Re: [webkit-dev] [PATCH] Errors from build-webkit --debug --minimal

2012-06-26 Thread Arthur O'Dwyer
Perhaps someone should set up a random bot that will select
uniformly at random from the set of *all* supported configurations. If
a configuration is known to be unsupported, or is found to be so buggy
that it is untenable, the configuration should be removed from the
documentation. After all, the point of having all these build switches
is so that the user (me) can customize the build. If the switches
aren't tested, then they don't work; and if they don't work, there's
no point in having them.

If I just want to build the core of WebKit2 (omitting optional
components where possible), what *is* the recommended approach?
Obviously --minimal is a bad idea right now. Also, should I continue
passing --debug? It seemed like some of the errors went away without
--debug.

-Arthur


On Mon, Jun 25, 2012 at 12:56 PM, Eric Seidel e...@webkit.org wrote:
 If someone wanted to set up a --minimal bot, I'm sure that would be
 welcome.  We used to have a --qt --minimal bot at some point.

 http://trac.webkit.org/wiki/BuildBot

 On Mon, 25 Jun 2012 12:03:12 -0700, Arthur O'Dwyer art...@push.am wrote:

 I'm trying to build WebKit by the simplest possible method. I checked
 out the SVN tree and did cd WebKit ; Tools/Scripts/build-webkit
 --debug --minimal. This produced compiler errors!

 I tried the same thing a while ago, and submitted patches to fix some of the
 issues, but what i was told is that there aren't really any bots building
 with --minimal, so it's not guaranteed to build.

 Last i remember, after you fix those issues you mention you're probably
 gonna come up against trickier ones in --web-audio, --geolocation,
 --netscape-plugin-api and --fullscreen-api [...]
___
webkit-dev mailing list
webkit-dev@lists.webkit.org
http://lists.webkit.org/mailman/listinfo.cgi/webkit-dev


Re: [webkit-dev] [PATCH] Errors from build-webkit --debug --minimal

2012-06-26 Thread Darin Adler
This is the wrong mailing list to discuss this kind of issue. On this list we 
discuss contributions to the WebKit project and collaborate on WebKit 
development. If you want to build WebKit and do not intend to contribute 
patches, you can ask questions on the webkit-help mailing list as described 
here http://www.webkit.org/contact.html, or elsewhere.

As far as configurations go, despite having some configuration options in the 
build scripts for convenience of the people working on the project, the 
contributors to the WebKit project do not intend to support arbitrary 
combinations of configuration options for WebKit users, even though I’m sure 
that would be handy. Generally we have a configuration for each port that is 
tested, and don’t spend our time on the additional work that would be needed to 
support the other combinations of options and keep them working.

-- Darin
___
webkit-dev mailing list
webkit-dev@lists.webkit.org
http://lists.webkit.org/mailman/listinfo.cgi/webkit-dev


Re: [webkit-dev] [PATCH] Errors from build-webkit --debug --minimal

2012-06-26 Thread Arthur O'Dwyer
On Tue, Jun 26, 2012 at 12:05 PM, Darin Adler da...@apple.com wrote:
 This is the wrong mailing list to discuss this kind of issue. On this list we 
 discuss contributions to
 the WebKit project and collaborate on WebKit development. If you want to 
 build WebKit and do not
 intend to contribute patches, you can ask questions on the webkit-help 
 mailing list as described
 here http://www.webkit.org/contact.html, or elsewhere.

If I could build WebKit, I wouldn't need to contribute patches. ;)
See my original post in this thread for an example of a patch that I
was forced to contribute. However, I'll go away now.

-Arthur
___
webkit-dev mailing list
webkit-dev@lists.webkit.org
http://lists.webkit.org/mailman/listinfo.cgi/webkit-dev


[webkit-dev] [PATCH] Errors from build-webkit --debug --minimal

2012-06-25 Thread Arthur O'Dwyer
I'm trying to build WebKit by the simplest possible method. I checked
out the SVN tree and did cd WebKit ; Tools/Scripts/build-webkit
--debug --minimal. This produced compiler errors!

