Age | Commit message (Collapse) | Author |
|
When using Yecc's operator precedence declarations (Right, Nonassoc,
Left), it is possible that some clauses of the generated code are
unreachable. Dialyzer complains, at least if SET_LIMIT in erl_types is
set sufficiently high (to avoid sets of integers to collapse to the
integer() type). Rather than trying to figure out exactly which
clauses are unreachable, Yecc generates a Dialyzer suppression.
|
|
into maint
* margnus1/dialyzer/fix_maps_opaque/ERL-249/PR-1161/OTP-13878:
erl_bif_types: Properly unopaque maps:merge/2 args
|
|
The files are not used in runtime of the snmp app, but they are still
part of the application and beam files are generated. Not including them
will make the general app test fail.
|
|
|
|
* ingela/diameter/cuddle-tests:
diameter: Do not test SCTP on sparc-sun-solaris2.10
|
|
The SCPT stack has a different behavior in solaris2.10 then in later versions
and linux. Diameter implementation does not support this behavior.
|
|
maint
* michaesc/crypto/honour-OPENSSL_NO_DES/ERL-203/OTP-13783:
Add '_cfb' alias for misspelled '_cbf' cipher
Fix badarg -> notsup and test cases
Honour OPENSSL_NO_DES
Conflicts:
lib/crypto/c_src/crypto.c
|
|
erl_bif_types:type/5 was calling erl_types:map_pairwise_merge/3 directly
with its (potentially opaque) arguments, causing Dialyzer crashes.
Bug (ERL-249) reported and minimised test case provided by Felipe
Ripoll.
|
|
* egil/hipe/fix-match-order-opaque-types/OTP-13876:
hipe: Fix erl_types opaque match order
|
|
* bjorn/compiler/lc-internal-failure/OTP-13863:
sys_core_fold: Don't move a fun into a guard
|
|
|
|
The path is only used when writing a comment so the correction
just allowed UTF-8 characters in that io_lib:format call.
IC don't support UTF-8 in the specs or their names
(generated modules contains coding: latin-1) but the path is
only used in a comment so it's not a problem if it can't be
written correctly as long as nothing fails.
|
|
OTP-13633
|
|
|
|
init_per_testcase timeout for renegotiation tests would be overridden
by local timeout in test case help function.
|
|
|
|
* ingela/ssl/dtls-progress/connection-states-as-maps:
dtls: fix encoding of client hello cookie
dtls: Prepare start of DTLS connection manager with SSL app
ssl: Refactor to use maps for the connection states
ssl, dtls: Refactor sni handling
dtls: Add close/5
dtls: Add renegotiate/2
dtls: Add protocol event handling
ssl: Refactor code so that tls and dtls can share more code
ssl, dtls: Disable V2 compatibility clause from ssl_handshake:update_handshake_history
ssl: Make sure common code for TLS and DTLS uses the TLS Version
ssl: remove unused RecordCB argument from master_secret
dtls: Add reinit_handshake_data/1 to dtls
dtls: replace tls_record with RecordCB in connection_info
Fix version numbers and dependencies
|
|
|
|
Conflicts:
lib/ssh/src/ssh_auth.erl
|
|
|
|
|
|
* ingela/ssl/ssl_to_openssl_SUITE:
ssl: Make sure to have correct default
|
|
* lukas/ts/no_flush_on_halt:
ts: Don't flush port when ending test run
|
|
|
|
* lukas/kernel/os_cmd_fix_stdin_close/OTP-13867:
kernel: Close stdin of commands run in os:cmd
|
|
into maint
* lukas/kernel/fix_os_SUITE_message_leak_testcase/OTP-13867:
kernerl: Remove infinite loop from testcase
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
We want to share more alert and application data handling code.
Some of the application data handling code, packet handling,
will not be relevant for dtls, but this code can be excluded from dtls
by options checking.
|
|
ssl_handshake:update_handshake_history
This proably a much bigger problem for DTLS than TLS, but should be
disabled for both unless explicitly configured for TLS.
|
|
When protocol version is proagated from the DTLS connection processes
state into general ssl functions it must be converted to the corresponding
TLS version.
|
|
Conflicts:
lib/ssl/src/ssl_handshake.erl
|
|
The callback is invoke before entering state 'connection'.
It allows a connection module to remove data from the
connection state that is no longer needed (e.g.
handshake history).
|
|
Conflicts:
lib/ssl/src/ssl_connection.erl
|
|
|
|
|
|
|
|
Moving a fun into a guard may cause code that is not accepted
by beam_validator.
|
|
* ingela/ssl/test-framework-checks:
ssl: Test ssl v2 clients rejection depending on configuration
ssl: Add check in test framework for crypto support
|
|
ssh_msg_userauth_info_request
This behavour is assumed by Codenomicon Defensics.
|
|
This is needed when running programs that only exit when
stdin has been closed, e.g. 'more'.
|
|
Even though v2 is never supported v2 hellos can be.
No support for v2 client hellos gives "handshake failiure" alert.
Support for v2 hello but no higher SSL/TLS version offered
gives "protocol version" alert.
|
|
|