| Age | Commit message (Collapse) | Author |
|---|
|
* bg/compiler-cover-and-clean:
v3_life: Remove clause that cannot match in match_fail/3
v3_life tests: Cover exception handling code in v3_life:function/1
beam_type: Remove redundant clause
v3_core tests: Cover make_bool_switch_guard/5
v3_core tests: Cover handling of pattern aliases
v3_core: Remove a clause in is_simple/1 that cannot match
v3_core: Remove unused support for generating compilation errors
Remove stray support for the put_literal/2 instruction
Remove stray support for the bs_bits_to_bytes2/2 instruction
Remove the bs_bits_to_bytes/3 instruction
Cover handling of 'math' BIFs
beam_bool: Remove a clause in live_regs/1 that cannot match
beam_bool: Cover handling of bs_context_to_binary in initialized_regs/2
beam_bool: Remove a clause in initialized_regs/2 that cannot match
beam_block: Remove a clause that will never be executed
erts: Stop supporting non-literal empty tuples
compile: Remove code that is only executed on Solaris
Do not cover-analyze core_scan
core_SUITE_data: Don't ignore *.core files in this directory
OTP-8636 bg/compiler-cover-and-clean
|
|
The clause that attempts to match a #k_literal{} where
the value is an atom cannot possibly match, since single
atoms are always encoded as #k_atom{}.
|
|
The clause does not server any useful purpose, since it does the
same as the default clause at the end.
|
|
The is_simple/1 function does not need to handle #c_binary{},
since it will never be called with a #c_binary{} argument.
|
|
The state record #core{} has a "es" field that can be used for
causing a compilation error. Since incorrect programs have been
rejected much earlier by erl_lint, no errors were actually stored
in the "es" fields. Therefore, get rid of the "es" field and
the support for generating a compilation error in the v3_core
module.
|
|
put_literal/2 was an experimental instruction added in R11 to
support literals, but before the R12 release support for literals
was implemented for all instruction, making the put_literal/2
instruction redundant. Although the beam_disasm module supports
dissambley of instructions in older releases, there is
no reason to have it support experimental instructions.
|
|
bs_bits_to_bytes2/2 was an experimental instruction added in R11,
but was removed in R12. Although the beam_disasm and beam_validator
modules do support instructions in older releases, there is
no reason to have them support experimental instructions.
|
|
The last compiler to generate code that uses the bs_bits_to_bytes/3
instruction was the R11 compiler. Since we don't support loading
R11 *.beam files in R14, removing the remaining support for the
instruction.
|
|
live_regs/1 folds over a list of tuples, so a clause matching
an element being an empty list can never match.
If the list for some unfathomable reason would contain an empty
list, there will be an internal compiler error and no *.beam file
will be created. Thus this change is safe.
|
|
When scanning a reversed instruction sequence for a function,
it is impossible to reach the end of the list, because each function
must have a fun_info/4 instruction followed by a label/1 instruction
at the beginning, and there is a clause that will handle those
instructions.
If for some unfathomable reason the end of the list would be
reached, with this change there will be an internal compiler error
and no *.beam file will be created. Thus this change is safe.
|
|
Since the introduction of literals in R12B, empty tuples
are literals. Thus the put_tuple/2 instruction is always
followed by at least one put/1 instruction. Therefore
the alloc_may_pass/1 function in beam_block no longer needs
a clause for for "put_tuple", because the clause for "put"
will always be reached first (since the instruction stream
is scanned in reverse execution order).
Note that if the compiler would generate a "put_tuple 0 Dst"
instruction for some unfathomable reason, we should still be
because the run-time system will now refuse to load a module
containing such an instruction.
|
|
On Solaris only, the 'time' option would include the process
size for the Erlang virtual machine in its output. Since few
compiler developers use Solaris nowadays, remove the code.
|
|
* pan/otp_8217_binary:
Add documentation for binary module
Add more tests and make some go easier on small systems
Correct Boyer More and trapping for longest_common_suffix
Add longer timetrap to testcases and add binary to app file
Add guard BIFs binary_part/2,3
Add binary:{encode,decode}_unsigned({1,2}
Add referenced_byte_size/1
Add binary:list_to_bin/1 and binary:copy/1,2
Add bin_to_list/{1,2,3}
Add binary:longest_common_prefix/longest_common_suffix
Add binary:part to erl_bif_binary.c
Move binary module bif's to erl_bif_binary.c
Count reductions for process even when not trapping
Add random compare testcase
Teach BIF's binary:match/matches interrupting/restarting
Teach binary.c the semantics to take longest instead of shortest match
Initial commit of the binary EEP
OTP-8217 Implement EEP31
The module binary from EEP31 (and EEP9) is implemented.
|
|
Add the gc_bif's to the VM.
