summaryrefslogtreecommitdiff
path: root/compile.c
AgeCommit message (Collapse)Author
2019-08-26merge revision(s) a6a26e42b15c46f117f4fce07a2050e9d727355d: [Backport #15906]usa
compile.c: Partially revert r63870 which caused wrong optimization [Bug #15906] git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/branches/ruby_2_5@67757 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
2019-05-02merge revision(s) 66324: [Backport #15385]usa
Modify insn list only when compiling git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/branches/ruby_2_5@67708 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
2019-04-30merge revision(s) 66326,66649: [Backport #15385]usa
Fix infinite loop by ensure * compile.c (iseq_insert_nop_between_end_and_cont): insert nop so that the end of rescue and continuing points are not same, to get rid of infinite loop. [Bug #15385] test_optimization.rb: increase timeout for RubyCI https://rubyci.org/logs/rubyci.s3.amazonaws.com/opensuseleap/ruby-trunk/log/20181230T040002Z.fail.html.gz git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/branches/ruby_2_5@67702 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
2018-12-10merge revision(s) 62621: [Backport #15396]nagachika
compile.c: do not truncate VALUE to long * compile.c (ibf_dump_object_regexp): do not truncate VALUE to long. it makes invalid VALUE on IL32LLP64 platforms where long is shorter than VALUE. git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/branches/ruby_2_5@66309 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
2018-12-05merge revision(s) 64736,65567: [Backport #15270]nagachika
iseq.c: prefix rb_ to non-static iseq functions I assume we always prefix rb_ to non-static functions to avoid conflict. These functions are not exported and safe to be renamed. iseq.h: ditto compile.c: ditto Fix TracePoint for nested iseq loaded from binary [Bug#14702] When loading iseq from binary while a TracePoint is on, we need to recompile instructions to their "trace_" variant. Before this commit we only recompiled instructions in the top level iseq, which meant that TracePoint was malfunctioning for code inside module/class/method definitions. * compile.c: Move rb_iseq_init_trace to rb_ibf_load_iseq_complete. It is called on all iseqs during loading. * test_iseq.rb: Test that tracepoints fire within children iseq when using load_from_binary. This patch is from: Alan Wu <XrXr@users.noreply.github.com> git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/branches/ruby_2_5@66225 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
2018-12-02merge revision(s) 63040: [Backport #14553]nagachika
compile.c: do not dump unused callinfos * compile.c (compile_if): rewind callinfo indexes used in unreachable paths, to get rid of dumping unused callinfos. [ruby-core:86399] [Bug #14553] git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/branches/ruby_2_5@66130 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
2018-11-14merge revision(s) 62772: [Backport #14738]nagachika
compile.c: fix load_from_binary * compile.c (ibf_load_iseq_each): realpath may be nil. follow up r59709. [fix https://github.com/Shopify/bootsnap/issues/132] From: nobu <nobu@b2dd03c8-39d4-4d8f-98ff-823fe69b080e> git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/branches/ruby_2_5@65728 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
2018-11-06merge revision(s) 65350: [Backport #15245]nagachika
compile.c: fix peephole optimization * compile.c (iseq_peephole_optimize): should `pop` before jump instruction which succeeds to `newarray` of a literal object, not after. [ruby-core:89536] [Bug #15245] git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/branches/ruby_2_5@65580 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
2018-10-21merge revision(s) 64515:nagachika
remove `const` warning. * compile.c (iseq_ibf_load): remove `const` to pass iseq as no `const` parameter. git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/branches/ruby_2_5@65274 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
2018-10-11merge revision(s) 64745: [Backport #15118]nagachika
Use opt_{aref,aset} over opt_{aref,aset}_with * compile.c (iseq_compile_each0): Use `opt_aref`/`opt_aset` over `opt_aref_with`/`opt_aset_with` when frozen_string_literal: true, not to resurrect the index string on non-Hash receiver. [Fix GH-1957] From: chopraanmol1 <chopraanmol1@gmail.com> git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/branches/ruby_2_5@65002 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
2018-10-11merge revision(s) 64514: [Backport #14702]nagachika
check trace flags at loading [Bug #14702] * iseq.c (iseq_init_trace): at ISeq loading time, we need to check `ruby_vm_event_enabled_flags` to turn on trace instructions. Seprate this checking code from `finish_iseq_build()` and make new function. `iseq_ibf_load()` calls this funcation after loading. * test/ruby/test_iseq.rb: add a test for this fix. git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/branches/ruby_2_5@64997 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
2018-10-01merge revision(s) 63868,63870: [Backport #14897]nagachika
