aboutsummaryrefslogtreecommitdiffstats
path: root/lib/compiler
AgeCommit message (Collapse)Author
2015-09-14Merge branch 'maint'Anders Svensson
2015-09-14Merge branch 'bjorn/cuddle-with-tests' into maintBjörn Gustavsson
* bjorn/cuddle-with-tests: (23 commits) rand_SUITE: Speed up basic_stats/1 base64_SUITE: Speed up roundtrip/1 lists_SUITE: Test lists:concat/2 lists_SUITE: Test lists:split/2 lists_SUITE: Add a test case for lists:prefix/2 lists_SUITE: Add hof/1 to test all high-order functions lists_SUITE: Add test for lists:takewhile/1 lists_SUITE: Run test cases in each group in parallel lists_SUITE: Test lists:keyreplace/4 lists_SUITE: Extend flatten/1 test to also test flatlength/1 lists_SUITE: Correct test of lists:flatten/2 id_transform_SUITE: Modernize test suite io_proto_SUITE: Speed up determination of default shell io_proto_SUITE: Refactor up rtnode() and friends gen_event_SUITE: Remove unnecessary sleep calls proc_lib: Improve coverage for crash/1 proc_lib_SUITE: Eliminate compiler warnings io_SUITE: Add coverage/1 to completely cover io_lib_pretty io_SUITE: Extend coverage of code for testing printable chars io_SUITE: Speed up test for bad +pc option ...
2015-09-11Merge branch 'maint'Zandra
2015-09-11Merge branch 'c-rack/fix-typo3' into maintZandra
* c-rack/fix-typo3: Fix typo in call_last/3 spec Fix typo Fix typo: message to send is in x(1) not x(0) Fix another small typo Fix typo
2015-09-10compile_SUITE: Add test of warningsBjörn Gustavsson
Make sure that all warnings produced when compiling the test suite contains filenames and line numbers.
2015-09-07Merge branch 'maint'Björn-Egil Dahlberg
2015-09-04compiler: Fix get_map_elements register corruptionBjörn-Egil Dahlberg
Instruction get_map_elements might destroy target registers when the fail-label is taken. Only seen for patterns with two, and only two, target registers. Specifically: we copy one register, and then jump. foo(A,#{a := V1, b := V2}) -> ... foo(A,#{b := V}) -> ... call foo(value, #{a=>whops, c=>42}). corresponding assembler: {test,is_map,{f,5},[{x,1}]}. {get_map_elements,{f,7},{x,1},{list,[{atom,a},{x,1},{atom,b},{x,2}]}}. %% if 'a' exists but not 'b' {x,1} is overwritten, jump {f,7} {move,{integer,1},{x,0}}. {call_only,3,{f,10}}. {label,7}. {get_map_elements,{f,8},{x,1},{list,[{atom,b},{x,2}]}}. %% {x,1} (src) is read with a corrupt value {move,{x,0},{x,1}}. {move,{integer,2},{x,0}}. {call_only,3,{f,10}}. The fix is to remove 'opt_moves' pass for get_map_elements instruction in the case of two or more destinations. Reported-by: Valery Tikhonov
2015-09-04compiler: Add extra checks for get_map_elements in validatorBjörn-Egil Dahlberg
2015-09-04compiler: Add test for corrupt register in get_map_elementsBjörn-Egil Dahlberg
2015-08-28Fix typo in call_last/3 specConstantin Rack
2015-08-25Fix typoConstantin Rack
2015-08-24Fix typo: message to send is in x(1) not x(0)Constantin Rack
2015-08-21beam_validator: Don't allow x(1023) to be usedBjörn Gustavsson
In 45f469ca0890, the BEAM loader started to use x(1023) as scratch register for some instructions. Therefore we should not allow x(1023) to be used in code emitted by the compiler.
2015-08-21v3_core: Improve code generation for guardsBjörn Gustavsson
When translating guards to Core Erlang, it is sometimes necessary to add an is_boolean/1 guard test. Here is an example when it is necessary: o(A, B) when A or B -> ok. That would be translated to something like: o(A, B) when ((A =:= true) or (B =:= true)) and is_boolean(A) and is_boolean(B) -> ok. The is_boolean/1 tests are necessary to ensure that the guard fails for calls such as: o(true, not_boolean) However, because of a bug in v3_core, is_boolean/1 tests were added when they were not necessary. Here is an example: f(B) when not B -> ok. That would be translated to: f(B) when (B =:= false) and is_boolean(B) -> ok. The following translation will work just as well. f(B) when B =:= false -> ok. Correct the bug to suppress those unnecessary is_boolean/1 tests.