Add infrastructure for gc_bif's (guard bifs that can gc) with two and.
three arguments in VM (loader and VM).
Add compiler support for gc_bif with three arguments.
Add compiler (and interpreter) support for new guard BIFs.
Add testcases for new guard BIFs in compiler and emulator.
|
|
* bg/compiler-attributes:
Remove opaque declarations from the attributes
|
|
-opaque declarations should not be retained in the attributes
(because they will be loaded along with the code and are not
useful).
While at it, filter away those Dialyzer attributes as early
as possible - in v3_kernel.
|
|
* bg/opt-receive:
Test that gen_server:call/2,3 are fast even with a huge message queue
erts: Add tests for the receive optimization
Update primary bootstrap
erts: Implement recv_mark/1 and recv_set/1 for real
compiler tests: Cover the error handling code in beam_receive
compiler test: Test optimization of receive statements
Optimize selective receives in the presence of a large message queue
Introduce the new recv_mark/1 and recv_mark/1 instructions
Compile tests that communicate with R12 nodes with the r12 option
Move p_run/2 to test_lib
gen: Inline wait_resp_mon/2 to help the compiler optimize
OTP-8623 bg/opt-receive
reveive statements that can only read out a newly created reference are now
specially optimized so that it will execute in constant time regardless of
the number of messages in the receive queue for the process. That
optimization will benefit calls to gen_server:call(). (See gen:do_call/4
for an example of a receive statement that will be optimized.)
|
|
If a gen_server process has many messages in its message queue and
calls another gen_server process, the selective receive in
gen_server:call() will have to go through the entire message queue.
Have the compiler generate the new mark_recv/1 and mark_recv/1
instructions that can avoid going through the entire message queue.
|
|
Make the recv_mark/1 and recv_mark/1 instructions known to the
compiler and run-time system. For the moment, make the loader ignore
any occurrences of those instructions in BEAM files.
Also update hipe_beam_to_icode to ignore those instructions.
|
|
* bg/compiler-suppress-result-ignored:
compiler tests: Eliminate "result of expression is ignored" warnings
Silence warnings for expressions that are assigned to "_"
OTP-8602 bg/compiler-suppress-result-ignored
It is now possible to suppress the warning in code such as
"list_to_integer(S), ok" by assigning the ignored value "_" like this: "_ =
list_to_integer(S), ok".
|
|
The following instruction sequence:
fmove {fr,Fr} {x,TempXreg}
move {x,TempXreg} {y,Dest}
is rewritten to:
fmove {fr,Fr} {y,Dest}
(Provided that {x,TempXreg} is killed by the instructions following
the sequence.)
Generalize the optimization to also handle:
fmove {fr,Fr} {x,TempXreg}
move {x,TempXreg} {_,Dest}
That is, the destination register can be either an X or Y register.
|
|
There is currently no zero-cost way to silence the warning
"the result of the expression is ignored", which is issued
for code such as:
list_to_integer(S),
ok
Such code can be useful for assertions or input validation.
Teach the compiler to silence the warning for expressions
that are explicitly assigned to to the "_" variable,
such as:
_ = list_to_integer(S),
ok
Implement it by having the v3_core pass annotate calls in
Core Erlang like this:
let <_> = ( call 'erlang':'list_to_integer'(S) -| ['result_not_wanted'] )
in 'ok'
and modifiy sys_core_fold to suppress the warning for any call
having the annotation.
