[pull] master from ruby:master - #1284
Merged
Merged
Conversation
tool/mkdepend.rb resolved dependency names against the current working
directory before falling back to the source root. When configure
re-runs in an already-built build directory nested inside the source
tree (e.g. .ruby under the checkout), leftover generated files such
as builtin_binary.rbbin got emitted as build-dir-relative paths:
builtin.$(OBJEXT): .ruby/builtin_binary.rbbin
Make satisfies such a prerequisite through VPATH as a plain existing
file, so it never matches the builtin_binary.rbbin rule and the file is
never regenerated. Linking a stale builtin_binary.rbbin with
up-to-date *.rbinc function tables then makes the built ruby fail to
boot during make install:
<internal:gc>: builtin function index (8) mismatch (expect _bi454 but _bi464) (ArgumentError)
The same defect also emitted .ruby/probes.h and
.ruby/vm_call_iseq_optimized.inc, and dropped the enc/trans/*.trans
dependencies of the generated transcoder sources.
Convert dependency names against the source root only, never the
current working directory, so that generated files keep the bare names
their Make rules use and the generated dependencies no longer vary with
leftover build artifacts.
There are actually 3 tokens in the Ruby grammar: the newline in a whitespace insensitive position that is ignored, a newline that is in a whitespace sensitive position that acts as a statement terminator, and a newline in a whitespace sensisitive position that acts as the terminator for an expression. The third one doesn't exist in our grammar at the moment, but I want to add it, because it makes working with translating the lex output easier. ruby/prism@3dbe592a76
JRuby must match the native implementation when captures become empty. ruby/strscan@535f9b43ba
`integer_at` must read captures from the scanner current string. ruby/strscan@435095d898
This reverts commit ruby/prism@3dbe592a76cb. ruby/prism@115d58f71b
Bumps the github-actions group with 3 updates in the / directory: [github/codeql-action/init](https://github.com/github/codeql-action), [github/codeql-action/analyze](https://github.com/github/codeql-action) and [github/codeql-action/upload-sarif](https://github.com/github/codeql-action). Updates `github/codeql-action/init` from 4.37.4 to 4.37.5 - [Release notes](https://github.com/github/codeql-action/releases) - [Changelog](https://github.com/github/codeql-action/blob/main/CHANGELOG.md) - [Commits](github/codeql-action@f205ea1...d1ba80a) Updates `github/codeql-action/analyze` from 4.37.4 to 4.37.5 - [Release notes](https://github.com/github/codeql-action/releases) - [Changelog](https://github.com/github/codeql-action/blob/main/CHANGELOG.md) - [Commits](github/codeql-action@f205ea1...d1ba80a) Updates `github/codeql-action/upload-sarif` from 4.37.4 to 4.37.5 - [Release notes](https://github.com/github/codeql-action/releases) - [Changelog](https://github.com/github/codeql-action/blob/main/CHANGELOG.md) - [Commits](github/codeql-action@f205ea1...d1ba80a) --- updated-dependencies: - dependency-name: github/codeql-action/init dependency-version: 4.37.5 dependency-type: direct:production update-type: version-update:semver-patch dependency-group: github-actions - dependency-name: github/codeql-action/analyze dependency-version: 4.37.5 dependency-type: direct:production update-type: version-update:semver-patch dependency-group: github-actions - dependency-name: github/codeql-action/upload-sarif dependency-version: 4.37.5 dependency-type: direct:production update-type: version-update:semver-patch dependency-group: github-actions ... Signed-off-by: dependabot[bot] <support@github.com>
Darwin 27 answers the retry connect(2) on a refused nonblocking socket with EISCONN, so the retry idiom in Addrinfo#connect_internal returned an unconnected socket. SO_ERROR still holds the real error, so consult it after wait_writable, as wait_connectable() in init.c already does. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Commit bb235bd ("Avoid build-time race condition when cross-compiling from a source tarball") turned out to be only a partial fix, we also need to cover the case where git is present but we're not building from a git checkout. We can simply check if the `git log` command returned any output, if it didn't then we're not in a git checkout.
[Bug #22123] Bundler evaluates gemspecs through TOPLEVEL_BINDING, which always belongs to the main box, so consuming BUNDLER_SETUP while the root box loads RubyGems runs Bundler code in the main box before RubyGems has finished loading there. Skip it in every box once error_highlight, did_you_mean and syntax_suggest are autoloaded, since nothing loaded here needs the bundle and RUBYOPT=-rbundler/setup sets Bundler up after the boot sequence. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Ruby 4.1 autoloads error_highlight, did_you_mean and syntax_suggest and deletes BUNDLER_SETUP while loading RubyGems, so the early bundler/setup is only there for Ruby 3.2 through 4.0. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Based on the test in #17323, extended to cover user boxes and the configuration where the decorator gems are not autoloaded, which is how Ruby 3.2 through 4.0 behave. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
error_highlight, did_you_mean and syntax_suggest are still loaded eagerly on Ruby 3.2 through 4.0, where BUNDLER_SETUP is loaded in the main box rather than skipped everywhere, so derive the expectation from the actual autoload state. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Add documentation-only deprecation notices to Module#ruby2_keywords, main.ruby2_keywords, Proc#ruby2_keywords, Hash.ruby2_keywords_hash?, and Hash.ruby2_keywords_hash, as the first phase of the schedule proposed at https://bugs.ruby-lang.org/issues/22205.
winnt_stat opened a real file handle (open_special +
GetFileInformationByHandle + GetFileType + get_handle_pathname +
CloseHandle, ~5 syscalls) for every existing file just to stat it.
require does this thousands of times per startup.
