| Age | Commit message (Collapse) | Author |
|---|
|
OTP-15897
|
|
OTP-15897
|
|
The nif code expects the socket address to be "complete",
that is all fields must exist. This was not the case for
a couple of functions, sendto, sendmsg and connect.
What was mssing was a call to 'ensure_sockaddr'.
Test: 1) Add a condition function to test if IPv6 is available
and works.
2) Needed to adjust a test case (actually a common function
used in several test cases) since the socket address
data type has two optional fields, which we do not
initiate (when sending), but the nif-code does fill in.
OTP-15897
|
|
|
|
|
|
|
|
* raimo/udp-send-TOS/OTP-15422:
Refine test cases
Remove test code that fails on Windows
|
|
|
|
* rickard/test-fixes-21:
Fix unstable test dump_SUITE:signal_abort
Fix unstable node_container_SUITE:magic_ref test
Fix unstable node_container_SUITE:node_controller_refc test
Fix unstable tests process_SUITE:no_priority_inversion{,2}
|
|
Sometimes processes have not had the time to spread
from one scheduler. Force spread by using the undocumented
'scheduler' option.
|
|
* rickard/test-fixes-20:
Fix unstable node_container_SUITE:magic_ref test
Fix unstable node_container_SUITE:node_controller_refc test
Fix unstable tests process_SUITE:no_priority_inversion{,2}
|
|
|
|
|
|
Sometimes processes have not had the time to spread
from one scheduler. Force spread by using the undocumented
'scheduler' option.
|
|
* sverker/valgrind-hipe-suppression:
erts: Suppress valgrind warning in offset_heap_ptr
|
|
Corrected calculation of buffer length for format(ed) string.
OTP-15817
|
|
The api_m_debug test case seems to trigger a gcc bug that
causes a segfault on an ancient Fedora 16 VM (with gcc
version 4.6.3-2). The simplest way to avoid this is to
test for this specific (gcc) version...
OTP-15817
|
|
Debug printouts is done to "dbgout", but the dbg macro
flushed "stdout" (which was the default). So, in order
to make the macro work properly in all cases, we rename
dbgout to esock_dbgout and make it extern (was static),
and then used that in the macro.
OTP-15817
|
|
|
|
* rickard/pcre-8.43/OTP-15889:
Updated stdlib runtime dependency to erts
Convert generated re replace and split tests from latin1 to UTF8
Adjust generated re replacement and split tests
Generate new re replacement and split tests using perl 5.22.1
Update internal PCRE to version 8.43
|
|
Infinite cycle fixed on try to change run queue (if it has already ch…
|
|
We use the netstat program to find a multicast address.
But that program write some stuff stderr which confused
the code. Add an attempt to filter out anything comming
on stderr. Also try to catch errors, end "translate"
them into skip.
|
|
* john/erts/lists_subtract_fixes/OTP-15938/OTP-15939:
erts: Fix integer overflow in loader
erts: Fix integer overflow in list subtraction
|
|
qsort expects the comparison function to return an int; returning
an `Sint` may yield nonsensical results.
|
|
CMP_TERM returned an `Sint`, which overflowed the `int` used in
erl_rbtree for storing the comparison, causing list subtraction to
behave strangely.
|
|
concurrently)
|
|
OTP-15818
|
|
The previous test amount could take up to 3 hours to finish!
|
|
|
|
|
|
|
|
Before this change erts used to crash dump and then abort, but a
crash dump is not really usefull at this point and it caused all
sort of lock problems when crash dumping that early in the system
boot, so now it is changed to only dump core instead.
Also in the process I cleaned up some the code so that it does
not to a lot of things that are not needed.
|
|
'bmk/erts/esock/incorrect_map_size_check_update_membership/OTP-15908' into maint
|
|
* lukas/erts/dist-optimize-large-data/PR-2291/OTP-15926:
erts: Don't copy binary data from dist data
erts: Move copy of payload to receiving process
|
|
when called by offset_nstack() for hipe native stack.
|
|
Add two simple (traffic) counter test cases, to test counter
actions for normal traffic (send and receive using sendmsg
and recvmsg).
One for IPv4 udp and one for unix-domain-socket stream (udp).
OTP-15818
|
|
Corrected both read pkg, byte and fail counter increment for
recvmsg (was actually missing).
OTP-15818
|
|
Add two simple (traffic) counter test cases, to test
counter actions for normal traffic (send and receive
using sendmsg and recvmsg).
One for IPv4 tcp and one for unix-domain-socket stream (tcp).
OTP-15818
|
|
Corrected both read pkg and byte counter increment for recvfrom
(was actually missing).
OTP-1581
|
|
Add two simple (traffic) counter test cases, to test
counter actions for normal traffic (send and receive
using sendto and recvfrom).
One for IPv4 udp and one for unix-domain-socket dgram (udp).
OTP-15818
|
|
When to count read failure and read package count
used number of bytes read instead of just increment (by 1).
OTP-1581
|
|
Add two simple (traffic) counter test cases, to test
counter actions for normal traffic (send and receive
using send and recv).
One for IPv4 tcp and one for unix-domain-socket stream (tcp).
OTP-15818
|
|
We already had the counters. But we did not entirely handle
counter wrap, especially in related to 'iow' (Inform On Wrap).
There was also no way to get to the counter values, so a socket
specific info function was implemented (piggy back on the existing
"global" info function). And since, we have it, we also add
other info in the returned map (number of readers, writers and
acceptors).
OTP-15818
|
|
|
|
Build failed on windows because of misplaced ifdef's
(of new command functions).
OTP-15817
|
|
rickard-green/rickard/re-unicode-validation/OTP-15831/OTP-15836/ERL-876
re unicode validation
|
|
* sverker/system_info-procs-bug/ERL-979/OTP-15909:
Fix fatal bug in erts_proc_sig_signal_size
|
|
OTP-15817
|
|
Add a command function to the socket API.
Currently only used for changing global debug.
OTP-15817
|
|
* bjorn/erts/unoptimized-plus:
Handle addition of bignum + variable in unoptimized code
|