The first two errors were easy to fix; it was just a case of features
being incompletely #ifdef'ed. I've quoted the patch at the bottom of
this message, since I'm not sure yet if this mailing list accepts
attachments.


This next error, related to ENABLE(JAVASCRIPT_DEBUGGER), I can't
figure out how to fix. So I'm passing build-webkit --debug --minimal
--javascript-debugger as a workaround.

/Users/ajo/Desktop/WebKit/Source/WebCore/bindings/js/JSConsoleCustom.cpp:41:20:
error: out-of-line definition of 'profile' does not match any
declaration in 'WebCore::JSConsole'
JSValue JSConsole::profile(ExecState* exec)
   ^~~
/Users/ajo/Desktop/WebKit/Source/WebCore/bindings/js/JSConsoleCustom.cpp:52:20:
error: out-of-line definition of 'profileEnd' does not match any
declaration in 'WebCore::JSConsole'
JSValue JSConsole::profileEnd(ExecState* exec)
   ^~
2 errors generated.


Then I get this fourth error, many many times. There's no
--dashboard-support option to build-webkit, so I'm stumped.

perl -I WebCore/bindings/scripts/
WebCore/bindings/scripts/generate-bindings.pl --include WebCore/dom
--include WebCore/fileapi --include WebCore/html --include WebCore/css
--include WebCore/page --include WebCore/notifications --include
WebCore/xml --include WebCore/svg --write-dependencies --outputDir .
--definesENABLE_DASHBOARD_SUPPORT
   ENABLE_JAVASCRIPT_DEBUGGER ENABLE_LEGACY_CSS_VENDOR_PREFIXES
ENABLE_TEXT_NOTIFICATIONS_ONLY  ENABLE_WEB_AUDIO
   LANGUAGE_OBJECTIVE_C --generator ObjC --supplementalDependencyFile
./SupplementalDependencies.txt WebCore/html/ImageData.idl
In file included from built-in:151:
command line:1:9In file included from built-in:151:
#define -DENABLE_DASHBOARD_SUPPORT 1
^
: error: macro names must be identifiers

Any helpful ideas?
-Arthur

Here is the patch for the first two errors:

$ svn diff
Index: Source/WebCore/inspector/InspectorResourceAgent.cpp
===
--- Source/WebCore/inspector/InspectorResourceAgent.cpp (revision 121171)
+++ Source/WebCore/inspector/InspectorResourceAgent.cpp (working copy)
@@ -64,7 +64,9 @@
 #include WebSocketHandshakeResponse.h

 #include wtf/CurrentTime.h
+#if ENABLE(WEB_SOCKETS)
 #include wtf/HexNumber.h
+#endif
 #include wtf/ListHashSet.h
 #include wtf/RefPtr.h
 #include wtf/text/StringBuilder.h
Index: Source/WebCore/css/CSSPropertyNames.in
===
--- Source/WebCore/css/CSSPropertyNames.in  (revision 121171)
+++ Source/WebCore/css/CSSPropertyNames.in  (working copy)
@@ -332,7 +332,9 @@
 -webkit-min-logical-width
 -webkit-min-logical-height
 -webkit-nbsp-mode
+#if defined(ENABLE_CSS3_FLEXBOX)  ENABLE_CSS3_FLEXBOX
 -webkit-order
+#endif
 -webkit-padding-after
 -webkit-padding-before
 -webkit-padding-end

$ svn info
Path: .
URL: http://svn.webkit.org/repository/webkit/trunk
Repository Root: http://svn.webkit.org/repository/webkit
Repository UUID: 268f45cc-cd09-0410-ab3c-d52691b4dbfc
Revision: 121171
Node Kind: directory
Schedule: normal
Last Changed Author: jon...@apple.com
Last Changed Rev: 121171
Last Changed Date: 2012-06-25 11:05:13 -0700 (Mon, 25 Jun 2012)
___
webkit-dev mailing list
webkit-dev@lists.webkit.org
http://lists.webkit.org/mailman/listinfo.cgi/webkit-dev


