Permalink
Commits on Sep 14, 2015
  1. Merge pull request #131 from aszlig/fix-features-on-required-starttls

    michalwski committed Sep 14, 2015
    connect: Check features when STARTTLS is required.
Commits on Sep 9, 2015
  1. Merge pull request #134 from esl/OTP18

    michalwski committed Sep 9, 2015
    Use OTP 18-fixed deps
  2. Use OTP 18-fixed deps

    fenek committed Sep 1, 2015
Commits on Sep 4, 2015
  1. connect: Check features when STARTTLS is required.

    aszlig committed Sep 3, 2015
    According to RFC6120 section 5.3.1, the server SHOULD NOT advertise any
    features before the client has done a successful <starttls/>:
    
    http://xmpp.org/rfcs/rfc6120.html#tls-rules-mtn
    
    Right now, MongooseIM returns the same feature list before and after the
    <starttls/>, which causes troubles in some clients like Gajim on
    Windows.
    
    While I haven't been able to fully debug why only the Windows version is
    trying to authenticate without STARTTLS, I believe we really _should_ do
    what the RFC recommends and that's what the check is for.
    
    Signed-off-by: aszlig <aszlig@redmoonstudios.org>
Commits on Aug 28, 2015
  1. Merge pull request #129 from esl/refactor-c2s-wait-for-stream

    michalwski committed Aug 28, 2015
    Refactor ejabberd_c2s:wait_for_stream/2
Commits on Aug 27, 2015
  1. Customize what modules to generate coverage stats for

    erszcz committed Aug 27, 2015
    This speeds iteration time a lot.
Commits on Aug 25, 2015
  1. Refactor invalid_host test

    erszcz committed Aug 25, 2015
Commits on Aug 24, 2015
Commits on Aug 21, 2015
  1. Merge pull request #121 from esl/riak-private

    ppikula committed Aug 21, 2015
    Riak private
Commits on Aug 20, 2015
  1. try a different approach

    ppikula committed Aug 20, 2015
  2. increase large_heap threshold in system_SUITE

    ppikula committed Aug 18, 2015
    This test sometimes fails, see the #127 issue in esl/ejabberd_tests. I found
    that the pid in assertion doesn't match. Propably some other tests leave large_heap
    and MiM generates 2 alarms in random order. So I increased threshold from
    100KB to 10MB to make sure, that only our designated process triggers the alarm.
Commits on Aug 19, 2015
  1. Merge pull request #128 from esl/log-path-fix

    ppikula committed Aug 19, 2015
    Log path fix
  2. Hopefully make metrics_api_SUITE:one_presence_error deterministic

    erszcz committed Aug 19, 2015
    Addressing the presence towards nbody@wronghost triggers the server
    to try a S2S outgoing connection which requires a DNS lookup.
    Such a lookup, it turns out, might take longer than the default timeout
    value in escalus:wait_for_stanza/1 (i.e. 1 second).
    
    Addressing the presence towards a still invalid (though causing a
    different error) JID, but one not triggering the federation mechanism
    should hopefully make the test more deterministic.
    
    Of course! In the boring case both variants (even the S2S one) can
    run for a hundred repetitions and nothing breaks...
Commits on Aug 18, 2015
  1. Expect updated status messages

    erszcz committed Jul 3, 2015
Commits on Aug 17, 2015
  1. Merge pull request #126 from esl/no-prepare-if-not-needed

    ppikula committed Aug 17, 2015
    Allow for easy skipping of the prepare Makefile dependency
  2. Merge pull request #118 from esl/simplify-mam-v1

    ppikula committed Aug 17, 2015
    Simplify mam v1
  3. Merge pull request #114 from esl/riak-base

    ppikula committed Aug 17, 2015
    Auth backends refactoring
Commits on Aug 13, 2015
  1. update riak config format

    ppikula committed Aug 12, 2015
  2. specify riak_config

    michalwski committed with ppikula Feb 12, 2015
  3. add riak_mnesia preset

    michalwski committed with ppikula Feb 11, 2015
Commits on Aug 6, 2015
  1. Allow for easy skipping of the prepare Makefile dependency

    erszcz committed Aug 6, 2015
    Just use `make quicktest PREPARE= TESTSPEC=default.spec`,
    i.e. assign an empty value to variable `PREPARE`.
  2. Merge pull request #125 from esl/escalus-rebase-specs

    michalwski committed Aug 6, 2015
    Escalus rebase specs
Commits on Aug 5, 2015
  1. Adhere to new Escalus API

    erszcz committed Aug 5, 2015
  2. Fix formatting

    erszcz committed Aug 5, 2015
  3. Update to Escalus with more specs and changed API (this needed fixes!)

    erszcz committed Aug 5, 2015
    The previous commits prepare for moving to this new version.