aboutsummaryrefslogtreecommitdiffstats
path: root/lib/compiler/src/beam_validator.erl
AgeCommit message (Collapse)Author
2019-05-27compiler: Fix broken type for erlang:'++'/2John Högberg
2019-04-23beam_validator: Don't infer types for dead valuesJohn Högberg
2019-03-25beam_validator: Remove uncovered lines in lists_mod_return_type/3Björn Gustavsson
Don't bother infering the return types for lists function that beam_ssa_type does not handle.
2019-03-18beam_validator: Infer types on both sides of '=:='John Högberg
2019-03-06beam_validator: Express test_arity/is_tagged_tuple as type testsJohn Högberg
This ensures that unreachable branches are properly ignored on repeated checks (although tuple type subtraction isn't complete yet).
2019-03-06beam_validator: Fix type subtraction on select_* and inequalityJohn Högberg
Type subtraction never resulted in the 'none' type, even when it was obvious that it should. Once that was fixed it became apparent that inequality checks also fell into the same subtraction trap that the type pass warned about in a comment. This then led to another funny problem with select_val, consider the following code: {bif,'>=',{f,0},[{x,0},{integer,1}],{x,0}}. {select_val,{x,0},{f,70},{list,[{atom,false},{f,69}, {atom,true},{f,68}]}}. The validator knows that '>=' can only return a boolean, so once it has subtracted 'false' and 'true' it killed the state because all all valid branches had been taken, so validation would crash once it tried to branch off the fail label.
2019-03-05beam_validator: Style fix for asserts in list comprehensionsJohn Högberg
2019-03-05beam_validator: Infer types from bs_put instructionsJohn Högberg
2019-03-05beam_validator: is_map no longer needs special treatmentJohn Högberg
2019-03-05beam_validator: Refactor type conflict resolutionJohn Högberg
The current type conflict resolution works well for the example case in the comment, but doesn't handle branched code properly, consider the following: {label,2}. {test,is_tagged_tuple,{f,ignored},[{x,0},3,{atom,r}]}. {allocate_zero,2,1}. {move,{x,0},{y,0}}. %% {y,0} is known to be {r, _, _} now. {get_tuple_element,{x,0},2,{x,0}}. {'try',{y,1},{f,3}}. %% ... snip ... {jump,{f,5}}. {label,3}. {try_case,{y,1}}. %% {x,0} is the error class (an atom), {x,1} is the error term. {test,is_eq_exact,{f,ignored},[{x,0},{y,0}]}. %% ... since tuples and atoms can't meet, the type of {y,0} is %% now {atom,[]} because the current code assumes the type %% we're updating with. {move,{x,1},{x,0}}. {jump,{f,5}}. {label,5}. %% ... joining tuple (block 2) and atom (block 3) means 'term', %% so the get_tuple_element instruction fails to validate %% despite this being unrechable from block 3. {test_heap,3,1}. {get_tuple_element,{y,0},1,{x,1}}. {put_tuple2,{x,0},{list,[{x,1},{x,0}]}}. {deallocate,2}. return. This commit kills the state on type conflicts, making unreachable instructions truly unreachable.
2019-03-05beam_validator: Refactor branch handlingJohn Högberg
While complex_test made certain branching instructions a lot easier to read, we're still using `branch_state` for many others which is hard to read and makes it impossible to "abort" branches on type conflicts. This commit replaces nearly all uses of `branch_state` with a general branching mechanism, improving readability and paving the way for proper type conflict resolution.
2019-03-04beam_validator: Fix element/2 BIF handlingJohn Högberg
The element type can not be extracted before the tuple type has been updated.
2019-03-04beam_validator: Rename get_move_term_type and clean up get_raw_typeJohn Högberg
2019-02-27beam_validator: Clarify a commentJohn Högberg
2019-02-27beam_validator: Make call argument validation stricterJohn Högberg
We used to cheat by checking if it were possible to meet the Given and Required types, which caught the most common problems but potentially let tuple element conflicts pass through. This was a compromise to let the thing "work" while we were refactoring the validator, but we can be a lot stricter now that its type tracking capabilities approach those of the type optimization pass.
2019-02-27beam_validator: Don't explode when building terms in receiveJohn Högberg
Building terms with fragile contents is okay because the GC is disabled during loop_rec, and the resulting term won't be reachable from the root set afterwards. ERL-862
2019-02-27beam_validator: Improve 'binary' type trackingJohn Högberg
2019-02-27beam_validator: Infer tuple element typesJohn Högberg
This is possible now that we track types on a per-value basis, and no longer need to care whether the source tuple's register has been clobbered by the time we infer the type.