Add a fast path that returns size/timestamps/attributes from a single
metadata syscall for regular files and directories:
- GetFileInformationByName (Windows 11 24H2+): one syscall, also gives
real FileId and link count -> accurate st_ino/st_nlink. Resolved via
GetProcAddress; Windows 10 and earlier fall through.
- GetFileAttributesExW (all supported Windows): one syscall for
size+times+attrs; st_ino/st_nlink left 0/1 (same compromise as the
existing stat_by_find fallback).
Reparse points (symlinks, AF_UNIX sockets) and unusual errors fall
through to the original handle-based path unchanged.
Single file, one function, no new dependencies. Measured on Ruby 4.0.5:
require "nokogiri" 224 ms -> 81 ms (2.77x); File.stat of an existing
file 112 us -> 21 us (5.4x).
Query FileStatBasicByNameInfo instead of FileStatByNameInfo for the 128-bit file ID; the 64-bit ID does not match the handle-based result on ReFS. Fall back to the handle-based path unless the volume serial matches the drive of the path, to keep st_dev of files behind cross-volume junctions and mount points, and also for non-disk devices and filesystems without file IDs. Drop the GetFileAttributesExW path that loses st_ino. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Since `prc.refined(*ms).refined(*ns)` is behaviorally equivalent to `prc.refined(*ms, *ns)`, the copy of the block is deferred until the first call and the memo is shared by these behaviorally equivalent Procs. A Proc that is never called is no longer copied at all. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
[Feature #22222] This allows to efficiently load iseq from mmaped files and other buffers without having to copy the bytes into a Ruby string an then invoke `RubyVM::InstructionSequence.load_from_binary`.
Up to `RHASH_AR_TABLE_MAX_SIZE` there's no need to initialize an `st_table`.
Introduce a streaming source hash API (rb_source_hash_init/update/ finalize) in ruby_parser.c, and use it in the lexer of parse.y to accumulate a hash of the source as each line is read. The hash is stored in the AST, and will be used to check whether a file still contains the same source code when it is re-parsed later. [Feature #21795] The hash algorithm (currently FNV-1a) is hidden behind the API as an implementation detail of the interpreter, so it can be changed freely between releases. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
After each prism parse, compute a source hash of the parsed source and keep it in the scope node. The data section after an __END__ marker is not part of the code, so the hash covers the source only up to the end of the __END__ line, which matches the range that parse.y hashes. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Record the source hash in each ISeq. For prism, the hash computed at parse time is propagated through the scope nodes; for parse.y, it is taken from the AST, and copied to child iseqs whose AST wrappers are created separately. Expose it along with the node id via RubyVM::InstructionSequence#source_hash and #node_id, include it in the misc hash of #to_a, and preserve it in ISeq binaries (bumping IBF_DEVEL_VERSION). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
An alias or a method installed via define_method(UnboundMethod) shares the original method definition, but the CME's owner and defined_class point at the site where the copy was installed. Combined with the method name (taken from the original definition), backtraces reported a "Class#method" pair that never existed: an alias in a subclass was shown as Child#original instead of Parent#original, and define_method(Original.instance_method(:m)) was shown as A#m instead of Original#m. Recover the defining module only for shared definitions (def->aliased), leaving plain, singleton and class methods untouched. Apply it to both Location#label and the backtrace string built by location_to_str. Distinguish the singleton-class cases via the singleton's attached object: define_method(SomeModule.instance_method(:m)) installed on a singleton class must report SomeModule#m, while an aliased class method (def self.m), whose iseq cref holds the lexical class rather than the singleton it lives on, must keep its owner. Only keep the owner when it is the singleton class of the cref's class; otherwise the cref names the genuine definition site, so use it. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Backtrace labels need the module where a method body was originally defined, but alias_method and define_method(UnboundMethod) overwrite the CME owner with the install site, producing "Class#method" pairs that never existed. The label code reconstructed the origin from a mix of owner, defined_class and the iseq cref with several heuristics. Store it directly instead: a new original_module field on rb_method_definition_t, set once when a body is created. Because the definition is reference-counted and shared, every alias/define_method copy inherits it, so the backtrace reader collapses to a field read. module_function is the one case that also installs an instance method's shared definition onto the module's singleton class; that copy is still labeled by its owner (M.f), detected as the copy whose owner is the singleton class of the definition module. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
This patch adds the following methods to String class: * String#bit_get(offset, lsb_first: true) -> 1 | 0 | nil * String#bit_set?(offset, lsb_first: true) -> true | false | nil * String#bit_set(offset, lsb_first: true) -> self * String#bit_clear(offset, lsb_first: true) -> self * String#bit_flip(offset, lsb_first: true) -> self * String#bit_count -> Integer * String#bitwise_not -> String * String#bitwise_not! -> self * String#bitwise_and(other) -> String * String#bitwise_and!(other) -> self * String#bitwise_or(other) -> String * String#bitwise_or!(other) -> self * String#bitwise_xor(other) -> String * String#bitwise_xor!(other) -> self Other than implementation, tests, specs, and docs are added. Link: [Feature #22118] ## Note In `string.c`, I wrote some big macro that create method functions and helper functions: * STR_DEFINE_BINARY_BITWISE_METHOD * STR_DEFINE_UNARY_BITWISE_KERNEL * STR_DEFINE_BINARY_BITWISE_KERNEL While using macros like this reduces maintainability, I believe it's acceptable because there are no plans to extend the `bitwise_*` methods beyond this proposal, and the logic is stable. On the other hand, other methods such as `bit_get` and `bit_count` are planned to have argument extensions in the future.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to subscribe to this conversation on GitHub.
Already have an account?
Sign in.
12 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
See Commits and Changes for more details.
Created by
pull[bot] (v2.0.0-alpha.4)
Can you help keep this open source service alive? 💖 Please sponsor : )