| Age | Commit message (Collapse) | Author |
|---|
|
In the v3_life pass, it is assumed that a 'match_fail' primop
only occur at the top-level and at the end of a function.
But this code:
do_split_cases(A) ->
case A of
x ->
Z = dummy1;
_ ->
Z = dummy2,
a=b
end,
Z.
will be optimized by sys_core_fold to the following code:
'split_cases'/1 =
fun (_cor0) ->
let <_cor7,Z> =
case _cor0 of
<'x'> when 'true' ->
< 'dummy1','dummy1' >
<_cor6> when 'true' ->
%% Here follows a 'match_fail' primop inside
%% multiple return values:
< primop 'match_fail'({'badmatch','b'}),'dummy2' >
end
in
Z
moving the 'match_fail' primop into a "values" construction.
In the future, we would like to get rid of the v3_life pass (it is
there for historical reasons), so in the mean-time we prefer to not
add more code to it by generalizing the handling of 'match_fail'.
Since the 'match_fail' primop can be simulated by erlang:error/{1,2},
the simplest solution is to translate 'match_fail' to a call to
erlang:error/{1,2} in v3_kernel and remove the handling of
'match_fail' in v3_life and v3_codegen.
It is tempting to get rid of 'match_fail' also in the Core Erlang
format, but there are two issues:
- Removing the support for 'match_fail' completely may break tools
that generate Core Erlang code. We should not do that in a minor
release.
- There is no easy way to generate a 'function_clause' exception
that will remain correct if it will be inlined into another
function. (Calling "erlang:error(function_clause, Args)" is
fine only if it is not inlined into another function.) A good
solution probably involves introducing new instructions, which
is better done in a major release.
Noticed-by: Håkan Matsson
Minimized-test-case-by: Erik Søe Sørensen
|
|
|
|
Compilation would fail with an error message similar to:
comma_splitter: function split_at_comma/2+32:
Internal consistency check failed - please report this bug.
Instruction: {put,{x,0}}
Error: {match_context,{x,0}}:
Noticed-by: René Kijewski
|
|
* bjorn/compiler/options/OTP-9752:
filename documentation: Recommend against using filename:find_src/1,2
Teach filename:find_src/1,2 to handle slim or stripped BEAM files
filename: Eliminate failing call to Mod:module_info(source_file)
filename.erl:filter_options/1: Remove handling of dead options
compiler: Don't include {cwd,_} in module_info(compile)
compiler: Don't include source code options in module_info(compile)
hipe: Teach hipe to handle slim or stripped BEAM files
|
|
As far as I know, the purpose of the compiler options included in
Mod:module_info(compile) has never been documented.
An educated guess is that they are there in case you want to
re-compile the module with the same options, and also to aid in
debugging when you need to know how a module was compiled.
In neither case is there any need to include options given in the
source itself in options included in Mod:module_info(compile).
Including those options will only waste memory.
Therefore, only include in the BEAM file the options that were
given to compile:file/2.
|
|
|
|
Starting in ff432e262e65243cbc983fcb002527f8fae8c78b, sys_pre_expand
passes external funs through to the downstream passes. It used to
translate external funs to a call to erlang:make_fun/3. Therefore, we
will now need to handle external funs in sys_expand_pmod.
Noticed-by: Stavros Aronis
|
|
Currently, the external fun syntax "fun M:F/A" only supports
literals. That is, "fun lists:reverse/1" is allowed but not
"fun M:F/A".
In many real-life situations, some or all of M, F, A are
not known until run-time, and one is forced to either use
the undocumented erlang:make_fun/3 BIF or to use a
"tuple fun" (which is deprecated).
EEP-23 suggests that the parser (erl_parse) should immediately
transform "fun M:F/A" to "erlang:make_fun(M, F, A)". We have
not followed that approach in this implementation, because we
want the abstract code to mirror the source code as closely
as possible, and we also consider erlang:make_fun/3 to
be an implementation detail that we might want to remove in
the future.
Instead, we will change the abstract format for "fun M:F/A" (in a way
that is not backwards compatible), and while we are at it, we will
move the translation from "fun M:F/A" to "erlang:make_fun(M, F, A)"
from sys_pre_expand down to the v3_core pass. We will also update
the debugger and xref to use the new format.
We did consider making the abstract format backward compatible if
no variables were used in the fun, but decided against it. Keeping
it backward compatible would mean that there would be different
abstract formats for the no-variable and variable case, and tools
would have to handle both formats, probably forever.
Reference: http://www.erlang.org/eeps/eep-0023.html
|
|
It is tempting to transform code such as:
fun(X) -> M:foo(X) end
to:
fun M:foo/1
Unfortunately, that transformation is not safe if
M happens to be a parameterized module. Add a test
case so that we don't attempt to do such an optimization
in the future.
|
|
The two noncharacter code points 16#FFFE and 16#FFFF were not
allowed to be encoded or decoded using the unicode module or
bit syntax. That causes an inconsistency, since the noncharacters
16#FDD0 to 16#FDEF could be encoded/decoded.
