Today
-
r6834
(Major refactor of CompilationState and associated classes to...) committed by scottb@google.com
- Major refactor of CompilationState and associated classes to fix concurrent refresh problems.
Key points:
- CompilationState is no longer a singleton-per-module. Instead, each call to
module.getCompilationState() produces a new CompilationState. These instances appear completely
independent, but internally may reuse some shared state. CompilationState cannot be refreshed, but
a new state must be created.
- To accomplish this, CompilationState is now created through a CompilationStateBuilder, which
compiles and validates incoming source files, possibly reusing internally-cached objects in the
process. A key aspect of this process is maintaining an accurate dependency graph between units, to
ensure that units are recompiled whenever their dependencies change. This ensures that things like
refresh (which is really building a new CompilationState with a hot cache) remains fast.
- CompilationUnit is now stateless. The formerly stateful aspects are now somewhat handled by
CompilationUnitBuilder, combined with CompilationStateBuilder.CompileMoreLater, and internal state
of JdtCompiler itself.
- Building compilation state now somewhat combines JsniCollector and JsniChecker. A full JSNI parse
is now performed, and any parse, JS, or custom JSNI errors are recorded in the compilation unit.
JSNI references to other units are also recorded.
Suggestions by: bruce, knorton, bobv, jat
Patch by: me, bobv (pair programming)
Review by: bobv
Major refactor of CompilationState and associated classes to fix concurrent refresh problems.
Key points:
- CompilationState is no longer a singleton-per-module. Instead, each call to
module.getCompilationState() produces a new CompilationState. These instances appear completely
independent, but internally may reuse some shared state. CompilationState cannot be refreshed, but
a new state must be created.
- To accomplish this, CompilationState is now created through a CompilationStateBuilder, which
compiles and validates incoming source files, possibly reusing internally-cached objects in the
process. A key aspect of this process is maintaining an accurate dependency graph between units, to
ensure that units are recompiled whenever their dependencies change. This ensures that things like
refresh (which is really building a new CompilationState with a hot cache) remains fast.
- CompilationUnit is now stateless. The formerly stateful aspects are now somewhat handled by
CompilationUnitBuilder, combined with CompilationStateBuilder.CompileMoreLater, and internal state
of JdtCompiler itself.
- Building compilation state now somewhat combines JsniCollector and JsniChecker. A full JSNI parse
is now performed, and any parse, JS, or custom JSNI errors are recorded in the compilation unit.
JSNI references to other units are also recorded.
Suggestions by: bruce, knorton, bobv, jat
Patch by: me, bobv (pair programming)
Review by: bobv
-
r6833
(Refactors JSNI parsing/error reporting from GenerateJavaAST ...) committed by scottb@google.com
- Refactors JSNI parsing/error reporting from GenerateJavaAST to JsniCollector.
This is to prepare for making JsniCollector the soure of truth for JSNI parsing.
Review by: bobv
Refactors JSNI parsing/error reporting from GenerateJavaAST to JsniCollector.
This is to prepare for making JsniCollector the soure of truth for JSNI parsing.
Review by: bobv
-
-
r6831
(Moving several files from com.google.gwt.dev.javac.impl -> c...) committed by scottb@google.com
- Moving several files from com.google.gwt.dev.javac.impl -> com.google.gwt.dev.javac
Having a subpackage was kind of pointless.
Review by: bobv (desk)
Moving several files from com.google.gwt.dev.javac.impl -> com.google.gwt.dev.javac
Having a subpackage was kind of pointless.
Review by: bobv (desk)
-
r6830
(Adding some useful PerfLogger messages.
Review by: bobv) committed by scottb@google.com
- Adding some useful PerfLogger messages.
Review by: bobv
Adding some useful PerfLogger messages.
Review by: bobv
-
r6829
(Remove spammy logging from SoycTest.
Review by: bobv) committed by scottb@google.com
- Remove spammy logging from SoycTest.
Review by: bobv
Remove spammy logging from SoycTest.
Review by: bobv
-
r6828
(Build improvements for Chrome plugin.
) committed by tamplinjohn
- Build improvements for Chrome plugin.
Build improvements for Chrome plugin.
-
r6827
(Move licence where it will be included in the plugin, update...) committed by tamplinjohn
- Move licence where it will be included in the plugin, update manifest.
Move licence where it will be included in the plugin, update manifest.
-
-
r6825
(Fix NPE when an unknown type is referenced in a class signat...) committed by tamplinjohn
- Fix NPE when an unknown type is referenced in a class signature (this could be
a binary-only annotation or an elided local class).
