aboutsummaryrefslogtreecommitdiffstats
Commit message (Collapse)AuthorAgeFilesLines
* reduce rb_mutex_t size from 160 to 80 bytes on 64-bitnormal2017-05-082-100/+62
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Instead of relying on a native condition variable and mutex for every Ruby Mutex object, use a doubly linked-list to implement a waiter queue in the Mutex. The immediate benefit of this is reducing the size of every Mutex object, as some projects have many objects requiring synchronization. In the future, this technique using a linked-list and on-stack list node (struct mutex_waiter) should allow us to easily transition to M:N threading model, as we can avoid the native thread dependency to implement Mutex. We already do something similar for autoload in variable.c, and this was inspired by the Linux kernel wait queue (as ccan/list is inspired by the Linux kernel linked-list). Finaly, there are big performance improvements for Mutex benchmarks, especially in contended cases: measure target: real name |trunk |built ----------------|------:|------: loop_whileloop2 | 0.149| 0.148 vm2_mutex* | 0.893| 0.651 vm_thread_mutex1| 0.809| 0.624 vm_thread_mutex2| 2.608| 0.628 vm_thread_mutex3| 28.227| 0.881 Speedup ratio: compare with the result of `trunk' (greater is better) name |built ----------------|------: loop_whileloop2 | 1.002 vm2_mutex* | 1.372 vm_thread_mutex1| 1.297 vm_thread_mutex2| 4.149 vm_thread_mutex3| 32.044 Tested on AMD FX-8320 8-core at 3.5GHz * thread_sync.c (struct mutex_waiter): new on-stack struct (struct rb_mutex_struct): remove native lock/cond, use ccan/list (rb_mutex_num_waiting): new function for debug_deadlock_check (mutex_free): remove native_*_destroy (mutex_alloc): initialize waitq, remove native_*_initialize (rb_mutex_trylock): remove native_mutex_{lock,unlock} (lock_func): remove (lock_interrupt): remove (rb_mutex_lock): rewrite waiting path to use native_sleep + ccan/list (rb_mutex_unlock_th): rewrite to wake up from native_sleep using rb_threadptr_interrupt (rb_mutex_abandon_all): empty waitq * thread.c (debug_deadlock_check): update for new struct (rb_check_deadlock): ditto [ruby-core:80913] [Feature #13517] git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@58604 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* relative path as VPATHnobu2017-05-071-2/+2
| | | | | | | * common.mk (ripper.c): use relative path as VPATH instead of absolute path of the current working directory. git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@58603 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* parse.y: export rb_parser_fatal for rippernobu2017-05-071-1/+1
| | | | git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@58602 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* stale targetsnobu2017-05-073-39/+4
| | | | | | * common.mk (update-mspec, update-rubyspec): empty stale targets. git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@58601 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* extract prereq.statusnobu2017-05-072-32/+55
| | | | | | | * tool/make-snapshot (package): extract static config values for prereq and override dynamic values at runtime. git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@58600 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* Update specs to follow newlines added by r58596eregon2017-05-073-3/+3
| | | | git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@58599 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* Rename spec/README to spec/README.mderegon2017-05-071-0/+0
| | | | git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@58598 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* * properties.svn2017-05-070-0/+0
| | | | git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@58597 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* * remove trailing spaces, append newline at EOF.svn2017-05-0730-31/+37
| | | | git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@58596 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* Add in-tree mspec and ruby/speceregon2017-05-074645-4/+230678
| | | | | | | | | | | * For easier modifications of ruby/spec by MRI developers. * .gitignore: track changes under spec. * spec/mspec, spec/rubyspec: add in-tree mspec and ruby/spec. These files can therefore be updated like any other file in MRI. Instructions are provided in spec/README. [Feature #13156] [ruby-core:79246] git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@58595 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* make update-rubyspec is no longer needederegon2017-05-072-8/+4
| | | | | | * Kept a no-op for compatibility. git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@58594 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* Update README about ruby/speceregon2017-05-071-31/+79
| | | | git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@58593 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* * spec/default.mspec: only require rbconfig if is it not already.eregon2017-05-071-1/+1
| | | | | | Useful when running spec/mspec/bin/mspec directly. git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@58592 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* vm_core.h (rb_thread_t): use 32-bit running_time_usnormal2017-05-072-3/+3
| | | | | | | | | | | | | | | | | | | | The current range based current values of: TIME_QUANTUM_USEC=100000 RUBY_THREAD_PRIORITY_MAX=3 RUBY_THREAD_PRIORITY_MIN=-3 Gives a range of 12500..800000, plenty enough for a 32-bit integer. Clamping this also reduces potential implementation bugs between 32-bit and 64-bit platforms. I may consider a further reduction to uint16_t in the future for M:N threading, but some users may want slightly larger time quantums. * vm_core.h (rb_thread_t): use 32-bit running_time_us git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@58591 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* io.c (do_fcntl): update max FD for F_DUPFD_CLOEXEC, toonormal2017-05-071-3/+9
