| Age | Commit message (Collapse) | Author |
|---|
|
When determining whether the delayed creation of sub-binaries
optimizations is applicable, this module some tests whether the
register containg the match state is killed. That is actually a
stronger condition than necessary; since the register is initialized,
it suffices to test whether the register is unused.
|
|
Generate slightly smaller and faster code.
|
|
* bjorn/asn1/clean-up-backends/OTP-10410:
Simplify the code for the generated info/0 function
Don't try to work around a non-loadable NIF library
Fix BER encoding when multiple levels of typedefs are used
Update megaco documentation
Update documentation for the asn1 application
Fix other applications
Fix use of asn1 in megaco
Remove the unused asn1ct_gen_ber module
Remove the unused asn1ct_constructed_ber module
Remove the unused asn1ct_per_bin module
Remove unused functions in asn1rt_ber_bin
Remove obsolete back-ends and simplify the options
Make the specialized decodes work with the 'nif' option
Remove support for the obsolete {Typename,Value} tuple notation
Remove tests for the obsolete {TypeName,Value} notation
Remove the obsolete and deprecated 'keyed_list' option
|
|
* bjorn/ct/separate-io-server/OTP-10101:
Fix race condition in test_server_io
|
|
* bjorn/ct/env-vars/OTP-10469:
Add support for passing environment variables in master mode
Remove redundant sleep in ct_master_SUITE
Test that test cases succeed
Add ct_test_support:verify_events/4 which takes a node name
|
|
While at it, also make the generated code for the attributes
more readable.
|
|
The NIF library is now mandatory.
The call to application:get_env/2 to find out whether the NIF
library is loaded is surprisingly expensive.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
* egil/fix-heart-unregister:
kernel: Heart port needs to be unregistered
|
|
|
|
* raimo/daily-build-fixes:
Fix erroneous skipping for jinterface, erl_interface and ic
|
|
|
|
When heart cycles we need to unregister the old port before
starting the new heart port program.
OTP-10591
|
|
|
|
|
|
|
|
|
|
Conflicts:
erts/doc/src/erlang.xml
|
|
* pan/process_info_spec_doc/r15/OTP-10584:
Correct doc of process_info(Pid,links)
OTP-10584
|
|
* pan/bs_append_crash:
Teach erts_bs_append not to dump core
OTP-10590
|
|
|
|
* anders/diameter/doc/OTP-10568:
Update doc for RFC 6733
Add copies of RFC's 6733 and 6737
|
|
* anders/diameter/doc/OTP-10471:
Add missing diameter_codec(3) content
Add content to diameter_codec(3) and diameter_make(3)
Add reference pages diameter_codec(3) and diameter_make(3)
|
|
|
|
|
|
|
|
|
|
* peppe/common_test/group_search_r15b03.fix:
Fix silly bug
|
|
|
|
|
|
|
|
|
|
It is time to clean up the mess of back-ends.
Remove all the obsolete back-ends and simplify the options used
to select them.
New Option Old Equivalent
---------- --------------
ber ber_bin,optimize,nif
per per,optimize,nif
uper uper_bin
The old options will still be recognized and translated to the
new options, but will also print a warning.
That implies that deprecated features that only are implemented
in the old 'ber' back-end will no longer work (e.g. the
{Typename,Value} notation).
Also make the return type for the generated encode/2 function
consistent. It used to be a binary for per and uper, and an iolist
for ber. Always make it a binary.
|
|
|
|
* egil/fix-crash_dump-lc-assert:
erts: Suppress lock assertion when dumping a crash
|
|
|
|
6733 deprecates 3588.
|
|
|
|
Not yet any content to speak of.
|
|
|
|
Of the back-ends that we are going to keep, only the UPER back-end
support the obsolete {Typename,Value} notation. For consistency
with the PER and BER encodings, remove the support for UPER
encoding too. Also remove vestiges of the support for the
notation in the other back-ends.
|
|
As a preparation for removing obsolete back-ends, remove tests
for the {TypeName,Value} notation to avoid having those test cases
fail.
|
|
The 'keyed_list' was only supported for the 'ber' and 'ber_bin'
backends and has been undocumented for a long time. Also remove
the note in the documentation about the feature.
|
|
|
