| Age | Commit message (Collapse) | Author |
|---|
|
|
|
* bjorn/cuddle-with-tests:
Stabilize call_with_huge_message_queue/1
same_time_yielding/1: Avoid failing if there are many schedulers
Remove obsolete old_scheduler_SUITE
Robustify process_SUITE:spawn_opt_max_heap_size/1
Contain damage cause by spawn_opt_max_heap_size failing
Make bulk_send_bigbig/1 purely informational
distribution_SUITE: Refactor bulk_send_bigbig/1
receive_SUITE: Remove ?line macros
Lengthen too short timetraps
Fix process_SUITE:low_prio/1
|
|
* bjorn/erts/fix-sys-task-cleanup:
Make sure that asynchronous replies are not lost
|
|
Fix typo
|
|
|
|
|
|
* ingela/inets/ERL-382:
inets: httpc - Gracefull normal shutdown
|
|
* ingela/dtls/retransmisssion-timers:
dtls: Fetch next DTLS record when dropping resent handshake data
dtls: Use enter actions
|
|
* lukas/erts/sendfile_sunos_fix/OTP-14424:
erts: Fix sendfile closeduring scenario on sunos
|
|
* dgud/observer/edit-pid-port-ref:
observer: Allow edit of pid, port and refs
|
|
* dgud/fix-unicode-atom:
mnesia: Print tables with ~ts to support unicode atoms
|
|
On Solaris, giving a too long sfv_len results in an
EINVAL error, but data is still transmitted and len is
correctly. So we translate this to a success with that
amount of data sent. This may hide some other errors
that causes EINVAL, but it is the best we can do for now.
|
|
|
|
Using enter actions for retransmission timers makes the code easier to
understand. Previously the retransmission timer was incorrectly started in
the connection state. Using enter actions feels like a cleaner approach
than bloating the state with more flags.
|
|
* lukas/erts/20_minor_fixes:
kernel: Iterate for correct time in os_SUITE:perf_counter
|
|
Conflicts:
OTP_VERSION
erts/vsn.mk
|
|
|
|
|
|
|
|
Time measurements are always tricky, particulary on virtual hosts.
On one particular virtual host, the measured times were often 0.
Measuring the time for 100 calls instead of 10 calls helps, but
0 can still be returned, so we will also need to discard
measurements that return 0 and try again.
|
|
On a computer with 32 schedulers, there would be 9632 (301*32)
'timeout' message in the receive queue. Receiving them with a
selective receive (matching on the timer ID) is quite slow.
Change the test case to read out the queue in the order the
messages are stored in the queue.
|
|
old_scheduler_SUITE was written before the SMP emulator.
It does not test anything that scheduler_SUITE does not
test. Keeping it wastes time when running the emulator
test suite, and does not help us find any bugs that
scheduler_SUITE would find.
|
|
process_SUITE starts os_mon in init_per_suite/1. Therefore,
there may be occasional alarm info messages received. Make
sure that they are ignored and don't cause the test fail to
fail.
Also, get rid of the flush/0 function that discards all
messages in the message queue. Instead, be more selective
and only discard {error, ...} messages.
|
|
If process_SUITE:spawn_opt_max_heap_size/1 failed, the default
value for the maximum heap size could have been changed. That
would cause other test cases that spawned huge processes to
fail.
Contain the damage by always restoring the default value
for max_heap_size in end_per_testcase/2.
|
|
bulk_send_bigbig/1 sometimes fails even if there is nothing wrong.
Stop testing the ratio, and just return information about the speed
and number of monitor messages. The testcase will still serve as a
smoke test of the +zdbbl option.
|
|
Refactor bulk_send_bigbig/1 to make it somewhat easier to follow.
|
|
|
|
c062dfc485a added timetraps to test suites that had no timetraps.
Some of added timetraps are only 5 or 10 seconds, which is can be
too short time on some slow computers. Change those times to one
or two minutes.
|
|
It is expected that low-priority processes should be allowed
to run about 1/8 of the time of normal-priority processes.
The proportion calculated was usually considerably lower than
1/8. It seems that the main reason is the punishement in
reductions when sending to a process that alreday has many
messages in the message queue. I have verified that by running
the emulator with the +snsp to turn off the send punishment.
To avoid that punishment, let the server process run at high
priority so that it can keep its message queue short.
It seems that the proportion is now very close to 1/8. Therefore,
we can tighten the margins for the proportion.
|
|
|
|
* lukas/erts/hipe_mkliterals_libs/ERIERL-14/OTP-14447:
hipe: Fix hipe_mkliterals make recipe
|
|
Make some C code static
|
|
erts: Remove unused functions from erl_cpu_topology
|
|
Update erl_msg_tracer example
|
|
|
|
The need for more trace patterns requires a somewhat different structure.
It was previoiusly a bit difficult to use in e.g. test suites. Now it is easier.
|
|
|
|
|
|
|
|
|
|
* lars/xmerl/fragmented-xml-directive/OTP-14430:
[xmerl] Update version
[xmerl] Add test cases
[xmerl] Fix endDocument bug in output backends
[xmerl] Fix fragmented xml directive bug and replace ets table
|
|
* lukas/erts/restore-sigterm-default/OTP-14418:
erts: Add undocumented option to do default SIGTERM
|
|
|
|
|
|
Correct close semantics for active once connections. This was a timing dependent bug the resulted in the close message not always reaching the ssl user process.
OTP-14443
|
|
|
|
|
|
In the test group public_key we need to control both the host key
and the user key. This new test case is aimed at faciliate the
user public key testing.
|
|
|
|
This reverts commit 4ee80fd8738393bf581e0393416befda1ca621b6.
|
