Age | Commit message (Collapse) | Author |
|
|
|
* refs/heads/fetch-and-merge.maint-r15-opu/FETCH_HEAD/11632: (73 commits)
Changed to correct errorcode in testcase
inets: Prepare for release
inets: Make test suites independent of each other
inets: Rewrite of test case to avoid timing issues in test code
Add missing brackets to report formatting on ftp_progress process exit
inets: Remove log message as it causes more harm than use at the moment
inets: Mend broken max_clients check
inets: Start CT'ify httpd_SUITE
inets: Remove use of default gen_server timeout
Fix http_request:http_headers/1 to send content-length when length is zero
Fix httpd config option 'keep_alive_timeout'
Fix httpd config option 'script_timeout'
inets: Restore ftp test files for the inets_{,sup_}SUITE to not fail
inets: Add crypto start check to ssl test cases
ftp: fix sockname dialyzer warning including ftp:sockname/1 bug
ftp,ssl: Fixes broken type link (ssloption).
ftp: Adds dynamic cert generation to tests.
ftp: Clean Makefile and conf file
ftp: Linking rfc-refs.
ftp: Add documentation.
...
|
|
|
|
|
|
|
|
Conflicts:
lib/inets/test/httpd_SUITE.erl
|
|
When progress reporting process exits, ftp client crashes
because of missing brackets in io_lib:format argument.
|
|
Conflicts:
lib/inets/src/http_lib/http_request.erl
|
|
Conflicts:
lib/inets/src/http_server/httpd_manager.erl
Conflicts:
lib/inets/test/inets_sup_SUITE.erl
|
|
Conflicts:
lib/inets/test/httpd_SUITE.erl
|
|
If you need a timeout between client and server use a server side
timeout instead of the default gen_server:call/2 timeout that is quite
useless. Also remove legacy code that has no further use.
|
|
In R16B01, the http_request:http_headers/1 function removes the content-length
field from the HTTP headers if the content length is zero. This results in
some (perhaps many) HTTP servers rejecting POSTs and PUTs without data with
a 411 status word. From RFC2616, section 14.13: "Any Content-Length greater than
or EQUAL to zero is a valid value".
|
|
The documentation states that the value is in seconds, which was true
when using the Apache like configuration file, but not true when using
the proplist style configuration.
|
|
This fixes the usage of the httpd configuration option 'script_timeout',
which got ignored before.
The documentation states that the value is in seconds, which was true
when using the Apache like configuration file, but not true when using
the proplist style configuration.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Conflicts:
lib/inets/test/ftp_suite_lib.erl
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Conflicts:
lib/eldap/test/eldap_basic_SUITE.erl
|
|
If a request handler terminates abnormally (terminate reason not
normal), the error-log entry was a bit obtuse ("Internal Server Error").
This has been improved so that the log entry now also include the
actual terminate reason.
|
|
Sometimes the size of a response could be as a string.
|
|
Remove logging when fo keep-alive connection timeout. This is
a normal event (feature) not an error.
|
|
|
|
Some error handling is only relevant for functions that
end up in gen_server:call and not in gen_server:cast.
|
|
Conflicts:
lib/inets/src/http_client/httpc_handler.erl
|
|
http_chunk:decode/4, the side-effecting chunk decoder, was only used
by httpc_handler:handle_http_body/2 but now it's completely unused.
Conflicts:
lib/inets/src/http_client/httpc_handler.erl
|
|
Previously the only difference between {stream, self} and {stream,
{self, once}} was an extra Pid in the stream_start message due to a
bug in httpc_handler. It was immediately sending a bunch of messages
till the end instead of waiting for httpc:stream_next/1 being called.
|
|
|
|
|
|
|
|
The option got ignored before 720721e.
|
|
|
|
The ones provided will fail inets test cases as default ssl will
negotiate stronger cipher suites now that will fail the with the weak keys.
|
|
Tests that the nocache options work and can be used independently from
each other. This was broken before c8ef69c.
Conflicts:
lib/inets/test/httpd_basic_SUITE.erl
|
|
|