| Age | Commit message (Collapse) | Author |
|---|
|
Remove unused variable warning in compile_SUITE
|
|
system/doc/reference_manual/: fix a typo
|
|
Fixed typo.
|
|
This reverts commit 4f03eba5294f1c1bc922b07f1267e6b0099fd4b5.
|
|
This reverts commit dc57404252c47520f352834ad9be45ad684f96c9.
|
|
p->fvalue will be set by BIFs that generate exceptions
(such as throw/1), and it will not be cleared before another
exception is generated. Potentially, p->fvalue may contain a
huge term (e.g. after throw(HugeTerm)) which will be kept in
the heap.
We can shorten the lifetime of f->value by clearing it in
the instructions that handle catches: catch_end and try_end.
That is safe because BEAM code will never access p->fvalue.
If BEAM code needs to rethrow an exception it will use a
reference to the value passed in an X register.
The reason that p->fvalue must not be cleared already in
handle_error() is that native code trap handlers will use
it. (See the comment before handle_error() and the comment
at the end of handle_error() in beam_emu.c for some more
information about exception handling.)
|
|
|
|
The undocumented compiler option 'slim' is used when compiling
the primary bootstrap. The purpose is to make the bootstrap smaller
and to avoid unnecessary churn in the git repository. That is,
the BEAM file should be different only if the actual code in the
file is different, and not if it has merely been re-compiled on
a different computer.
Two commits have fattened the 'slim' option. In 36f7087ae0f,
extra chunks are included even in slim BEAM files. In dfb899c0229f7,
the "Dbgi" were added as an extra chunk, causing it to be included
in slim files.
Make 'slim' slim again by only including the essential chunks and
the attribute chunk (as was the case before the {extra,...} option
was added).
|
|
|
|
Try doing three consecutive chunked receive with some changed timing in the ending sequence
|
|
|
|
|
|
|
|
Local trace on first function in module
made code:is_module_native/1 return true.
Use new erts_is_function_native() to make a proper check.
|
|
|
|
|
|
|
|
When removing typer as application, it was also removed from the list
of applications to exclude from upgrade tests. However, since typer
still exist in older OTP releases, it must remain in the exclude list
as long as these releases are used in the test.
|
|
|
|
|
|
* IPv6 address FFFF::192.168.42.2 is an invalid form as an IPv4-mapped address.
This should be fixed as ::FFFF:192.168.42.2.
The assigned address block is ::ffff:0:0/96.
See RFC6890 Table 20 in the Section 2.2.2 for the further details.
And for this IPv4-mapped address use *only*,
the 32bit LSB of the IPv6 address can be written
with the respective IPv4 address.
See RFC4291 Section 2.2 Form 3 for the further details.
|
|
|
|
|
|
* rand/arbitrary_normal_distributions:
Disable heavy test case
|
|
* raimo/rand-dev/OTP-14295:
Adjust timetrap
|
|
|
|
It turns out that the excessive memory usage is cause by the
test framework printing all the test vectors into the log output.
A similar proplem was already diagnosed for long_msg/0. The root
cause was not mentioned in the SUITE, but the same fix applies
to the CAPV test vector data.
Switch all CAPV data to lazy evaluation and have the test itself
read the data.
|
|
* dgud/mnesia/snmp-sup:
mnesia: Remove not used mnesia_snmp_sup
Do not call which_applications()
|
|
* ingela/ssl/releasenote-fix:
ssl: Move missing release note fix to correct place
|
|
Commit 87584ae85893df917ca83cb0c40748fd4da0f3bc added missing release note
but not in the correct place.
|
|
|
|
|
|
|
|
NIST's Cryptographic Algorithm Validation Program provides
validation testing of FIPS-approved and NIST-recommended
cryptographic algorithms.
Instead of hard coding a limited set of test vectors, use
their comprehensive validation set to test AES cipher modes.
|
|
in the httpd_util:rfc1123_date/1 function.
|
|
* ingela/ssl/missing-relnote:
ssl: Add missing release note
|
|
|
|
* hm/escript-doc:
Improvements of escript documentation
|
|
* maint:
Updated OTP version
Prepare release
inets: Add missing release note
dialyzer: Fix a bug regarding unknown types
erl_mseg.c: don't use invalid indices in - > cache_powered_node[]
Fix release note for OTP-14290
Conflicts:
OTP_VERSION
erts/emulator/sys/common/erl_mseg.c
erts/vsn.mk
lib/dialyzer/src/dialyzer_analysis_callgraph.erl
|
|
* origin/hasse/correct_unknown_types:
hipe: Fix unknown type
dialyzer: Fix unknown type
|
|
It is not used anymore, snmp indecies are now ets tables and not processes.
|
|
* maint-19:
Updated OTP version
Prepare release
inets: Add missing release note
dialyzer: Fix a bug regarding unknown types
erl_mseg.c: don't use invalid indices in - > cache_powered_node[]
Fix release note for OTP-14290
|
|
Which may hang if an application start does wait for tables
during initilization.
|
|
* ingela/ssl/prepare-release:
ssl: Prepare for release
|
|
|
|
|
|
* dgud/wx/fixup-batch-break:
Fixup batch break
|
|
|
|
in case of active timers exists.
Added unit tests for hibernate_after functionality combined with gen_statem timers.
|
|
It was done because "hibernate_after" option already used in ssl for the same reason.
|
