| Commit message (Collapse) | Author | Age | Files | Lines |
| |
|
|\
| |
| |
| | |
into maint
|
| |
| |
| |
| |
| |
| |
| |
| |
| | |
On some old linux platforms the errqueue does not
explicitly include the linux types include file.
Therefor we (always) do this to avoid the problem of
trying to figure out if we need to.
OTP-16302
|
|\ \
| | |
| | |
| | |
| | |
| | |
| | | |
* lukas/erts/ets-smp-1-optimization/OTP-16315/OTP-16316:
erts: Optimize dec_term for atoms when used by ets compressed
erts: Optimize meta table lock to not be taken when +S 1
ets: Remove table locking when using smp 1
|
| | |
| | |
| | |
| | |
| | |
| | | |
We don't need to do a size check of the atoms as we know
that they will be valid atoms when extracted from ets as
they were valid when we put them into ets.
|
| | |
| | |
| | |
| | |
| | |
| | | |
ets tables can only be accessed from normal schedulers, so
when there is only one normal scheduler we don't have to take
any locks.
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
If we only have one normal scheduler we know that only
one thread can access an ETS table at a time, so we don't
need to optimize away that.
This approach reads and checks the global variable
erts_no_schedulers each time, it could be optimized
to use the table flags field instead. Though I'm not sure
how much that would help.
With this change doing ets:lookup/2 och a small value is
about twice as fast.
|
|\ \ \
| | | |
| | | |
| | | |
| | | | |
* ingela/inets/warnings:
inets: Remove legacy call to random module
|
| | | | |
|
|\ \ \ \
| | | | |
| | | | |
| | | | |
| | | | | |
* peterdmv/eunit/remove_warnings/OTP-16313:
eunit: Remove compiler warnings
|
| | | | | |
|
|\ \ \ \ \
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
* dgud/wx/fix-ifdef:
Fix awk warning about missing file
Fix deprecated warning
Add wxMouseEvent:getWheelAxis function
Fix enum that have changed values
Fix bad compiler test
|
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
For edoc all files does not exists when doc is not built yet.
And we don't want to build the doc in the normal src build.
|
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
Modernize usage, i.e. copy from current gen_server.
|
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
Allows users see if the events was from the horizontal mouse wheel.
Have not changed the wxMouse record, would be incompatible, but this
allows the user to check for it in an event callback.
|
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
Some more enums have been added/changed definitions and can no longer
be constants in wx.hrl
|
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
Turned the tests the wrong way and compilation
failed with wxWidget-3.0 and wxMac.
|
|\ \ \ \ \ \
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | | |
maint
* peterdmv/inets/httpc_chunked_encoding/ERL-1090/OTP-16300:
inets: Fix handling of HTTP HEAD requests in httpc
|
| | |/ / / /
| |/| | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
httpc could not handle HEAD requests when the response contained
the header 'Transfer-Encoding: chunked' and no message-body.
According to RFC2616 a server MUST NOT return a message-body in
the response.
This commit fixes httpc while preserving its robust behavior to
gracefully handle a faulty response with chunked-encoded empty
message-body.
|
|\ \ \ \ \ \
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | | |
* hans/ftp/fix_warnings/OTP-16318:
ftp: Make var anonymous to remove erlc warning
|
| | | | | | | |
|
|\ \ \ \ \ \ \
| |/ / / / / /
|/| | | | | |
| | | | | | |
| | | | | | |
| | | | | | | |
* hans/tftp/fix_warnings/OTP-16317:
tftp: Fix -behaviour warnings
tftp: Fix erlang:get_stacktrace warning
|
| | | | | | | |
|
|/ / / / / / |
|
|\ \ \ \ \ \
| |_|_|/ / /
|/| | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
* rickard/nodedown-reason/OTP-16216:
net_kernel: save connection pending owners in map
net_kernel: save connection owners in map
Fix race causing nodedown reason to be lost
|
| | | | | | |
|
| | | | | | |
|
| | | | | | |
|
|\ \ \ \ \ \
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | | |
* hans/crypto/cuddle_tests:
crypto: Add prop__crypto_init_update to property tests
crypto: New common libfile for property tests in crypto
|
| | | | | | | |
|
| | | | | | | |
|
|\ \ \ \ \ \ \
| |_|_|_|_|_|/
|/| | | | | |
| | | | | | | |
maint
|
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | | |
OTP-16212
|
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | | |
Moved the net test suite into the kernel app,
where it belongs (since thats where the net
module exist).
OTP-16212
|
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | | |
Updated/changed according to review comments.
OTP-16212
|
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | | |
OTP-16212
|
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | | |
Added the sockaddr_ll used by family packet. And also documentation
for same.
OTP-16212
|
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | | |
OTP-16212
|
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | | |
OTP-16212
|
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | | |
Make it possible to *also* filter on flags in getifaddrs/1,2.
OTP-16212
|
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | | |
Added the function getifaddrs/1,2 that takes a filter
argument, which makes it possible to filter out only
those interfaces that "belong" to that specific
(socket) address family.
OTP-16212
|
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | | |
Add getifaddrs to the net module.
OTP-16212
|
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | | |
It works, but some of the packet address fields are still
not translated (just integers for now).
OTP-16212
|
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | | |
OTP-16212
|
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | | |
Some of the (internal) functions wher prefixed 'n' (net).
This has been changed to 'enet_' instead (this is to avoid
possible name clashes). And also to make the code slightly
more readable.
OTP-16212
|
|\ \ \ \ \ \ \
| | | | | | | |
| | | | | | | |
| | | | | | | | |
'bmk/kernel/enet/20191125/make_net_doc_always_included/OTP-16294' into maint
|
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | | |
When building kernal from its top directory, an
awk warning is generated (cd lib/kernel && make)
from make in the doc/src directory even though
nothing is actually built there.
The root cause is the target 'xmllint' and the
variables that are evaluated (in the otp_release_target.mk),
containing all xml files in kernel doc. Some of which
are *generated* (at a later stage) and therefor does
not yet exist...
Trying to always include the net documentation to
avoid the above generation.
OTP-16294
|
|\ \ \ \ \ \ \ \
| | | | | | | | |
| | | | | | | | |
| | | | | | | | |
| | | | | | | | | |
* ingela/inets/warnings:
inets: Remove legacy code
|
| | |_|/ / / / /
| |/| | | | | |
| | | | | | | |
| | | | | | | | |
Use module rand instead of random
|
|\ \ \ \ \ \ \ \
| | | | | | | | |
| | | | | | | | |
| | | | | | | | |
| | | | | | | | | |
* ingela/ssl/typo:
ssl: Fix typo
|