Patch by: jat
Review by: scottb (desk)
Fix NPE when an unknown type is referenced in a class signature (this could be
a binary-only annotation or an elided local class).
Patch by: jat
Review by: scottb (desk)
-
r6824
(Update Chrome windows binary, update MIME type for name chan...) committed by tamplinjohn
- Update Chrome windows binary, update MIME type for name change, fix missed
name change in exports.
Update Chrome windows binary, update MIME type for name change, fix missed
name change in exports.
-
r6823
(Checkin remaining Linux binaries and rebuilt XPI.
Patch by:...) committed by tamplinjohn
- Checkin remaining Linux binaries and rebuilt XPI.
Patch by: jat
Review by: jlabanca
Checkin remaining Linux binaries and rebuilt XPI.
Patch by: jat
Review by: jlabanca
-
r6822
(Update windows-specific preference code (used in IE and Chro...) committed by tamplinjohn
- Update windows-specific preference code (used in IE and Chrome) to new name.
Patch by: jat
Review by: jlabanca (TBR)
Update windows-specific preference code (used in IE and Chrome) to new name.
Patch by: jat
Review by: jlabanca (TBR)
-
-
-
-
r6820
(Update Linux FF alternate-library binaries.
) committed by tamplinjohn
- Update Linux FF alternate-library binaries.
Update Linux FF alternate-library binaries.
-
r6819
(Update Windows FF binaries.
) committed by tamplinjohn
- Update Windows FF binaries.
Update Windows FF binaries.
-
r6818
(Replace tabs with spaces.
Patch by: jat
Unreviewed
) committed by tamplinjohn
- Replace tabs with spaces.
Patch by: jat
Unreviewed
Replace tabs with spaces.
Patch by: jat
Unreviewed
-
r6817
(Change top-window calculation to not follow the opener trail...) committed by tamplinjohn
- Change top-window calculation to not follow the opener trail, and checkin
Mac binaries with this change.
Note that this change will affect the URL that is returned to the devmode UI
in the case where a GWT app opens a new window containing a different GWT app.
Previously they would have been collapsed as two modules under the same URL,
but now they will be separate top-level URLs. The reason for this change is
that we have run into too many issues with Selenium and other unusual ways
apps get opened, and getting it nailed down solidly will take more time than
we have.
Patch by: jat
Review by: jlabanca (TBR)
Change top-window calculation to not follow the opener trail, and checkin
Mac binaries with this change.
Note that this change will affect the URL that is returned to the devmode UI
in the case where a GWT app opens a new window containing a different GWT app.
Previously they would have been collapsed as two modules under the same URL,
but now they will be separate top-level URLs. The reason for this change is
that we have run into too many issues with Selenium and other unusual ways
apps get opened, and getting it nailed down solidly will take more time than
we have.
Patch by: jat
Review by: jlabanca (TBR)
-
-
-
-
issue 4214
(DynaTable RPCs fail outside dev mode) changed by bobv@google.com
-
Status:
FixedNotReleased
Labels:
Milestone-2_0_RC1 Category-RPC
Status:
FixedNotReleased
Labels:
Milestone-2_0_RC1 Category-RPC
-
-
r6815
(Missed update to GenerateJavaAST for final/volatile consiste...) committed by bobv@google.com
- Missed update to GenerateJavaAST for final/volatile consistency.
Patch by: bobv
Review by: scottb (TBR)
Missed update to GenerateJavaAST for final/volatile consistency.
Patch by: bobv
Review by: scottb (TBR)
-
-
r6814
(Mark artificially-rescued fields as volatile.
Patch by: bob...) committed by bobv@google.com
- Mark artificially-rescued fields as volatile.
Patch by: bobv
Review by: scottb (desk)
Mark artificially-rescued fields as volatile.
Patch by: bobv
Review by: scottb (desk)
-
r6813
(<g:StackLayoutPanel> is now consistent with <g:TabLayoutPane...) committed by rjrjr@google.com
- <g:StackLayoutPanel> is now consistent with <g:TabLayoutPanel> and
<g:DisclosurePanel> in its use of <g:header> and <g:customHeader>.
The parsers for StackLayoutPanel, DockLayoutPanel and TabLayoutPanel
now all treat their unit (or barUnit) attributes as optional, and
default them to PX
Also, yet another pass at making the generator more strict: if any extra
attributes, elements or body text are left after all the parsers are run, the
user is scolded.
<g:StackLayoutPanel> is now consistent with <g:TabLayoutPanel> and
<g:DisclosurePanel> in its use of <g:header> and <g:customHeader>.
