as the PHP does.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
|
|
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
|