Re: [webkit-dev] [PATCH] Errors from build-webkit --debug --minimal

2012-06-25 Thread Pablo Flouret

On Mon, 25 Jun 2012 12:03:12 -0700, Arthur O'Dwyer art...@push.am wrote:


I'm trying to build WebKit by the simplest possible method. I checked
out the SVN tree and did cd WebKit ; Tools/Scripts/build-webkit
--debug --minimal. This produced compiler errors!


I tried the same thing a while ago, and submitted patches to fix some of  
the issues, but what i was told is that there aren't really any bots  
building with --minimal, so it's not guaranteed to build.


Last i remember, after you fix those issues you mention you're probably  
gonna come up against trickier ones in --web-audio, --geolocation,  
--netscape-plugin-api and --fullscreen-api (although i haven't checked  
recently).


You can still run --minimal and enable the features that are giving you  
trouble (e.g. build-webkit --debug --minimal --javascript-debugger  
--geolocation ... etc.)


Cheers,

--
pablo flouret
motorola | webkit / browser team
___
webkit-dev mailing list
webkit-dev@lists.webkit.org
http://lists.webkit.org/mailman/listinfo.cgi/webkit-dev


Re: [webkit-dev] [PATCH] Errors from build-webkit --debug --minimal

2012-06-25 Thread Eric Seidel
If someone wanted to set up a --minimal bot, I'm sure that would be
welcome.  We used to have a --qt --minimal bot at some point.

http://trac.webkit.org/wiki/BuildBot

On Mon, Jun 25, 2012 at 12:28 PM, Pablo Flouret pab...@motorola.com wrote:
 On Mon, 25 Jun 2012 12:03:12 -0700, Arthur O'Dwyer art...@push.am wrote:

 I'm trying to build WebKit by the simplest possible method. I checked
 out the SVN tree and did cd WebKit ; Tools/Scripts/build-webkit
 --debug --minimal. This produced compiler errors!


 I tried the same thing a while ago, and submitted patches to fix some of the
 issues, but what i was told is that there aren't really any bots building
 with --minimal, so it's not guaranteed to build.

 Last i remember, after you fix those issues you mention you're probably
 gonna come up against trickier ones in --web-audio, --geolocation,
 --netscape-plugin-api and --fullscreen-api (although i haven't checked
 recently).

 You can still run --minimal and enable the features that are giving you
 trouble (e.g. build-webkit --debug --minimal --javascript-debugger
 --geolocation ... etc.)

 Cheers,

 --
 pablo flouret
 motorola | webkit / browser team

 ___
 webkit-dev mailing list
 webkit-dev@lists.webkit.org
 http://lists.webkit.org/mailman/listinfo.cgi/webkit-dev
___
webkit-dev mailing list
webkit-dev@lists.webkit.org
http://lists.webkit.org/mailman/listinfo.cgi/webkit-dev


Re: [webkit-dev] [PATCH] Errors from build-webkit --debug --minimal

2012-06-25 Thread Eric Seidel
It appears we still have a Qt-Minimal builder, I'm just not sure how
minimal it is:
http://build.webkit.org/builders/Qt%20Linux%20Release%20minimal

On Mon, Jun 25, 2012 at 12:56 PM, Eric Seidel e...@webkit.org wrote:
 If someone wanted to set up a --minimal bot, I'm sure that would be
 welcome.  We used to have a --qt --minimal bot at some point.

 http://trac.webkit.org/wiki/BuildBot

 On Mon, Jun 25, 2012 at 12:28 PM, Pablo Flouret pab...@motorola.com wrote:
 On Mon, 25 Jun 2012 12:03:12 -0700, Arthur O'Dwyer art...@push.am wrote:

 I'm trying to build WebKit by the simplest possible method. I checked
 out the SVN tree and did cd WebKit ; Tools/Scripts/build-webkit
 --debug --minimal. This produced compiler errors!