2015-08-21Move rewriting of select_val to is_boolean from beam_peep to beam_deadBjörn Gustavsson
We can rewrite more instances of select_val to is_boolean because it is not necessary that a particular label follows the select_val.
2015-08-21Put 'try' in blocks to optimize allocation instructionsBjörn Gustavsson
Put 'try' instructions inside block to improve the optimization of allocation instructions. Currently, the compiler only looks at initialization of y registers inside blocks when determining which y registers that will be "naturally" initialized.
2015-08-21Reorder instructions across try/catchBjörn Gustavsson
Simplify further optimizations by moving safe instructions to before the 'try' or 'catch' instruction.
2015-08-21Delay get_tuple_element instructions until they are neededBjörn Gustavsson
When matching tuples, the pattern matching compiler would generate code that would fetch all elements of the tuple that will ultimately be used, *before* testing that (for example) the first element is the correct record tag. For example: is_tuple Fail {x,0} test_arity Fail {x,0} 3 get_tuple_element {x,0} 0 {x,1} get_tuple_element {x,0} 1 {x,2} get_tuple_element {x,0} 2 {x,3} is_eq_exact Fail {x,1} some_tag If {x,2} and {x,3} are not used at label Fail, we can re-arrange the code like this: is_tuple Fail {x,0} test_arity Fail {x,0} 3 get_tuple_element {x,0} 0 {x,1} is_eq_exact Fail {x,1} some_tag get_tuple_element {x,0} 1 {x,2} get_tuple_element {x,0} 2 {x,3} Doing that may be beneficial in two ways. If the branch is taken, we have eliminated the execution of two unnecessary instructions. Even if the branch is never or rarely taken, there is the possibility for more optimizations following the is_eq_exact instructions. For example, imagine that the code looks like this: get_tuple_element {x,0} 1 {x,2} get_tuple_element {x,0} 2 {x,3} move {x,2} {y,0} move {x,3} {y,1} Assuming that {x,2} and {x,3} have no further uses in the code that follows, that can be rewritten to: get_tuple_element {x,0} 1 {y,0} get_tuple_element {x,0} 2 {y,1} When should we perform this optimization? At the very latest, it must be done before opt_blocks/1 in beam_block which does the elimination of unnecessary moves. Actually, we want do the optimization before the blocks have been established, since moving instructions out of one block into another is cumbersome. Therefore, we will do the optimization in a new pass that is run before beam_block. A new pass will make debugging easier, and beam_block already has a fair number of sub passes.
2015-08-21Optimize get_tuple_element instructions by moving them forwardBjörn Gustavsson
2015-08-21beam_block: Improve the move optimizationsBjörn Gustavsson
Here is an example of a move instruction that could not be optimized away because the {x,2} register was not killed: get_tuple_element Reg Pos {x,2} . . . move {x,2} {y,0} put_list {x,2} nil Any We can do the optimization if we replace all occurrences of the {x,2} register as a source with {y,0}: get_tuple_element Reg Pos {y,0} . . . put_list {y,0} nil Dst
2015-08-21beam_block: Clean up optimization of move optimizationsBjörn Gustavsson
The 'move' optimization was relatively clean until GC BIFs were introduced. Instead of re-thinking the implementation, the existing code was fixed and patched. The current code unsuccessfully attempts to eliminate 'move' instructions across GC BIF and allocation instructions. We can simplify the code if we give up as soon as we encounter any instruction that allocates.
2015-08-21beam_block: Eliminate redundant wasteful call to opt/1Björn Gustavsson
opt_alloc/1 makes a redundant call to opt/1. It is redundant because the opt/1 function has already been applied to the instruction sequence prior to calling opt_alloc/1.
2015-08-09Fix another small typoConstantin Rack
2015-08-09Fix typoConstantin Rack
2015-08-06Teach the compiler the 'da' and 'dz' optionsBjörn Gustavsson
Add the 'da' option to create a list after the beam_a pass. Seeing how the code looks after beam_a, but before the blocks have been established, is sometimes useful. For symmetry, add the 'dz' option, even though it is just a synonym for 'S'.
2015-06-23Prepare releaseErlang/OTP
2015-06-22fix errors caused by changed line numbersBruce Yinhe
2015-06-18Change license text to APLv2Bruce Yinhe
2015-06-17compiler: Add regressions_SUITEBjörn-Egil Dahlberg
regressions_SUITE will have code snippets which previously crashed the compiler. This commits includes a test for Maps crash in beam_bool.