| | | | | | | | | | Somebody may pass 1030 (the value of F_DUPFD_CLOEXEC) to IO#fcntl because they copied code from somewhere else. Ensure we know about FDs created that way. * io.c (do_fcntl): update max FD for F_DUPFD_CLOEXEC, too git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@58590 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* NEWS: Integer#{round,floor,ceil,truncate} [Bug #13420]stomar2017-05-071-0/+2
| | | | git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@58589 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* * properties.svn2017-05-070-0/+0
| | | | git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@58588 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* variable.c: cleanup waitq upon thread deathnormal2017-05-073-8/+57
| | | | | | | | | | | | | * variable.c (autoload_reset): use idempotent list_del_init (autoload_sleep): moved code from rb_autoload_load (autoload_sleep_done): cleanup for use with rb_ensure (rb_autoload_load): ensure list delete happens in case the thread dies during sleep * test/ruby/bug-13526.rb: new script for separate execution * test/ruby/test_autoload.rb (test_bug_13526): new test [ruby-core:81016] [Bug #13526] git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@58587 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* make Integer#{round,floor,ceil,truncate} always return integerstomar2017-05-062-36/+32
| | | | | | | | | | | | | | | * numeric.c (int_round): return integer (self) instead of float for Integer#round with positive ndigits argument, because conversion to float introduces errors for large integers. * numeric.c (int_floor): ditto for Integer#floor. * numeric.c (int_ceil): ditto for Integer#ceil. * numeric.c (int_truncate): ditto for Integer#truncate. * test/ruby/test_integer.rb: adjust test cases and add some more. [ruby-core:80645] [Bug #13420] git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@58586 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* * 2017-05-07svn2017-05-061-1/+1
| | | | git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@58585 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* ripper/lexer.rb: nested indented heredocnobu2017-05-063-1/+20
| | | | | | | | | * ext/ripper/lib/ripper/lexer.rb (on_heredoc_dedent): insert stripped leading spaces as `on_ignored_sp` elements, so that the original source can be reconsructed. [ruby-core:80977] [Bug #13536] git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@58584 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* fix up r58573nobu2017-05-061-1/+1
| | | | | | | * common.mk: make the directory for extension shared library with the timestamp directory. git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@58583 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* math.c: improve docs for Math.sqrtstomar2017-05-061-1/+1
| | | | | | * math.c: [DOC] mention Integer.sqrt in docs for Math.sqrt. git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@58582 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* numeric.c: remove mention of Bignum from docsstomar2017-05-061-9/+8
| | | | | | | * numeric.c: [DOC] remove mention of Bignum from Integer#{+,-,*,/}, the return type does not depend on magnitude anymore. git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@58581 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* numeric.c: revise docsstomar2017-05-061-291/+257
| | | | | | | | | | | | * numeric.c: [DOC] revise docs for Numeric, Integer, Float: * nodoc Numeric#singleton_method_added * mention that result for Integer#** might also be Complex * add / simplify / fix some examples * mention aliases * fix rdoc formatting, typos, grammar * clarifications and other improvements git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@58580 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* delete enc/prelude.rb, because no longer neededduerst2017-05-061-4/+0
| | | | git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@58579 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* delete lib/unicode_normalize.rb, because no longer neededduerst2017-05-061-11/+0
| | | | | | (all content has been moved to string.c) git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@58578 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* remove enc/prelude.rb from list of prelude files in common.mkduerst2017-05-061-1/+1
| | | | | | | | (This is a retry of r58559, it should work now that rubyspec does no longer check for unicode_normalized to be required. See also https://github.com/ruby/spec/commit/41176ead68b14011658c8abd0ebd67df2ae632fb.) git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@58577 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* share `@@accept_charset`nobu2017-05-063-2/+17
| | | | | | | | * lib/cgi/{core,util}.rb: include CGI::Util not only extending, to share `@@accept_charset` class variable, so that it is always accessible. [ruby-core:80986] [Bug #13539] git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@58576 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* string.c: fix typesnobu2017-05-061-3/+3
| | | | | | | | | * string.c (id_normalize, id_normalized_p): fix types, IDs should be ID. * string.c (unicode_normalize_common): ditto. git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@58575 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* * 2017-05-06svn2017-05-061-1/+1
| | | | git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@58574 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* timestamp for extensionsnobu2017-05-061-4/+4
| | | | | | | * common.mk: use the same timestamp file for architecture specific directory as the file used in Makefiles under ext. git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@58573 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* * 2017-05-05svn2017-05-041-1/+1
| | | | git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@58572 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* benchmark/bm_vm_thread_pass_flood.rb: add commentnormal2017-05-041-0/+2