 I tried the same thing a while ago, and submitted patches to fix some of the
 issues, but what i was told is that there aren't really any bots building
 with --minimal, so it's not guaranteed to build.

 Last i remember, after you fix those issues you mention you're probably
 gonna come up against trickier ones in --web-audio, --geolocation,
 --netscape-plugin-api and --fullscreen-api (although i haven't checked
 recently).

 You can still run --minimal and enable the features that are giving you
 trouble (e.g. build-webkit --debug --minimal --javascript-debugger
 --geolocation ... etc.)

 Cheers,

 --
 pablo flouret
 motorola | webkit / browser team

 ___
 webkit-dev mailing list
 webkit-dev@lists.webkit.org
 http://lists.webkit.org/mailman/listinfo.cgi/webkit-dev
___
webkit-dev mailing list
webkit-dev@lists.webkit.org
http://lists.webkit.org/mailman/listinfo.cgi/webkit-dev


Re: [webkit-dev] [PATCH] Errors from build-webkit --debug --minimal

2012-06-25 Thread Patrick Gansterer
IMHO only setting up a bot won't help that much. Someone must keep it green. 
From my experience with the WinCE build slave (which has some common features 
turned of by default) I don't see many people caring about it.
E.g. WinCE build slaves builds with JIT turned of, which leads to much redness 
of the bot. IMHO that's ok, since the bots exist to find failures with other 
configurations (even if EWS does much work in that area already), but most of 
the redness of the WinCE bot stays until I have the time to fix it [1]. BTW: 
It's not that much fun to fix all failures (and find the corresponding 
revision) if you pause for a few weeks.
Do we have any defined responsibility, when someone breaks a port? Should the 
committer watch landing of the patch? (and ping the port maintainers if it has 
some port specific issues) Can we make the green/red status more meaningful, to 
make it easier to see if a patch cases real problems? (IMHO many bots are red 
by default this time and so don't provide quick information in the buildbot 
overview pages)

-- Patrick

[1] http://trac.webkit.org/search?q=paroga+ENABLE%28JIT%29

Am 25.06.2012 um 21:56 schrieb Eric Seidel:

 If someone wanted to set up a --minimal bot, I'm sure that would be
 welcome.  We used to have a --qt --minimal bot at some point.
 
 http://trac.webkit.org/wiki/BuildBot
 
 On Mon, Jun 25, 2012 at 12:28 PM, Pablo Flouret pab...@motorola.com wrote:
 On Mon, 25 Jun 2012 12:03:12 -0700, Arthur O'Dwyer art...@push.am wrote:
 
 I'm trying to build WebKit by the simplest possible method. I checked
 out the SVN tree and did cd WebKit ; Tools/Scripts/build-webkit
 --debug --minimal. This produced compiler errors!
 
 
 I tried the same thing a while ago, and submitted patches to fix some of the
 issues, but what i was told is that there aren't really any bots building
 with --minimal, so it's not guaranteed to build.
 
 Last i remember, after you fix those issues you mention you're probably
 gonna come up against trickier ones in --web-audio, --geolocation,
 --netscape-plugin-api and --fullscreen-api (although i haven't checked
 recently).
 
 You can still run --minimal and enable the features that are giving you
 trouble (e.g. build-webkit --debug --minimal --javascript-debugger
 --geolocation ... etc.)
 
