diff options
author | Pierre-Charles David | 2017-06-29 08:18:17 +0000 |
---|---|---|
committer | Pierre-Charles David | 2017-06-29 08:23:30 +0000 |
commit | 6d22567e96e95ebd9ee69382ac5de55c9e7c5d12 (patch) | |
tree | 09ed8a9a729cb22079368a8fd0297f920d4b8fb3 /packaging/org.eclipse.sirius.samples/pom.xml | |
parent | fc8a1dc3b93a3e8f0fda2a1f17a7da4b894c9f98 (diff) | |
download | org.eclipse.sirius-6d22567e96e95ebd9ee69382ac5de55c9e7c5d12.tar.gz org.eclipse.sirius-6d22567e96e95ebd9ee69382ac5de55c9e7c5d12.tar.xz org.eclipse.sirius-6d22567e96e95ebd9ee69382ac5de55c9e7c5d12.zip |
[518259] Update Session1MillionTests and move it out of the Gerrit suite
Commit 8cf04acad92e8646077e69a5c869923168c30f54 forces the
installation of the cross-referencer during the session opening to fix
potential bugs and more generally have a more reliable/predictable
behavior. It causes performances regressions when opening sessions
with big semantic models which happen to have no viewpoint currently
enabled. This is probably rare in practice, but it affects at least
the test session used by Session1MillionTests, causing it to take
around 100s to open (from less that 15s normally).
The "regression" only concerns artificial corner cases, which happen
to apply to Session1MillionTest, but:
* it fixes a real bug that can be visible to end-users;
* it only highlights that the conditions tested by Session1MillionTest
does not really match normal use cases situation, and the times it
checks do not correspond to what real users would see anyway.
Note that even without commit 8cf04acad92e864, enabling a single
viewpoint in the test aird increases the session opening time to
~100s.
Bug: 518259
Change-Id: I4f3bbf03480d5c787f429ad8240de981d9b37fea
Signed-off-by: Pierre-Charles David <pierre-charles.david@obeo.fr>
Diffstat (limited to 'packaging/org.eclipse.sirius.samples/pom.xml')
0 files changed, 0 insertions, 0 deletions