There is two ways to fix that inconsistency.
We have chosen to allow 16#FFFE and 16#FFFF to be encoded and
decoded, because the noncharacters could be useful internally
within an application and it will make encoding and decoding
slightly faster.
Reported-by: Alisdair Sullivan
|
|
|
|
* hl/beam_disasm-no_attri_chunk:
beam_disasm: Handle stripped BEAM files
OTP-9571
|
|
beam_disasm:file/1 would crash if asked to disassemble a stripped
BEAM file without an "Attr" chunk.
|
|
Conflicts:
lib/asn1/doc/src/asn1ct.xml
|
|
|
|
Also update the r12 and r13 options so that they imply no_line_info.
|
|
This commit is a preparation for introducing location information
(filename/line number) in stacktraces in exceptions. Currently
a stack trace looks like:
[{Mod1,Function1,Arity1},
.
.
.
{ModN,FunctionN,ArityN}]
Add a forth element to each tuple that can be used indication
the filename and line number of the source file:
[{Mod1,Function1,Arity1,Location1},
.
.
.
{ModN,FunctionN,ArityN,LocationN}]
In this commit, the fourth element will just be an empty list,
and we will change all code that look at or manipulate stacktraces.
|
|
|
|
|
|
|
|
|
|
Since the introduction of improved record optimizations in
1858cb81391d2bce29b4b7620574ca60128cebf7 and
470c91d43eae54f63661645acbce4b92d73287cc, the optimization of
a is_record/3 call with a known correct type in
beam_type:simplify_basic_1/3 has not been covered.
|
|
In 3d0f4a3085f11389e5b22d10f96f0cbf08c9337f (an update to conform
with common_test), in all test_lib:recompile(?MODULE) calls, ?MODULE
was changed to the actual name of the module. That would cause
test_lib:recompile/1 to compile the module with the incorrect
compiler options in cloned modules such as record_no_opt_SUITE,
causing worse coverage.
|
|
* hw/call-chmod-without-f:
Call chmod without the "-f" flag
Conflicts:
erts/emulator/test/Makefile
lib/asn1/test/Makefile
lib/crypto/test/Makefile
lib/debugger/test/Makefile
lib/docbuilder/test/Makefile
lib/edoc/test/Makefile
lib/erl_interface/test/Makefile
lib/inviso/test/Makefile
lib/parsetools/test/Makefile
lib/percept/test/Makefile
lib/ssl/test/Makefile
lib/syntax_tools/test/Makefile
lib/test_server/test/Makefile
lib/tools/test/Makefile
OTP-9170
|
|
The compiler (sys_core_fold) tries to avoid constructing tuples
in case expressions. The following code:
c(A, B) ->
case {A,B} of
{ok,X} -> X;
{_,_} -> error
end.
will be rewritten so that no tuple is built. If a clause
requires a tuple to be built as in this code:
c(A, B) ->
case {A,B} of
{ok,X} -> X;
V -> V %The tuple will be built here
end.
the tuple will be built in the clause(s) in which it is needed.
If the value returned from the case is not used as in this code:
c(A, B) ->
case {A,B} of
V -> V %Warning: a term is constructed, but never used
end,
ok.
there will be an incorrect warning. Basically, what happens is
that the code is reduced to:
c(A, B) ->
{A,B}, %Warning: a term is constructed, but never used
ok.
and the optimizer sees that the {A,B} tuple can't possibly be used.
Eliminate the warning by adding a 'compiler_generated' annotation
to the tuple.
Reported-by: Kostis Sagonas
|
|
In the following code:
m(<<Sz:8,_:Sz/binary>>) ->
Sz = wrong.
the Sz variable is supposed to be bound in the function header and the
matching "Sz = wrong" should cause a badarg exception. But what
happens is that the Sz variables seems to be unbound and the matching
succeds and the m/1 function returns 'wrong'.
If the Sz variable is used directly (not matched), it will have
the expected value. Thus the following code:
m(<<Sz:8,_:Sz/binary>>) ->
Sz.
will correctly return the value of Sz that was matched out from
the binary.
Reported-by: Bernard Duggan
|
|
|
|
|
|
* jp/dependencies_makefile:
Add dependencies Makefile generation to erlc(1) and compile(3)
Conflicts:
lib/compiler/test/compile_SUITE.erl
OTP-9065
|
|
This is useful when a project is built with Makefiles and erlc(1)
instead of EMakefiles. Tracking dependencies by hand is error-prone and
it becomes painful when using external application headers like EUnit's
one.
A dependencies Makefile will look like this:
module.beam: module.erl \
/usr/local/lib/erlang/lib/eunit-2.1.4/include/eunit.hrl \
header.hrl
When included in the main Makefile, 'module' will be recompiled only
when needed.