The parsers for StackLayoutPanel, DockLayoutPanel and TabLayoutPanel
now all treat their unit (or barUnit) attributes as optional, and
default them to PX
Also, yet another pass at making the generator more strict: if any extra
attributes, elements or body text are left after all the parsers are run, the
user is scolded.
-
-
issue 3969
(ServerSerializationStreamReader has a critical bug when deal...) commented on by rares.axz
-
-
r6812
(Merging trunk@6789, 6790, 6791 into this branch
for i in 67...) committed by amitman...@google.com
- Merging trunk@6789, 6790, 6791 into this branch
for i in 6789 6790 6791 ; do svn merge --ignore-ancestry -c $i https://google-web-
toolkit.googlecode.com/svn/trunk .; done
Patch by: scottb
Merging trunk@6789, 6790, 6791 into this branch
for i in 6789 6790 6791 ; do svn merge --ignore-ancestry -c $i https://google-web-
toolkit.googlecode.com/svn/trunk .; done
Patch by: scottb
-
issue 4175
(Custom PluralRule is silently ignored when it's a nested cla...) Status changed by tamplinjohn
-
-
-
r6810
(Until soft perms land and we can do this properly, a hack to...) committed by rjrjr@google.com
- Until soft perms land and we can do this properly, a hack to keep Ie7 users from
paying the AlphaImageLoader penalty.
Reviewed by jgw
Until soft perms land and we can do this properly, a hack to keep Ie7 users from
paying the AlphaImageLoader penalty.
Reviewed by jgw
-
r6809
(Fix custom plural rules that are inner classes, and adds a t...) committed by tamplinjohn
- Fix custom plural rules that are inner classes, and adds a test for that case.
Issue: 4175
Patch by: jat
Review by: rice
Fix custom plural rules that are inner classes, and adds a test for that case.
Issue: 4175
Patch by: jat
Review by: rice
-
r6808
(Merging trunk@6782, 6783, 6784, 6785 into this branch.
for...) committed by amitman...@google.com
- Merging trunk@6782, 6783, 6784, 6785 into this branch.
for i in 6782 6783 6784 6785 ; do svn merge --ignore-ancestry -c $i https://google-web-
toolkit.googlecode.com/svn/trunk .; done
Patch by: scottb
Merging trunk@6782, 6783, 6784, 6785 into this branch.
for i in 6782 6783 6784 6785 ; do svn merge --ignore-ancestry -c $i https://google-web-
toolkit.googlecode.com/svn/trunk .; done
Patch by: scottb
-
issue 4175
(Custom PluralRule is silently ignored when it's a nested cla...) changed by tamplinjohn
-
Status:
ReviewPending
Labels:
Category-I18N Milestone-2_0_RC1 Type-Defect ReviewBy-rice
Status:
ReviewPending
Labels:
Category-I18N Milestone-2_0_RC1 Type-Defect ReviewBy-rice
-
r6807
(Rebuild mac binaries for FF plugin, add convenience make tar...) committed by tamplinjohn
- Rebuild mac binaries for FF plugin, add convenience make target.
Rebuild mac binaries for FF plugin, add convenience make target.
-
r6806
(Updates branch-info.txt to record the merge of trunk@6804 in...) committed by rdayal@google.com
- Updates branch-info.txt to record the merge of trunk@6804 into this branch.
Updates branch-info.txt to record the merge of trunk@6804 into this branch.
-
-
-
-
-
-
r6804
(Fixes a problem where the DevMode server was hanging when th...) committed by rdayal@google.com
- Fixes a problem where the DevMode server was hanging when the remote UI was shut
down in an unclean fashion. We now set a keepalive on the Remote UI socket, and
when an exception occurs, a callback (passed in to the MessageTransport
constructor) is invoked which causes the DevMode server to terminate.
Also fixed a bug in BrowserListener where the setReuseAddress option was being
set on the wrong socket.
Review by: mmendez, jat
Fixes a problem where the DevMode server was hanging when the remote UI was shut
down in an unclean fashion. We now set a keepalive on the Remote UI socket, and
when an exception occurs, a callback (passed in to the MessageTransport
constructor) is invoked which causes the DevMode server to terminate.
Also fixed a bug in BrowserListener where the setReuseAddress option was being
set on the wrong socket.
Review by: mmendez, jat
-
-
r6803
(Extended timeouts of HttpSuite tests to 15 seconds because t...) committed by jlabanca@google.com
- Extended timeouts of HttpSuite tests to 15 seconds because they sometimes fail if the server is
busy.
Patch by: jlabanca
Review by: fabbott
Extended timeouts of HttpSuite tests to 15 seconds because they sometimes fail if the server is
busy.
Patch by: jlabanca
Review by: fabbott
-
-
|
