| Age | Commit message (Collapse) | Author |
|---|
|
Commits 53bd4974a101 and 726f6e4c7afe simplified the handling of
match_fail (used to generated exceptions such as 'function_clause')
by first rewriting them to a call to erlang/error{1,2} and later
rewriting them to specialized BEAM instructions (to reduce the
code size).
There was one flaw, though, which only was exposed when more
aggressive optimizations were added in c3b60f86c622. Here is an
example to explain it:
t(V) ->
fun(get) -> V end.
The following BEAM code will be initially generated for the fun:
{function, '-t/1-fun-0-', 2, 5}.
{label,1}.
{line,[{location,"t.erl",5}]}.
{func_info,{atom,t},{atom,'-t/1-fun-0-'},2}.
{label,2}.
{test,is_eq_exact,{f,2},[{x,0},{atom,get}]}.
{move,{x,1},{x,0}}.
return.
{label,2}.
{test_heap,2,1}.
{put_list,{x,0},nil,{x,1}}.
{move,{atom,function_clause},{x,0}}.
{line,[{location,"t.erl",5}]}.
{call_ext_only,2,{extfunc,erlang,error,2}}.
Translating back to Erlang code, that would be roughly:
'-t/1-fun-0-'(get, V) -> V;
'-t/1-fun-0-'(Arg1, _) -> erlang:error(function_clause, [Arg1]).
Note that the second argument (the free variable V) is not included
in the call to erlang:error/2.
The beam_except pass will simplify the code to:
{function, '-t/1-fun-0-', 2, 8}.
{label,1}.
{line,[{location,"t.erl",5}]}.
{func_info,{atom,t},{atom,'-t/1-fun-0-'},2}.
{label,2}.
{test,is_eq_exact,{f,1},[{x,0},{atom,get}]}.
{move,{x,1},{x,0}}.
return.
The code has been shortened by jumping to the func_info/3 instruction.
Translating back to Erlang:
'-t/1-fun-0-'(get, V) -> V;
'-t/1-fun-0-'(Arg1, Arg2) -> erlang:error(function_clause, [Arg1,Arg2]).
it is clear that both arguments are now included in the
'function_clause' exception, even though the initially generated
code only included the first argument.
That is no problem in this particular case, but for some more complex
funs, optimizing the first version based on variable usage could make
the second version unsafe.
I rejected the following potential solutions:
- Including the free arguments in the call to erlang:error/2:
'-t/1-fun-0-'(get, V) -> V;
'-t/1-fun-0-'(Arg1, Arg2) -> erlang:error(function_clause, [Arg1,Arg2]).
Unfortunately, that is tricky. The free variables are only known
after the second pass in v3_kernel when variable usage has been
calculated. We would need to add a third pass (only for funs) that
would the free arguments to the second argument for erlang:error/2
*and* update the variable usage information.
- Calling beam_except earlier, from within beam_block before any
optimizations based on variable usages are done. But means that the
problem could reappear in some other form in the future when other
updates are done to the code generator and/or optimization passes.
The solution I have chosen is to modify beam_except to only replace
a call to erlang:error(function_class, Args) if the length of Args
is the same as the arity in the func_info/3 instruction. The code
will be slightly larger. Also, the free variables for funs and list
comprehensions will no longer be included in the function_clause
exception (that could be less confusing, but it also means less
information during debugging).
|
|
* sverk/packet-http-string-maxlen:
erts: Increase length of well formed header names from {packet,http}
OTP-10824
|
|
* sverk/erl_interface-enum-typedef:
erl_interface: Change enum erlang_char_encoding to a typedef
|
|
* sverk/r16/erl_interface-fixup:
erl_interface: Fix ei_skip_term
|
|
|
|
|
|
|
|
* hb/stdlib/bug_fixes/OTP-10622:
[stdlib] Fix a bug that could cause the Erlang shell to loop
[stdlib] Fix bugs in eval_bits
|
|
* fredrik/odbc/update-vsn:
Bumped version number
|
|
|
|
* kp/odbc-empty-params/OTP-10798:
Add testcases for odbc:param_query
Fix odbc:param_query/3 and odbc:param_query/4.
|
|
|
|
Unicode related.
|
|
to make the API independent of the actual type of erlang_char_encoding
|
|
* siri/ts-and-tp-in-test-logs/OTP-10780:
[test_server] Add error printouts in case writing to test case html log fails
[test_server] Don't write unicode strings to latin1 log files
|
|
* bmk/snmp/snmp4222_integration/r15:
[snmp/compiler] Improved debug printouts
[snmp/compiler] Add test case and test mib
[snmp/compiler] Proper release nodes
[snmp/compiler] Add the mib (ALARM-MIB)
[snmp/compiler] MIB compiler did not handle forward index ref
[snmp/compiler] Added test case for BITS import
[snmp/compiler] Release notes, appup and proper version
[snmp/compiler] Corrected test mibs
[snmp/compiler] MIB compiler did not handle import BITS
|
|
* siri/sasl/appup-R16/OTP-10806:
Update sasl.appup.src with R16 release and remove R13 versions
|
|
* maint:
Prepare release
Don't run testX420/1 on old slow Sparc systems
testX420: Pass Options to the ASN.1 compiler
asn1_erl_nif: Correct broken length encoding
asn1_SUITE: Mend broken test_modified_x420/1
Revert "Prepare release"
Conflicts:
lib/asn1/doc/src/notes.xml
lib/asn1/test/asn1_SUITE.erl
|
|
* maint-r15:
Prepare release
Don't run testX420/1 on old slow Sparc systems
testX420: Pass Options to the ASN.1 compiler
asn1_erl_nif: Correct broken length encoding
asn1_SUITE: Mend broken test_modified_x420/1
Fix a bug for multiple extension addition groups
|
|
Errors are printed to unexpected_io log.