Fix a bug of peephole optimization ``` if L1 L0: jump L2 L1: ... L2: ``` was wrongly optimized to: ``` unless L2 L0: L1: ... L2: ``` To make it conservative, this optimization is now disabled when there is any label between `if` and `jump` instructions. Fixes [Bug #14897]. compile.c: remove unreachable jump only * compile.c (iseq_peephole_optimize): remove unreachable jump instruction only. if it is labeled and referred from other instructions, it is reachable and must not be removed. [ruby-core:87830] [Bug #14897] git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/branches/ruby_2_5@64893 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
2018-08-17merge revision(s) 63255: [Backport #14708]nagachika
compile.c: fix unconditional branch optimization * compile.c (iseq_peephole_optimize): add dummy `putnil` after a `jump` replacing an unconditional branch, to adjust removed `dup`. [ruby-core:86666] [Bug #14708] git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/branches/ruby_2_5@64419 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
2018-03-24merge revision(s) 61587,61617,61618: [Backport #14273]naruse
compile.c: next label * compile.c (compile_next): label for jump to the end of block is removable. compile.c: remove more unreachable chunk * compile.c (remove_unreachable_chunk): remove beyond labels to be removed. compile.c: fix stack consistency error * compile.c (iseq_peephole_optimize): fix stack consistency error from return in loop, by adding extra `pop` when replacing `jump` with `leave`, which is never reached but needed to adjust sp calculation. [ruby-core:84589] [Bug #14273] git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/branches/ruby_2_5@62911 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
2018-02-22merge revision(s) 62333: [Backport #14459]naruse
compile.c: popped hash * compile.c (compile_array): skip creating new hash if preceeding elements are popped all. [ruby-core:85486] [Bug #14459] git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/branches/ruby_2_5@62527 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
2018-02-19merge revision(s) 61711:naruse
compile.c (iseq_compile_each0): remove irrelevant tracecoverage This change removes tracecoverage instructions on a line that has any NODE but is non-significant, such as, just one literal. This fixes the following failure that occurs only when coverage is enabled: 1) Failure: TestISeq#test_to_a_lines [.../ruby/test/ruby/test_iseq.rb:56]: <[3, 4, 7, 9]> expected but was <[3, 4, 5, 6, 7, 8, 9]>. git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/branches/ruby_2_5@62481 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
2018-02-18Revert r62426naruse
it cause test failure git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/branches/ruby_2_5@62470 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
2018-02-16merge revision(s) 61617,61618: [Backport #14273]naruse
compile.c: remove more unreachable chunk * compile.c (remove_unreachable_chunk): remove beyond labels to be removed. compile.c: fix stack consistency error * compile.c (iseq_peephole_optimize): fix stack consistency error from return in loop, by adding extra `pop` when replacing `jump` with `leave`, which is never reached but needed to adjust sp calculation. [ruby-core:84589] [Bug #14273] git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/branches/ruby_2_5@62426 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
2018-02-16merge revision(s) 61603: [Backport #14279]naruse
check array for zsuper. [Bug #14279] * compile.c (iseq_compile_each0): for zsuper (NODE_ZSUPER), we need to check given argument is Array or not. * test/ruby/test_super.rb: add a test for this bug. git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/branches/ruby_2_5@62420 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
2017-12-25Set first line numbers for empty iseqs.ko1
* compile.c (iseq_compile_each): for empty method, block and so on, `last_line` is not set so that line number of `putnil` instruction will be zero. This patch set `first_lineno` for such `putnil`. Problem is reported by deivid-rodriguez via Yuichiro Kaneko. * test/ruby/test_iseq.rb: add a test for this spec. git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@61457 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
2017-12-24compile.c: toplevel return argumentsnobu
* compile.c (compile_return): evaluate arguments to top-level return but ignore the results. git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@61437 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
2017-12-23fix last commitko1
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@61428 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
2017-12-23RubyVM::InstructionSequence#trace_points.ko1
* iseq.c (iseqw_trace_points): add `RubyVM::InstructionSequence#trace_points` method for tools which want to manipulate ISeq (and traces). * test/ruby/test_iseq.rb: add a test for this method. git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@61427 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
2017-12-23revert line number spec of *return events.ko1
