Matti Tahvonen [Fri, 4 Oct 2013 12:41:06 +0000 (15:41 +0300)]
Only fetch rows if there are some (#11189)
IE hacks cause calls to onScroll in situations where the cache row fetch
logic is not working correctly (causes JS exception). This change has an
optimization to pass this logic if there are no rows available and this
way fixes the JS exception as well.
Artur Signell [Mon, 7 Oct 2013 16:56:43 +0000 (19:56 +0300)]
Makes test stable and adds helper comparison methods
* Takes into account that the timer can be triggered multiple times before the initial request is done (especially if the server is slow)
* Adds assertLessThan, assertLessThanOrEqual, assertGreaterThan, assertGreaterThanOrEqual to ease test creation
Artur Signell [Mon, 7 Oct 2013 15:12:06 +0000 (18:12 +0300)]
Update TB3+ convention to use a separate *Test file
For any issue there should be a UI class, e.g. LabelSomething and a
Test class which has the name of the UI + Test, e.g. LabelSomethingTest.
The LabelSomethingTest contains ALL tests which depend on LabelSomething
and are named accordingly to what they test.
For special and legacy cases also LegacyApplication and UIProvider are
supported in addition to UI classes
John Ahlroos [Mon, 7 Oct 2013 09:37:50 +0000 (12:37 +0300)]
Fixed Table range selection IE regression #12407
After fixes for #12407 the range selection did not work in the case
where the selection start had previously been removed. This caused
MultiSelectWithRemovedRow test to fail on IE.
Artur Signell [Thu, 12 Sep 2013 17:57:01 +0000 (20:57 +0300)]
Prepare for Atmosphere Javascript 2.0 (#12241)
* Atmosphere 2.0 has changed enableProtocol to default to true.
* Reopening a connection in Atmosphere 2.0 is signalled through an onReopen event (not present at all in Atmosphere 1.0.x)
Fix was backported from master
Change-Id: I6ed258087a0b3a06440ab9d19b621560fa4f998f
Merge: no
Removed some obsolete (hopefully!?) code doing some odd things with caption height calculation and some refactoring to make that part of code slightly more readable.
Artem Godin [Wed, 2 Oct 2013 14:07:14 +0000 (17:07 +0300)]
Fix OptionGroup elements losing focus on value change (#10451)
The misbehavior was caused by VOptionGroup.buildOptions recreating
associated panel on every change by removing and adding new elements.
With this fix applied it tries to update existing elements,
distinguishing them by assigned keys. It will recreate panel though if
elements are reordered or new elements were added/removed.
Artur Signell [Fri, 27 Sep 2013 10:14:36 +0000 (13:14 +0300)]
Refactor how TB3 tests are written (#12572)
* openTestUrl() must be called in the beginning of each test. Before that setPush and setDebug can be used to determine whether /run-push or ?debug should be used in the URL
John Ahlroos [Thu, 19 Sep 2013 13:42:50 +0000 (16:42 +0300)]
Fixes issue with Table not scrolling completely to the end #12651
Made the Table notice if the user is trying to scroll to an item on the
last "page" and in those cases actually scroll to that item, not just to
the page's first item as it did before.
Artur Signell [Mon, 16 Sep 2013 11:31:37 +0000 (14:31 +0300)]
Refactored build scripts to support TB2, TB3 + integration tests (#12572)
* Main build configuration triggers unit tests for all modules and uitest/build.xml testbench tests in parallel
* uitest/build.xml triggers Jetty startup and integration (server) tests in parallel. After the server has started, TB2 and TB3 tests are run in parallel.
* Server integration tests for servlet containers are run using TB3 and the com.vaadin.tests.tb3.ServletIntegrationTests test suite.
* Portlet integration tests are still run using TB2 test scripts
Artur Signell [Wed, 18 Sep 2013 21:18:19 +0000 (00:18 +0300)]
Update release notes to more closely be like 7.0.x (#12544)
* Fixed ticket link and removed duplicate link
* Reordered sections to changes in this version comes before changes in old versions
* Updated browser versions
NullPointerException in DateToSqlDateConverter (#12284)
DateToSqlDateConverter throws a NullPointerException when the provided
value is null, thus violating the interface contract. If the provided
value is null, then the methods should return null.
Fabian Lange [Thu, 12 Sep 2013 12:16:37 +0000 (14:16 +0200)]
fixed incorrect name for close-pressed.png for windows in black theme (#12563)
The scss file and the light and normal style suggest that the name should be
"-pressed". This is what this patch fixes. I have considered renaming all
"-pressed" to "-active", to be consistent with e.g. "maximize-active.png"
but decided against due to compatibility concerns.
Fabian Lange [Mon, 2 Sep 2013 13:49:53 +0000 (15:49 +0200)]
Allow creating TextBox or SuggestionPopup when extending VFilterSelect (#12491)
This patch adds two methods, that one can override when extending VFilterSelect.
By doing so, the developer can for example change the behavior of setting text,
or performing clicks in the suggestion popup.
This change is backwards compatible, as it just offers two new methods to
override.
TestHideTimeAndSeparator didn't specify a locale, which
caused it to fail if the locale on the system where it
ran was different from the one on the system where the
test was created.
Henrik Paul [Tue, 20 Aug 2013 12:47:54 +0000 (15:47 +0300)]
Table ignores Container updates while painting (#12258)
Vaadin threw an IllegalStateException if a Container was updated while
Table was being painted. SQLContainer was known to invalidate its cached
size during a Table repaint, resulting in an ItemSetChangeEvent. This fix
has been copied over from how ComboBox handles this situation.
Avoid leaking memory from inherited ThreadLocales. Fixes #12401
The issue is fixed by changing the normal HashMap inside the inheritable
thread local to a map implementation holding only weak references to the
values (WeakValueMap).
Also included is a test UI that starts threads, which run until the JVM
is quit. This along with VisualVM was used to reproduce the issue and
verify the fix.
Henri Sara [Wed, 28 Aug 2013 09:34:53 +0000 (12:34 +0300)]
Use the add() path of CssLayout only when appending to the end (#11284)
This should get the benefits of the previous optimization on initial
rendering without negatively affecting the performance when modifying a
layout with complex non-leaf children.