Age | Commit message (Collapse) | Author |
|
* egil/fix-atom-hash/OTP-10860:
tests: Testing hash values of unicode atoms
erts: Fix atom hash for latin1 characters
tests: Add latin1 chars atom test
|
|
* sverk/is_module_native-fix:
erts: Fix code:is_module_native
OTP-10870
|
|
* fredrik/sws-doc-fix-R16B:
Changed documentation for +sws option
|
|
|
|
* lukas/hipe/fix-utf32-binary/OTP-10867:
noncharacter code points are allowed
|
|
* 'ᚠᚡᚢᚣᚤᚥᚦᚧᚨᚩᚪᚫᚬᚭᚮᚯᚰᚱᚲᚳᚴᚵᚶᚷᚸᚹᚺᚻᚼᚽᚾᚿᛀᛁᛂᛃᛄᛅᛆᛇᛈᛉᛊᛋᛌᛍᛎᛏᛐᛑᛒᛓᛔᛕᛖᛗᛘᛙᛚᛛᛜᛝᛞᛟᛠᛡᛢᛣᛤᛥᛦᛧᛨᛩᛪ᛫᛬᛭ᛮᛯᛰ'
* 'åäö'
* '゙゚゛゜ゝゞゟ゠ァアィイゥウェエォオカガキギクグケゲコゴサザシジスズ'
* '😃'
|
|
Non ASCII latin1 characters were hashed as utf8 which breaks
compatability.
|
|
|
|
* pan/unicode_filename_warnings:
Update file_name_SUITE
Fix broken file:read_link_all/1
Fix broken prim_file:list_dir_all/1
|
|
* sverk/valgrind-cuddling:
test_server: Increase timeout for VALGRIND
erts: Use native atomics and fallback spinlock to mutex for VALGRIND
|
|
code:is_module_native returned false for hipe compiled module if
the first function in the module was a BIF stub
|
|
|
|
* lukas/erts/file_read_line_opt/OTP-10859:
invoke_read_line now remembers read_ahead settings
|
|
* lukas/erts/writev_osx_fix/OTP-10858:
Do not use writev on osx snow leopard
|
|
* bjorn/asn1/bug-fixes/OTP-10853:
asn1ct: Eliminate use of obsolete size/1
Eliminate use of obsolete size/1 in generated code
BER: Fix broken table constraints inside a SET OF/SEQUENCE OF
Smoke test the 'der' option
der: Correct code generation for checking of empty sets
per,uper: Fix breakage of compilation of InformationFramework
Add NullTest
asn1_SUITE: Shorten the path to case-specific directory
asn1_SUITE: Don't add priv_dir to the code path
asn1 tests: Remove obsolete duplicates of ASN.1 specs
|
|
6d516de001dde82c02fe050db8e3aab47914fa90 added prim_file:list_dir_all/1.
Unfortunately, only the first element in the list would be handled
as intended.
|
|
Also teach otp_SUITE:call_to_size/1 that calls to size/1 is no
longer allowed within the asn1 application.
|
|
* lukas/erts/add_sha_to_shell_title/OTP-10838:
Remove could not find warning
|
|
|
|
|
|
writev will for large files return 0 indefinetely causing
the efile driver to go into an infinite loop.
|
|
* rickard/erl_drv_consume_timeslice/OTP-10810:
Implement erl_drv_consume_timeslice()
|
|
* rickard/win-drv-bugfix/OTP-10807:
Fix flush of fd-driver on Windows
|
|
* rickard/win-drv-bugfix/OTP-10803:
Fix unsafe accesses to driver data from win32 fd/vanilla/spawn drivers
|
|
* rickard/win-drv-bugfix/OTP-10802:
Fix unsafe driver_select() from win32 vanilla/spawn drivers
|
|
* rickard/r16b/port-optimizations-fixes/OTP-10809:
Add erl_drv_busy_msgq_limits() to driver API on Windows
Fix driver port accesses
Fix port exit
|
|
* rickard/r16b/notes-fix:
Fix erts-5.10 release notes
|
|
* lukas/erts/add_sha_to_shell_title/OTP-10838:
Remove surplus echos
|
|
|
|
|
|
* lukas/erts/binary_conv_bifs/OTP-10300:
erts: Remove ?line macro from hash_SUITE
Add float_to_binary and binary_to_float
Add new binary conversion bifs
|
|
* lukas/erts/update-config-guess-sub/OTP-10848:
Add +x rights to config.sub
|
|
|
|
|
|
|
|
Added: binary_to_integer/1,2, integer_to_binary/1,2
|
|
* lukas/erts/update-config-guess-sub/OTP-10848:
Update config.guess and config.sub to latest versions
|
|
|
|
* sverk/nif-cut-timeslice:
erts: Add enif_consume_timeslice
OTP-10810
|
|
|
|
* egil/enhance-mseg-cache/OTP-10840:
erts: Utilize even more cached sbc segments
erts: Prefer sbc segment caching over mbc segments
erts: Segment allocator CircleQ API
erts: Increase default #cached segments to 10
erts: Evict old cached segments for newer ones
erts: Refactor mseg cache
|
|
* sverk/efile_drv-memleak:
erts: Fix documentation about 10 default async threads.
erts: Fix memory leak in efile_drv.c
OTP-10841
|
|
|
|
* lukas/erts/add_sha_to_shell_title/OTP-10838:
Include git sha in prompt if available
|
|
The sha will only be included if there is no tag
starting with OTP_R* associated with the sha. This
is because we do not want the sha to show on offical
releases.
|
|
* saleyn/float_to_list_2/OTP-10837:
Fix memory leak in error case
Use macros instead of constants
float_to_list/2 changed rounding and cosmetic cleanup
|
|
|
|
This is needed as C90 (and therefore the win32 compiler) does
not allow constants to be calculations based on other constants.
|
|
|
|
|