We deliberately do not make it possible to silence the warnings
for expressions like:
{build,an,unnecessary,term}, ok
or
is_list(L), ok
because we don't know of any real-world scenarios in which that would
be useful.
|
|
* bg/compiler-inliner:
pmod_SUITE: Again test inlining parameterized modules
compiler tests: Cope with missing args in function_clause for native code
compiler tests: Compile a few more modules with 'inline'
Consistently rewrite an inlined function_clause exception to case_clause
compiler tests: Test the 'inline' option better
compiler: Suppress bs_context_to_binary/1 for a literal operand
compiler: Fix binary matching bug in the inliner
sys_core_inline: Don't generated multiple compiler_generated annos
OTP-8552 bg/compiler-inliner
Several problems in the inliner have been fixed.
|
|
A function_clause exception is generated by jumping to a func_info/3
instruction at the beginning of the function. The x registers are
assumed to contain the arguments for the function. That means
that a func_info/3 instruction copied from another function
(or even from the same function if not at the top level) will
not work, so it must be replaced with an instruction that
generates a case_clause exception.
In Core Erlang, a func_info/3 instruction is represented as
a the primop match_fail({function_clause,Arg1,...ArgN}).
The current mechanism that is supposed to replace the
primop match_fail(function_clause) with match_fail(case_clause)
will fail to do that in the following circumstances:
1. If the inliner has inlined a function into itself.
Fix that by having the inliner clear the function_name annotations
on all match_fail primops in functions that are inlined. (To simplify
doing that, the annotation is now on the primop node itself and not on
the 'function_clause' atom inside it.)
2. If the inliner has rewritten the tuple node in the primop node
to a literal (when inlining a function call with literal arguments),
v3_kernel would not recognize the match_fail(function_clause) primop
and would not rewrite it. Fix it by making v3_kernel smarter.
Also simplify the "old" inliner (sys_core_inline) to only clear
the function_name annotations instead of rewriting function_clause
execptions to case_clause execptions itself.
|
|
The inliner can cause illegal uses of the bs_context_to_binary/1
instruction, such as:
bs_context_to_binary {literal,"string"}
or
bs_context_to_binary {integer,1}
Remove the bs_context_to_binary/1 instruction if the operand
is not a register (it is clearly not needed).
|
|
The inliner incorrectly assumes that a literal cannot
match a binary in code such as:
t() ->
bc(<<"string">>).
bc(<<A:1,T/bits>>) -> [A|bc(T)];
bc(<<>>) -> [].
The bug was introduced when binary literals were introduced.
|
|
Multiple compiler_generated annotations are harmless, but makes
listing files harder to read during debugging.
|
|
* bg/compiler-remove-r11-support:
compiler: Don't support the no_binaries option
erts: Don't support the put_string/3 instruction
compiler: Don't support the no_constant_pool option
compiler: Don't support the r11 option
test_server: Don't support communication with R11 nodes
binary_SUITE: Don't test bit-level binary roundtrips with R11 nodes
erts: Test compatibility of funs with R12 instead of R11
OTP-8531 bg/compiler-remove-r11-support
|
|
The no_binaries option terminates the compiler with an error
if any bit syntax is used in the module being compiled.
(It used to be implied by the removed r11 option.)
|
|
Since R14 does not need to load code that can also be loaded
in an R11 run-time system, support for the put_string/3
instruction can be removed.
|
|
The no_constant_pool option was implied by the r11 option. It turns
off the usage of the constant (literal) pool, so that BEAM
instructions that use constants can be loaded in an R11 system.
Since the r11 option has been removed, there is no need to
retain the no_constant_pool option.
|
|
The r11 option was used to generate BEAM modules that could
loaded both on both R11 and R12/R13 to facilitate testing
that R11 and R13 nodes could communicate with each other.
Since R14 is only required to be compatible with R12 and
R13 nodes, the r11 option is no longer needed.
|
|
Print the name and arity of the function being compiled
if the v3_life compiler pass crashes to facilitate finding
out which part of the source code that triggered the problem.
|
|
modules. (Thanks to Jebu Ittiachen.)
|
|
* ks/compiler:
compiler: keep line numbers for attributes
compiler Makefile: alphabetize module names
compile.erl: eliminate compiler warning
|
|
|
|
In the future, we might want to generate warnings
for attributes, referring to them with line numbers.
sys_pre_expand used to replace line number for attributes
with 0. Change sys_pre_expand to retain the real line
number.
v3_core used to throw away the line numbers. Change
v3_core so that it retains the line numbers in annotations.