GCC offers the same feature and new erlc(1) options are compatible with
it.
More informations at:
http://wiki.github.com/dumbbell/otp/dependencies-makefile
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Sometimes the beam_bool pass wants to know whether an
y register will be killed by the code that follows and
will do (effectively):
beam_utils:is_killed({y,Y}, Code, L)
When asked to calculate the liveness for an y register,
beam_utils:is_killed/3 will loop forever if the code
includes a receive loop.
Since this rarely occurs, fix the problem in the simplest
and most conservative way.
Reported-by: Christopher Williams
|
|
When gc_bif instructions occurred outside of a block,
beam_utils:check_liveness/3 did not take into account
that the instruction could do a garbage collection, and
could falsely report that an x register would be killed.
That could cause the beam_dead pass to make the code
unsafe by removing the assignment to an x register that
would subsequently be referenced by the garbage collector.
Reported-by: Christopher Williams
|
|
"-f" is a non-standard chmod option which at least SGI IRIX and HP UX do
not support. As the only effect of the "-f" flag is to suppress warning
messages, it can be safely omitted.
|
|
Moving of allocation instructions upwards in the instruction
stream (in order to enable further optimizations) in beam_block,
is implemented with the assumption that if a register {x,X}
contains a valid term, then all other x register with lower
numbers than X also contain valid terms. That assumption is
true after code generation.
The beam_utils:live_opt/1 optimization, however, may invalidate
that assumption. For instance, if a receive statement exports a
variable that is used, but the return value of the receive statement
is not used, then {x,1} but not {x,0} contains a valid term at the
end of the receive statement. If the receive statement is
followed by
{bif,self,{f,0},[],{x,0}}.
{test_heap,NumberOfWords,2}.
moving the allocation upwards will produce
{test_heap,NumberOfWords,2}.
{bif,self,{f,0},[],{x,0}}.
which will cause the beam_validator pass to scream loudly that
{x,0} is not live at the test_heap instruction.
Fix the problem by doing the optimizations in reverse order.
Reported-by: Jim Engquist
|
|
* bjorn/compiler-bin-generators/OTP-8864:
core_lint: Enforce that tail segments only occur at the end
Don't generate multiple tail segments in binary matching
Factor out some of the code for binary generators
Add tests for tail segments in binary generators
|
|
Code such as
foo(A) -> <<A:0>>.
would cause a compiler crash.
|
|
|
|
The code for generating the string table (which is now
only used for bit syntax matching) in a BEAM file is quite
complicated and potentially expensive when compiling modules
with many thousands of clauses doing bit syntax matching.
Simplify and optimize the code using bit syntax and
binary:match/2 instead of the list operations in the
original code.
|
|
warn_unused_import works correctly (does not give warnings when overridden).
Local call in guard gives its own error pointing out the local/imported
function.
Use of the phrase "overridden auto-imported bif" instead of "redefined
auto-imported bif" in textual error messages.
|
|
|
|
Wrote and changed some tests in stdlib:erl_lint_SUITE
nowarn_bif_clash is obsoleted but will remove warnings and errors about bif
clashes. The recommended way is to use no_auto_import directives instead.
Hopefully erlang.erl is the only user in the world of nowarn_bif_clash.
|
|
Added only a few testcases in compiler:error_SUITE and guard_SUITE
The new behaviour of warnings and errors when overriding autoimported BIF's:
Bifs that were autoimported before R14 are dangerous because old code
using them and overriding them in exports can start behaving
differently. For newly added autoimports this can't happen to the new
code that wants to (or dont want to) use them, why only warnings are
added for the BIFs autoimported after the compilator change. Errors
are issued only for code that could have worked in one way in R13 and
now will behave in a different way.
If overriding autoimport with local function:
- if explicit -compile directive supresses autoimport
-> no message
else
- if called from inside module
- if pre R14 autoimported bif
-> error
else
-> warning
else
-> no message
If overriding autoimport with import directive
- if explicit -compile directive supresses autoimport
-> no message
else (regardless of actual usage)
- if pre R14 autoimported bif
-> error
else
-> warning
Calls of local functions or imports overriding autoimported functions
(either post R14 or by using explicit -compile supressions of
autoimport) always goes to the local function or the imported.
The compileation errors are added to not let code like this silently
and disastrously change its semantic (probably to an infinite loop)
between R13 and R14:
----------
-module(m).
-export([length/1]).
length(X) ->
...
Y = length(Z),
....
----------
The user has to select if he/she wants to call length in 'erlang' explicitly
or if the overriding semantics is desired, in which case the -compile
directive has to be used.
-compile({no_auto_import,[F/A]}). Is added to allow to override the
autoimports so that code gets unanbiguous. The directive will remove
an autoimport even if there is no local function or import overriding,
because any other behaviour would be inconsistent and confusing.
record_info and module_info can never be overridden.
|