* compile.c (rb_iseq_compile_node): line number spec of :return, :b_return and :end events of 'TracePoint` is changed for [Feature #14104]. Quoted from [Feature #14104]: > Line numbers on :return/:b_return events show the last executed lines, > instead of end lines (without return statement). Note that :end event also affected. However, "buybug", a well-known ruby debugger depends on previous behavior so that I reverted this specification. * test/ruby/test_settracefunc.rb: catch up this fix. git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@61424 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
2017-12-21revert r61371 for [Bug #14214]ko1
Reverted patch doesn't check overwritten value. git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@61387 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
2017-12-20compile.c: branch on block paramnobu
* compile.c (compile_branch_condition): replace block param value in branch conditions with `defined(yield)`, to get rid of creating proc value. git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@61371 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
2017-12-20compile.c: add a RUBY_EVENT_COVERAGE_LINE event for line coveragemame
2.5's line coverage measurement was about two times slower than 2.4 because of two reasons; (1) vm_trace uses rb_iseq_event_flags (which takes O(n) currently where n is the length of iseq) to get an event type, and (2) RUBY_EVENT_LINE uses setjmp to call an event hook. This change adds a special event for line coverage, RUBY_EVENT_COVERAGE_LINE, and adds `tracecoverage` instructions where the event occurs in iseq. `tracecoverage` instruction calls an event hook without vm_trace. And, RUBY_EVENT_COVERAGE_LINE is an internal event which does not use setjmp. This change also cancells lineno change due to the deletion of trace instructions [Feature #14104]. So fixes [Bug #14191]. git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@61350 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
2017-12-19compile.c: pop splat resultnobu
* compile.c (compile_array): pop splat result without creating a new hash if the whole hash is popped. [ruby-core:84340] [Bug #14201] From: Nobuyoshi Nakada <nobu@ruby-lang.org> git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@61331 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
2017-12-19compile.c: side effect in splatnobu
* compile.c (compile_array): splat which may have side effects should be compiled even if the result will be popped. [ruby-core:84340] [Bug #14201] From: Nobuyoshi Nakada <nobu@ruby-lang.org> git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@61329 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
2017-12-18iseq.c (finish_iseq_build): fix coverage leakage [Bug #14191]mame
Before this change, coverage.so had failed to measure some multiple-line code fragments. This is because removing trace instructions (#14104) changed TracePoint's lineno (new lineno), and coverage counter array was based on old lineno. This change initializes coverage counter array based on new lineno. git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@61313 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
2017-12-14node.[ch], parse.y, compile.c: rename nd_column to nd_first_columnmame
nd_set_column -> nd_first_set_column nd_lineno -> nd_first_lineno nd_set_lineno -> nd_first_set_lineno git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@61224 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
2017-12-06compile.c (rb_iseq_compile_node): Move the check for imemo_ifunc to topmame
Applying nd_type to imemo_ifunc object seems harmless fortunately, but very dirty (to me). git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@61050 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
2017-12-06vm_core.h (RUBY_EVENT_COVERAGE_BRANCH): renamedmame
This change moves RUBY_EVENT_COVERAGE from include/ruby/ruby.h to vm_core.h and renames it to RUBY_EVENT_COVERAGE_BRANCH. git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@61049 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
2017-12-06insns.def (tracebranch): renamed from `trace2`mame
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@61047 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
2017-12-06thread.c (update_branch_coverage): renamed from `update_coverage`mame
Now this function only deals with branch events, so this change renames it and remove complexity that is no longer needed. git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@61046 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
2017-12-06thread.c (update_line_coverage): Use RUBY_EVENT_LINEmame
This change makes coverage use the general event type RUBY_EVENT_LINE instead of a special event type RUBY_EVENT_COVERAGE. Just a refactoring. git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@61043 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
2017-12-05vm_core.h (rb_iseq_locatoin_t): add a field `code_range`mame
This change makes each ISeq keep NODE's code range. This information is needed for method coverage. git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@61025 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
2017-12-05Revamp method coverage to support define_methodmame