2019-02-27beam_validator: Tolerate the 'receive' hack in prim_evalJohn Högberg
2019-02-27beam_validator: Track types by value rather than by registerJohn Högberg
This is a rather subtle but important distinction. While tracking types on a per-register basis is fairly effective, it forces us to track which registers alias each other, and makes it tricky to infer types over large blocks of code as instruction arguments may have been clobbered between definition and inference. Tracking types on a per-value basis makes us immune to these problems.
2019-02-27beam_validator: Disregard 'none' on joinJohn Högberg
2019-02-27beam_validator: Handle is_number, and join(float,int) -> numberJohn Högberg
I have no idea how this escaped us for so long...
2019-02-27beam_validator: Treat is_nil as is_eq_exact with nilJohn Högberg
2019-02-27beam_validator: Simplify get_element_typeJohn Högberg
2019-02-27beam_validator: Fix literal handling in meet/2John Högberg
2019-02-27beam_validator: Use literals as keys in container (tuple) elementsJohn Högberg
2019-02-27beam_validator: Refactor try/catch handling, againJohn Högberg
2019-02-27beam_validator: Refactor register initializationJohn Högberg
2019-02-27beam_validator: Refactor stack allocationJohn Högberg
2019-02-26beam_validator: Handle argument/return types for more functionsJohn Högberg
2019-02-26beam_validator: Don't forget last element when using put_tupleJohn Högberg
2019-02-21beam_ssa_type: Use types from some 'lists' functionsBjörn Gustavsson
This commit lets the compiler know about the return type of some of the functions in the `lists` module. Knowing about the return will allow the compiler to emit fewer type test instructions, and also fewer instructions for throwing `case_clause` or `badmatch` exceptions, thus producing slightly faster and more compact code. This change makes the `lists` module a part of the language, but it could be argued that it already is because several functions (e.g. `member/2` and `keymember/3`) are implemented in as BIFs in the runtime system. Therefore, a user cannot simply change the `lists` module and expect everything to continue working as before. The compiler will now know the return types for the following functions: all/2 any/2 keymember/3 member/2 prefix/2 suffix/2 dropwhile/2 duplicate/2 filter/2 flatten/1 map/2 mapfoldl/3 mapfoldr/3 partition/2 reverse/1 sort/1 splitwith/1 takewhile/1 unzip/1 usort/1 zip/2 zipwith/3
2019-02-21beam_validator: Refactor call argument validationJohn Högberg
2019-02-21beam_validator: Refactor liveness/stack initialization checksJohn Högberg
2019-02-21beam_validator: Refactor try/catch handlingJohn Högberg
2019-02-20beam_validator: Remember definitions on assignmentJohn Högberg
2019-02-20beam_validator: Refactor stack trimmingJohn Högberg
2019-02-20beam_validator: Track definitions of all termsJohn Högberg
2019-02-20beam_validator: Remove special handling of map_get/is_map_keyJohn Högberg
Neither can be used for type subtraction, so the default BIF handler suits them just fine.
2019-02-20beam_validator: Refactor select_tuple_arityJohn Högberg
2019-02-20beam_validator: Treat select_val as a series of '=:='John Högberg
2019-02-20beam_validator: Treat all bs_get instructions as extractionsJohn Högberg
While this is strictly only relevant for bs_get_binary2, we should never build anything while matching a message, so it ought to be safe to remove this last raw use of propagate_fragility.
2019-02-20beam_validator: Separate BIF/call types more clearlyJohn Högberg
2019-02-20beam_validator: Assert that no tuple elements are out of boundsJohn Högberg
2019-02-20beam_validator: Get rid of the last uses of set_aliased_typeJohn Högberg
Granted, it's replaced with a thin wrapper, but it'll simplify migration to the new type format.
2019-02-20beam_validator: Minor cosmetic refactoringJohn Högberg
2019-02-18beam_validator: Infer types from result of all type test BIFsJohn Högberg
The compiler will usually optimize these into test instructions, but they'll nevertheless pop up in some cases.
2019-02-18beam_validator: Infer BIF argument typesJohn Högberg
If we know that a BIF will badarg unless its arguments have certain types, we can infer that we have at least those types on success. Note that we can't do this in the general case as the BIF could fail for reasons other than bad arguments.
2019-02-08beam_validator: type_test in BIFs that only fail on invalid typesJohn Högberg
2019-02-08beam_validator: Simplify complex branchesJohn Högberg
Branches where the state is altered on both success and failure are hard to follow and require juggling the current state, so this commit adds a helper function to make it easier.