While at it, do some tidying as suggested by tidier.
|
|
|
|
|
|
We must also catch exits, not only errors, since the hipe
compilers does an exit/1 if an internal error is found.
|
|
Implement the 'no_native' option to disable native-code
compilation. If given in a module like this:
-compile(no_native).
it will override a 'native' option given on the command
line.
|
|
* cf/compile_warning_as_error:
Add option -Werror in erlc(1)
compile: add flag warnings_as_errors to treat warnings as errors
compile.erl: remove trailing whitespace
OTP-8382 The -Werror option for erlc and the compiler option
warnings_as_errors will cause warnings to be treated as errors.
(Thanks to Christopher Faulet.)
|
|
With this flag, warnings are treated as errors, like gcc flag '-Werror'.
|
|
|
|
* bg/compiler-beam_validator:
beam_validator: fix incorrect assumptions about GC guard BIFs
OTP-8378 In rare circumstances when using garbaging collecting guard BIFs,
the validation pass (beam_validator) would signal that the code
was unsafe, when it in fact was correct. (Thanks to Kiran
Khaladkar.)
|
|
The beam_validator pass incorrectly assumes that a GC guard
BIF (such as length/1) may first do a garbage collection
and then fail. That assumption is not correct (guards BIF
only do garbage collection when it is known that the BIF
call will succeed), and will cause the compiler to reject
valid programs.
Modify the beam_validator to assume that if the branch is
taken for a gc_bif instruction, all registers are unchanged
and no garbage collection has occurred. Also add a comment
in the emulator about that assumption.
|
|
* bg/compiler-bopt-bug:
beam_bool: Fix generation of code that does not validate
Fix crash in beam_bool
OTP-8338 Using complex boolean expressions in ifs could cause the compiler
to either crash or teminate with an internal error. (Thanks to
Simon Cornish.)
|
|
The following code (by Simon Cornish)
bad(XDo1, XDo2, Do3) ->
Do1 = (XDo1 =/= []),
Do2 = (XDo2 =/= []),
CH1 = if Do1 == true;
Do1 == false,Do2==false,Do3 == blah ->
ch1;
true ->
no
end,
CH2 = if Do1 == true;
Do1 == false,Do2==false,Do3 == xx ->
ch2;
true ->
no
end,
{CH1,CH2}.
is optimized by beam_bool even though the optimization is not
safe. The trouble is that an assignment to {y,0} no longer
occurs on all paths leading to its use.
The bug is in dst_regs/2 which is supposed to return a set
of all registers assigned in a code block, but it ignores
registers assigned in 'move' instructions.
Fix the bug by taking 'move' instructions into account. This change
is safe since it can only cause more registers to be added
to the MustBeKilled and MustBeUnused sets in ensure_opt_safe/6,
which means that it can only cause the optimization to be turned
off for code that used to be optimized.
|
|
The following code crashes beam_bool:
bad(XDo1, XDo2, Do3) ->
Do1 = (XDo1 =/= []),
Do2 = (XDo2 =/= []),
if Do1 =:= true;
Do1 =:= false, Do2 =:= false, Do3 =:= delete ->
no
end.
(Reported by Simon Cornish; minimized by Kostis Sagonas.)
For the moment fix the bug in the simplest and safest way possible
(basically, instead of crashing just don't do the optimization).
In a future major release (e.g. R14), the following improvements
could be considered:
* In beam_bool, it should be possible to move the Do1 and Do2
expressions to the pre-block and still optimize the expression
in the 'if' statement.
* In sys_core_fold, it should be possible to eliminate the
try/catch around the guard expression in the 'if', because
none of the guard tests can actually fail.
|
|
The original intention for the undocumented 'slim' option was
to omit non-essential parts of *.beam files to reduce the
size of the primary bootstrap. Therefore, debug information and
local function information (only used by xref, not by the loader)
are omitted, but information about the compilation time and
compiler version are still included.
Including compilation information is troublesome, however, when
committing the bootstrap into a revision control system, because
every beam file is guaranteed to be changed every time the bootstrap
is updated.
Therefore, change the 'slim' option to also omit compilation
information.
|