Traditionally, method coverage measurement was implemented by inserting `trace2` instruction to the head of method iseq. So, it just measured methods defined by `def` keyword. This commit drastically changes the measuring mechanism of method coverage; at `RUBY_EVENT_CALL`, it keeps a hash from rb_method_entry_t* to runs (i.e., it counts the runs per method entry), and at `Coverage.result`, it creates the result hash by enumerating all `rb_method_entry_t*` objects (by `ObjectSpace.each_object`). git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@61023 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
2017-11-18introduce `trace_events' info for iseq.ko1
* vm_core.h (rb_iseq_t::aux): add `trace_events` which represents which events are enabled on this iseq. With this information, we can skip useless trace-on changes for ISeqs. * vm_trace.c (RUBY_EVENTS_TRACE_BY_ISEQ): moved to iseq.h and rename it with ISEQ_TRACE_EVENTS. * iseq.h: introduce ISEQ_USE_COMPILE_DATA iseq (imemo) flag to represent COMPILE_DATA is available. In other words, iseq->aux.trace_events is not available when this flag is set. * ISEQ_COMPILE_DATA() is changed from a macro. * ISEQ_COMPILE_DATA_ALLOC() is added. * ISEQ_COMPILE_DATA_CLEAR() is added. * iseq.c: use them. git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@60838 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
2017-11-16Refactoring out the direct accesses of NODE's u1, u2, and u3mame
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@60812 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
2017-11-16Add a last location to branch coverageyui-knk
* compile.c (DECL_BRANCH_BASE, ADD_TRACE_BRANCH_COVERAGE): Add a last location to arguments. * compile.c (compile_if, compile_case, compile_case2, compile_loop, iseq_compile_each0): Pass a last location to macros. * ext/coverage/coverage.c (branch_coverage): Add a last location to a return value. * test/coverage/test_coverage.rb: Follow-up these changes. git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@60785 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
2017-11-15fix uninitialized memory reference.ko1
* compile.c (iseq_set_sequence): clear kwargs (in ci_entries) memory area. kwargs ci entries are initialized by compiler. However, sometimes these initializations are skipped because corresponding calls are eliminated by some optimizations (for example, `if true` syntax elimnates else code). git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@60771 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
2017-11-14compile.c: fixup TRACEnobu
* compile.c (remove_unreachable_chunk): ignore TRACE elements. git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@60768 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
2017-11-14fix prefix.ko1
* compile.c: the following functions accept LINK_ELEMENT so rename to `ELEM_` prefix names: * INSERT_ELEM_NEXT -> ELEM_INSERT_NEXT * INSERT_ELEM_PREV -> ELEM_INSERT_PREV * REPLACE_ELEM -> ELEM_REPLACE * REMOVE_ELEM -> ELEM_REMOVE git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@60764 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
2017-11-14remove `trace` instruction. [Feature #14104]ko1
* tool/instruction.rb: create `trace_` prefix instructions. * compile.c (ADD_TRACE): do not add `trace` instructions but add TRACE link elements. TRACE elements will be unified with a next instruction as instruction information. * vm_trace.c (update_global_event_hook): modify all ISeqs when hooks are enabled. * iseq.c (rb_iseq_trace_set): added to toggle `trace_` instructions. * vm_insnhelper.c (vm_trace): added. This function is a body of `trace_` prefix instructions. * vm_insnhelper.h (JUMP): save PC to a control frame. * insns.def (trace): removed. * vm_exec.h (INSN_ENTRY_SIG): add debug output (disabled). git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@60763 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
2017-11-14compile.c: comments for concatstrings optimization [ci skip]nobu
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@60759 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
2017-11-13compile.c: fixup r60727nobu
* compile.c (iseq_peephole_optimize): skip next `freezestring` instruction after `concatstrings` instruction when frozen string literal is enabled. git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@60756 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
2017-11-09compile.c: optimize nested string interpolationsnobu
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@60727 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
2017-11-09insn_info/insns_infoko1
* iseq.h (iseq_line_info_entry): rename to iseq_insn_info_entry. * vm_core.h (rb_iseq_constant_body): rename field name line_info_table to insns_info and also from line_info_size to insns_info_size. * compile.c (INSN): add struct insn_info to contain per insn information. * compile.c (add_insn_info): added to add new insn_info entry. git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@60726 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
2017-11-09fix peephole optimization.ko1
* compile.c (iseq_peephole_optimize): do not need to put `pop` instruction. * test/ruby/test_optimization.rb (test_peephole_optimization_without_trace): This code "def foo; 1.times{|(a), &b| nil && a}; end" fails to compile by stack underflow because of above bug (fixed by this patch). git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@60724 b2dd03c8-39d4-4d8f-98ff-823fe69b080e