2015-06-17compiler: Fix beam_bool pass for get_map_elementsBjörn-Egil Dahlberg
Before beam_split the get_map_elements instruction is still in blocks and the helper function in beam_jump did not reflect this. Reported-by: Quviq twitter account
2015-06-04Merge branch 'bjorn/compiler/spurious-warning'Björn Gustavsson
* bjorn/compiler/spurious-warning: sys_core_fold: Eliminate warnings for unused terms in effect context sys_core_fold: Eliminate warnings for unused terms
2015-05-27Merge branch 'bjorn/compiler/doc'Björn Gustavsson
* bjorn/compiler/doc: Update compiler documentation (Sneaking in OTP-12769 here which is a release note for syntax_tools. Sorry about that.)
2015-05-27Update compiler documentationBjörn Gustavsson
Language cleaned up by the technical writer tmanevik from Combitech. Proofreading and corrections by Björn Gustavsson.
2015-05-27Update application versionsHans Bolinder
2015-05-26Merge branch 'egil/opt-compile-time/OTP-12774'Björn-Egil Dahlberg
* egil/opt-compile-time/OTP-12774: stdlib: Relax erl_anno_SUITE:is_anno/1 test Update primary bootstrap compiler: Use Maps as type information compiler: Use Maps instead of dict in beam_jump compiler: Use cerl_sets instead of gb_sets in beam_type compiler: Use Maps instead of gb_trees in beam_dead compiler: Use cerl_sets instead of gb_sets in beam_jump compiler: Use cerl_sets instead of sets in v3_kernel compiler: Use cerl_sets instead of gb_sets in sys_core_fold compiler: Add cerl_sets module compiler: Scope uses gb_sets not gb_trees beam_dict: Use Maps to map function name indices beam_dict: Use Maps to map line indices beam_dict: Use Maps to map atom indices v3_codegen: Use Maps to map local functions v3_life: Refactor variable db compiler: Use lc instead of map/1 in v3_codegen stdlib: Optimize erl_anno:is_string/1 Conflicts: bootstrap/lib/kernel/ebin/inet_dns.beam bootstrap/lib/stdlib/ebin/erl_anno.beam bootstrap/lib/stdlib/ebin/erl_lint.beam
2015-05-22sys_core_fold: Eliminate warnings for unused terms in effect contextBjörn Gustavsson
The optimization introduced in 0a0d39d351fc could cause spurious warnings of the type: "a term is constructed, but never used". That would happen for constructs in effect context. To avoid those warnings, we will need to apply warning suppression also in effect context.
2015-05-22Merge branch 'egil/fix-compiler-beam_bsm/OTP-12758'Björn-Egil Dahlberg
* egil/fix-compiler-beam_bsm/OTP-12758: compiler: Add tests for beam_bsm get_map_elements compiler: Teach beam_bsm get_map_elements instruction
2015-05-21compiler: Use Maps as type informationBjörn-Egil Dahlberg
Using Maps as type information container speedups files like cow_http_hd.erl by ~500ms. Previously spent ~60% of the time in orddict:store/3.
2015-05-21compiler: Use Maps instead of dict in beam_jumpBjörn-Egil Dahlberg
2015-05-21compiler: Use cerl_sets instead of gb_sets in beam_typeBjörn-Egil Dahlberg
2015-05-21compiler: Use Maps instead of gb_trees in beam_deadBjörn-Egil Dahlberg
2015-05-21compiler: Use cerl_sets instead of gb_sets in beam_jumpBjörn-Egil Dahlberg
2015-05-21compiler: Use cerl_sets instead of sets in v3_kernelBjörn-Egil Dahlberg
2015-05-21compiler: Use cerl_sets instead of gb_sets in sys_core_foldBjörn-Egil Dahlberg
2015-05-21compiler: Add cerl_sets moduleBjörn-Egil Dahlberg
A sets implementation based on maps.
2015-05-21compiler: Scope uses gb_sets not gb_treesBjörn-Egil Dahlberg
2015-05-21beam_dict: Use Maps to map function name indicesBjörn-Egil Dahlberg
2015-05-21beam_dict: Use Maps to map line indicesBjörn-Egil Dahlberg
2015-05-21beam_dict: Use Maps to map atom indicesBjörn-Egil Dahlberg
2015-05-21v3_codegen: Use Maps to map local functionsBjörn-Egil Dahlberg