 Cheers,
 
 --
 pablo flouret
 motorola | webkit / browser team
 
 ___
 webkit-dev mailing list
 webkit-dev@lists.webkit.org
 http://lists.webkit.org/mailman/listinfo.cgi/webkit-dev
 ___
 webkit-dev mailing list
 webkit-dev@lists.webkit.org
 http://lists.webkit.org/mailman/listinfo.cgi/webkit-dev

___
webkit-dev mailing list
webkit-dev@lists.webkit.org
http://lists.webkit.org/mailman/listinfo.cgi/webkit-dev


Re: [webkit-dev] [PATCH] Errors from build-webkit --debug --minimal

2012-06-25 Thread Adam Barth
Hopefully this email won't come off as rude, but it's not clear to me
whether the WinCE port being actively developed.
http://trac.webkit.org/wiki/WinCE says that the port is maintained
by Torch Mobile, but http://www.torchmobile.com/ says Torch Mobile
is focused on BlackBerry these days.

Looking at the last year of commits to these directories:

http://trac.webkit.org/log/trunk/Source/WebKit/wince
http://trac.webkit.org/log/trunk/Source/WebCore/platform/wince

The changes appear to be build fixes and patches that incidentally
touch WinCE as part of project-wide maintenance.

Patrick, it's clear that you're putting in the effort to keep the port
building, but I suspect the reason folks from the community don't put
much effort into keeping the WinCE bot green is because the port seems
to be somewhat in a holding pattern.

Adam


On Mon, Jun 25, 2012 at 1:45 PM, Patrick Gansterer par...@paroga.com wrote:
 IMHO only setting up a bot won't help that much. Someone must keep it green. 
 From my experience with the WinCE build slave (which has some common features 
 turned of by default) I don't see many people caring about it.
 E.g. WinCE build slaves builds with JIT turned of, which leads to much 
 redness of the bot. IMHO that's ok, since the bots exist to find failures 
 with other configurations (even if EWS does much work in that area already), 
 but most of the redness of the WinCE bot stays until I have the time to fix 
 it [1]. BTW: It's not that much fun to fix all failures (and find the 
 corresponding revision) if you pause for a few weeks.
 Do we have any defined responsibility, when someone breaks a port? Should the 
 committer watch landing of the patch? (and ping the port maintainers if it 
 has some port specific issues) Can we make the green/red status more 
 meaningful, to make it easier to see if a patch cases real problems? (IMHO 
 many bots are red by default this time and so don't provide quick information 
 in the buildbot overview pages)

 -- Patrick

 [1] http://trac.webkit.org/search?q=paroga+ENABLE%28JIT%29

 Am 25.06.2012 um 21:56 schrieb Eric Seidel:

 If someone wanted to set up a --minimal bot, I'm sure that would be
 welcome.  We used to have a --qt --minimal bot at some point.

 http://trac.webkit.org/wiki/BuildBot

 On Mon, Jun 25, 2012 at 12:28 PM, Pablo Flouret pab...@motorola.com wrote:
 On Mon, 25 Jun 2012 12:03:12 -0700, Arthur O'Dwyer art...@push.am wrote:

 I'm trying to build WebKit by the simplest possible method. I checked
 out the SVN tree and did cd WebKit ; Tools/Scripts/build-webkit
 --debug --minimal. This produced compiler errors!


 I tried the same thing a while ago, and submitted patches to fix some of the
 issues, but what i was told is that there aren't really any bots building
 with --minimal, so it's not guaranteed to build.

 Last i remember, after you fix those issues you mention you're probably
 gonna come up against trickier ones in --web-audio, --geolocation,
 --netscape-plugin-api and --fullscreen-api (although i haven't checked
 recently).

 You can still run --minimal and enable the features that are giving you
 trouble (e.g. build-webkit --debug --minimal --javascript-debugger
 --geolocation ... etc.)