|
|
|
|
* bjorn/compiler/dialyzer-warnings:
compile: Eliminate warnings for unmatched return values
beam_receive: Eliminate dialyzer warning for unmatched return
beam_validator: Eliminate dialyzer warnings for unmatched returns
|
|
|
|
* bjorn/stdlib/erl_lint-coverage:
Test calling deprecated or removed functions
Test deprecating list_to_atom/1 to cover more code
Cover binary syntax errors and warnings
Remove unused error message {bittype_mismatch,_,_,_}
Remove unused error message for importing an auto-imported BIF
Remove the error message for future reserved keyword
Test a few more kinds of illegal guard expressions
Remove error handling for calling a BIF that is not auto-imported
Provoke basic semantic errors to improve test coverage
|
|
* nox/fix-seq-opt/OTP-10818:
Add two tests for unused multiple values in effect context
Forbid multiple values in Core Erlang sequence arguments
|
|
* mh/duplicate_name_error_message/OTP-10797:
Slightly nicer error message when node start fails due to duplicate name
|
|
* fredrik/tools/version_bumps:
Bumped version number
|
|
|
|
To ease matching of unrecognized header field names we convert them
from case insensitive to the format Sec-Websocket-Version
with capital letters only first and after hyphens.
Earlier only header names up to 20 characters were converted to this
format due to internal buffer limitation. Raising this limit to 50
is a pragmatic solution for existing long header names such as
Sec-Websocket-Version, while valid header names longer than 50 characters
are not very probable.
|
|
|
|
|
|
The assert_fls/2 and assert_type/3 functions both return the
Vst passed to them, but all callers ignore the return value.
Given the name of the functions, they are not expected to return
anything. Make it so by changing the return value to 'ok'.
There are two calls to bsm_get_context/2 used only to validate
that the match context is valid. Call bsm_validate_context/2
instead.
In bsm_validate_context/2, explicitly match the return value of
bsm_get_context/2 to '_' to make it clear that it is not used.
|
|
|
|
* bjorn/asn1/open-type-error/OTP-10805:
asn1_erl_nif: Correct broken length encoding
asn1_SUITE: Mend broken test_modified_x420/1
|
|
* bjorn/asn1/extension-addition-groups/OTP-10811:
Fix a bug for multiple extension addition groups
|
|
* bjorn/asn1/test-cases:
Don't run testX420/1 on old slow Sparc systems
testX420: Pass Options to the ASN.1 compiler
|
|
One and a half hour is not enough for it to finish.
|
|
When the caller passed the 'der' option, it was ignored.
|
|
bmk/snmp/snmp4231_integration/r16
Conflicts:
lib/snmp/doc/src/notes.xml
lib/snmp/src/app/snmp.appup.src
lib/snmp/vsn.mk
|
|
bmk/snmp/snmp4222_integration/r15
|
|
'bmk/snmp/compiler/handling_import_of_BITS_in_mibs/r15/OTP-10799' into bmk/snmp/snmp4222_integration/r15
|
|
|
|
|
|
|
|
|
|
The MIB compiler could not handle a table index refering to
an object defined later in the MIB.
|
|
|
|
* jv/cover-patches/OTP-10778:
Fix a bug in cover when used with no_auto_import
Ensure cover keeps the proper file source
cover now relies on the compile info to find file sources
|
|
The unicode update of test_server for R16A introduced a few potential
errors when logging to files. Sometimes ~tp or ~ts was used for
formatting also when writing to files that were not opened with the
{encoding,utf8} option. If then the argument contained unicode
characters above 255, the file descriptor would crash. This has been
corrected by the following modifications:
* Since the 'unexpected_io' log file is used only when the test case
HTML file is not available (e.g. between test cases), this file is
now also a HTML file and as other test_server HTML logs it is always
UTF-8 encoded
* Since it is possible to change which information is going to which
log file (with test_server_ctrl:set_levels/3), we do not have full
control over which information is written to which file. This means
that any printout could be written to the 'major' log file
(suite.log), which was earlier encoded as latin1. To avoid crashing
this file descriptor due to unicode strings, the 'major' log file is
now also encoded in UTF-8 (possible incopatibility).
* The cross_cover.info file is no longer a text file which can be read
with file:consult/1, instead it is written as a pure binary file
using term_to_binary when writing and binary_to_term when reading.
* The encoding of the file named 'last_name', which only content is
the path to the last run.<timestamp> directory, is now dependent on
the file name mode of the VM. If file names are expected to be
unicode, then the 'last_name' file is UTF-8 encoded, else it is
latin1 encoded.
Also, ~tp is changed back to ~p unless it is somehow likely that the
argument includes strings. It is not obvious that this is the correct
thing to do, but some decission had to be taken...
|
|
The ber_bin_v2 backend calls a NIF function to decode all tags
and lengths. Even open types that should not be decoded will also be
decoded, which makes it necessary to later re-encode the data using
a NIF function.
The NIF function incorrectly encoded lengths.
|
