Age | Commit message (Collapse) | Author |
|
|
|
|
|
If a child fails to start, supervisor relies upon error_logger which does not
work when IO is inhibited. Instead pass the error up the chain and let someone
else use a proper Reason for any possible printouts.
|
|
* ia/ssh/sending-large-data-and-eof/OTP-10467:
ssh: Make test case platform independant
ssh: ssh_connection:channel_data() and send_eof() now return {error, closed} for closed or invalid channels.
ssh: Ensure that all data is sent before ssh:send returns.
ssh: add test suite for testing channel send/receive behavior.
|
|
* ia/ssl/sslv2-test:
ssl: Better test to determine if sslv2 is supported by openssl
|
|
The test case interrupted_send did not work on all platforms.
Also ct-ify the test suite
|
|
for closed or invalid channels.
ssh_connection:handle_msg(#ssh_msg_channel_close...) will now cause any send() that was in progress to immediately return {error,closed}.
|
|
|
|
|
|
* gp/make_erlang_mode_elpa_friendly:
Make erlang-mode more compatible with package.el
OTP-10465
|
|
|
|
* sa/dialyzer-fix-r15b02:
Fix a crash in race condition detection
Remove old untested experimental extension
Respect {plt_check,false} option when using dialyzer:run/1
Fix handling of tuple set remote types appearing in tuple sets
OTP-10464
|
|
* bjorn/test_server/fix-test-cases:
Provided better information when the test cases fail
Remove the useless test_server_line_SUITE test suite
test_server_SUITE: Remove the undefined_functions/1 test case
|
|
* raimo/inets/https-proxy/OTP-10256:
inets httpc: Document that RFC 2817 is not supported
inets: Dialyzer fixes
inets httpc: TLS via proxy
inets: Remove obsolete testcases that are now in the proxy suite
inets: Add tests using external proxy running on localhost
|
|
|
|
|
|
Introduces new option htts_proxy so that it is possible
to have different proxies for http and https. The new option
defaults to the old proxy option.
Implements HTTP-1.1 extension method CONNECT to establish SSL/TLS tunnel
We choose not to implement "TLS upgrade" as defined by RFC 2817 as
this method of upgrade is vulnerable to man in the middle attacks,
can be easily broken by proxies and does not seem to be widely adopted.
|
|
|
|
|
|
|
|
* dg/keyboard_interactive:
fix typo in keyboard-interactive string
OTP-10456
|
|
While at it, also make the path to the directory with the log files
a clickable link.
|
|
* siri/common_test/ct_snmp-tests:
[common_test] Added tests and corrected bugs in ct_snmp
OTP-10454
OTP-10434
|
|
* siri/os_mon/undefined-snmp-funcs/OTP-10448:
[os_mon] Remove call to deprecated/undefined functions in snmp
|
|
* vs/re_back_reference:
extend re back reference syntax with \g escape sequence
OTP-10455
|
|
* bjorn/test_server/dialyzer-fixes:
test_server_ctrl: Don't use undocumented features of io:format()
Fix obsolete assumption for the return type of os:type/0
|
|
OTP-10454
OTP-10434
Function register_users/2, register_agents/2 and register_usm_users/2,
and the corresponding unregister_*/1 functions were not
executable. These are corrected/rewritten.
Function update_usm_users/2 is removed, and an unregister function is
added instead. Update can now be done with unregister_usm_users and
then register_usm_users.
Functions unregister_*/2 are added, so specific users/agents/usm users
can be unregistered.
Function unload_mibs/1 is added.
Overriding configuration files did not work, since the files were
written in priv_dir instead of in the configuration dir
(priv_dir/conf). This has been corrected.
Minor updates to documentation.
|
|
* hb/stdlib/opaque_warnings/OTP-10436:
Refine a few opaque types
Warn for underspecified opaque types
Warn for opaque types that are not exported
|
|
* hb/dialyzer/nowarn_unused_fun/OTP-10433:
dialyzer: Do not output warnings for unused funs
|
|
* dgud/ct/fix-trashed-logs/OTP-10369:
[ct] Shorten testcase names
Shorten testcase names
[ct] Block output into single io call
|
|
* ia/ssl/npn/OTP-10361:
ssl: Shorten test case names to workaround ct shortcomings on windows
ssl: SSL 3.0 does not support next protocol negotiation
ssl: Dialyzer fixes and code cleaning
ssl: Changed default behaviour of next protocol negotiation to make more "sense" (be true to the specification).
ssl: Update SSL docs for SSL Next Protocol Support
ssl: Support for SSL Next Protocol Negotiation * http://technotes.googlecode.com/git/nextprotoneg.html
|
|
|
|
|
|
|
|
The Format string argument for io:format/3 is not documented
to accept an iolist, so we should not depend on it. Fix the
problem by using io:put_chars/2 instead.
|
|
The test_server_line parse transform module was removed in R15
(in f43c0a51cd15b2b0f8adba4bb9ec5531dd9d8820), but not the
corresponding test cases. (Strangely enough, the test cases did not
fail; it was because init_per_testcase/2 failed and the corresponding
test case was skipped.)
|
|
The undefined_functions/1 test case will fail if it is run
using an uninstalled Erlang/OTP system (i.e. in a source
tree/git repository).
Since the test_server application is now a part of Erlang/OTP
(it was not originally), the system-wide test case that looks
for use of undefined functions will catch undefined functions.
Therefore the easiest solution is to remove the test case.
|
|
|
|
* siri/reltool/spawn_executable/OTP-10358:
[reltool] Use spawn_executable instead of spawn when starting nodes in test
|
|
So it works on windows otherwise crashes the testrun when
path is longer than 255 chars.
|
|
* ia/ssl/reuse_session_expired-test:
ssl: Adopt test case to not take so long
|
|
* ia/ssh/empty-binary/OTP-9478:
ssh: Will not hang if you attempt to send empty binary
|
|
Also, rewrite test suite (os_mon_mib_SUITE) to use ct_snmp intead of
using snmp application directly.
|
|
* bjorn/common_test/dialyzer-fixes:
ct_master_logs: Don't use io:format/3 with an empty variable list
common_test: Don't use undocumented features of io:format()
|
|
Analysis is not always able to find the variable labels and names for any kind
of arguments passed to the ets module functions.
|
|
This extension allowed minimal detection of race conditions that were passing
through behaviour API calls, but had no tests in any suite and also caused a
crash in R15B02.
|
|
|
|
Originally reported by Joseph Wayne Norton, simplified testcase provided by Hans
Bolinder.
|
|
os:type/0 could (a long, long time ago) return atoms such as
'vxworks' instead of a two-tuple.
|
|
|