| | | | | | | | | | | | | I was about to write off this benchmark while working on GVL improvements on multi-core systems. However I noticed it exposes a weakness in my work-in-progress code when I tested on an old single CPU system. Further testing reveals setting CPU affinity ("schedtool -a 0x1" on Linux) on a modern multi-core system is enough to reproduce the problem exposed by this benchmark. git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@58571 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* time.c: avoid taking a pointer to a member of packed structmame2017-05-041-7/+23
| | | | | | | | clang 4.0.0 emitted a warning: "taking address of packed member 'subsecx' of class or structure 'vtm' may result in an unaligned pointer value [-Waddress-of-packed-member]". git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@58570 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* string.c: [DOC] improve docs for String.newstomar2017-05-041-8/+10
| | | | git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@58569 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* internal.h (rb_gc_resurrect): remove stale declarationktsj2017-05-041-2/+0
| | | | | | rb_gc_resurrect is no longer defined since r47444. git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@58568 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* string.c: [DOC] Properly refer to keyword argument by its namektsj2017-05-041-2/+2
| | | | git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@58567 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* dir.c: [DOC] Properly refer to keyword argument by its name [Fix GH-1543]ktsj2017-05-041-4/+4
| | | | | | | | | | | | enc is the name of the variable used in the example, not the name of the keyword argument (encoding). The documentation used to wrongly suggest that the keyword argument name was "enc" which could cause people try try to call `Dir.open("thing", enc: "utf-8")` Author: Olivier Lacan <hi@olivierlacan.com> git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@58566 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* benchmark/bm_vm1_gc_wb_*.rb: fix GC.start args for invalid keywordnormal2017-05-044-4/+4
| | | | | | | | | | | | | | | | | | "lazy_sweep" does not appear to have ever been a valid kwarg for GC.start, however the opposite of "lazy_sweep" appears to be "immediate_sweep". So use immediate_sweep, and flip the boolean value of each arg. I guess this only started failing with r56981 in Dec 2016 ("class.c: missing unknown_keyword_error", commit e3f0cca2f26ba44c810ac980cdff7dda129ae533) * benchmark/bm_vm1_gc_wb_ary.rb: "lazy_sweep: false" => "immediate_sweep: true" * benchmark/bm_vm1_gc_wb_ary_promoted.rb: ditto * benchmark/bm_vm1_gc_wb_obj.rb: ditto * benchmark/bm_vm1_gc_wb_obj_promoted.rb: ditto git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@58565 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* gems/bundled_gems: Update to power_assert 1.0.2ktsj2017-05-041-1/+1
| | | | git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@58564 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* .gitignore: ignore benchmark/bm_require.datanormal2017-05-041-0/+1
| | | | | | | | Followup-to r58562 in SVN * .gitignore: ignore benchmark/bm_require.data git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@58563 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* benchmark: ignore bm_require.data/normal2017-05-040-0/+0
| | | | git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@58562 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* parse.y: make shared string modifiablenobu2017-05-042-1/+15
| | | | | | | | * parse.y (dedent_string): ensure that the string is modifiable, not to set the length of shared string. [ruby-core:80987] [Bug #13540] git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@58561 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* revert 58559, because it fails on travis (rubyspec problem)duerst2017-05-041-1/+1
| | | | git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@58560 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* remove enc/prelude.rb from list of prelude files in common.mkduerst2017-05-041-1/+1
| | | | git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@58559 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* refactor common parts of unicode normalization functions into ↵duerst2017-05-041-31/+20
| | | | | | | | | | | unicode_normalize_common In string.c, refactor the common parts (requiring of unicode_normalize/normalize.rb, check of number of arguments) of the unicode normalization functions (rb_str_unicode_normalize, rb_str_unicode_normalize_bang, rb_str_unicode_normalized_p) into the new function unicode_normalize_common. git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@58558 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* * properties.svn2017-05-040-0/+0
| | | | git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@58557 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* benchmark: new single-threaded read/write benchmark with pipenormal2017-05-041-0/+13
| | | | | | | | | | | This is currently for testing GVL performance in the uncontended case: IO#write and IO#read unconditionally release GVL for blocking I/O with pipe. It will also be interesting to see how this changes if we switch to M:N threading model. git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@58556 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* move definition of String#unicode_normalized? to C to make sure it is documentedduerst2017-05-043-30/+38
| | | | | | | | | | | | * lib/unicode_normalize.rb: Remove definition of String#unicode_normalized? (including documentation). Leave a comment explaining that the file is now empty. * string.c: Define String#unicode_normalized? in rb_str_unicode_normalized_p in C, (including documentation) * lib/unicode_normalize/normalize.rb: Remove (re)definition of String#unicode_normalized? to avoid warnings (when $VERBOSE==true) and problems when String is frozen git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@58555 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
generated by cgit v1.2.3 (git 2.39.1) at 2024-07-26 10:33:13 +0000