 Cheers,

 --
 pablo flouret
 motorola | webkit / browser team

 ___
 webkit-dev mailing list
 webkit-dev@lists.webkit.org
 http://lists.webkit.org/mailman/listinfo.cgi/webkit-dev
 ___
 webkit-dev mailing list
 webkit-dev@lists.webkit.org
 http://lists.webkit.org/mailman/listinfo.cgi/webkit-dev

 ___
 webkit-dev mailing list
 webkit-dev@lists.webkit.org
 http://lists.webkit.org/mailman/listinfo.cgi/webkit-dev
___
webkit-dev mailing list
webkit-dev@lists.webkit.org
http://lists.webkit.org/mailman/listinfo.cgi/webkit-dev


Re: [webkit-dev] [PATCH] Errors from build-webkit --debug --minimal

2012-06-25 Thread Patrick Gansterer
As already mentioned in my previous mail (maybe net clear enough): It's ok if 
someone breaks WinCE specific stuff, and it's nice that it's still accepted in 
trunk, but what I wanted to say is: WinCE bot has a different 
_feature_configuration_ (e.g. !ENABLE(JIT) isn't WinCE specific) and that's 
what breaks most of the time.

BTW: My current main feature target is to merge the WinCE stuff directly into 
the windows, so it's more in the win directory. ;-)

-- Patrick

Am 25.06.2012 um 23:06 schrieb Adam Barth:

 Hopefully this email won't come off as rude, but it's not clear to me
 whether the WinCE port being actively developed.
 http://trac.webkit.org/wiki/WinCE says that the port is maintained
 by Torch Mobile, but http://www.torchmobile.com/ says Torch Mobile
 is focused on BlackBerry these days.
 
 Looking at the last year of commits to these directories:
 
 http://trac.webkit.org/log/trunk/Source/WebKit/wince
 http://trac.webkit.org/log/trunk/Source/WebCore/platform/wince
 
 The changes appear to be build fixes and patches that incidentally
 touch WinCE as part of project-wide maintenance.
 
 Patrick, it's clear that you're putting in the effort to keep the port
 building, but I suspect the reason folks from the community don't put
 much effort into keeping the WinCE bot green is because the port seems
 to be somewhat in a holding pattern.
 
 Adam
 
 
 On Mon, Jun 25, 2012 at 1:45 PM, Patrick Gansterer par...@paroga.com wrote:
 IMHO only setting up a bot won't help that much. Someone must keep it green. 
 From my experience with the WinCE build slave (which has some common 
 features turned of by default) I don't see many people caring about it.
 E.g. WinCE build slaves builds with JIT turned of, which leads to much 
 redness of the bot. IMHO that's ok, since the bots exist to find failures 
 with other configurations (even if EWS does much work in that area already), 
 but most of the redness of the WinCE bot stays until I have the time to fix 
 it [1]. BTW: It's not that much fun to fix all failures (and find the 
 corresponding revision) if you pause for a few weeks.
 Do we have any defined responsibility, when someone breaks a port? Should 
 the committer watch landing of the patch? (and ping the port maintainers if 
 it has some port specific issues) Can we make the green/red status more 
 meaningful, to make it easier to see if a patch cases real problems? (IMHO 
 many bots are red by default this time and so don't provide quick 
 information in the buildbot overview pages)
 
 -- Patrick
 
 [1] http://trac.webkit.org/search?q=paroga+ENABLE%28JIT%29
 
 Am 25.06.2012 um 21:56 schrieb Eric Seidel:
 
 If someone wanted to set up a --minimal bot, I'm sure that would be
 welcome.  We used to have a --qt --minimal bot at some point.
 
 http://trac.webkit.org/wiki/BuildBot
 
 On Mon, Jun 25, 2012 at 12:28 PM, Pablo Flouret pab...@motorola.com wrote:
 On Mon, 25 Jun 2012 12:03:12 -0700, Arthur O'Dwyer art...@push.am wrote:
 
 I'm trying to build WebKit by the simplest possible method. I checked
 out the SVN tree and did cd WebKit ; Tools/Scripts/build-webkit
 --debug --minimal. This produced compiler errors!
 
 
 I tried the same thing a while ago, and submitted patches to fix some of 
 the
 issues, but what i was told is that there aren't really any bots building
 with --minimal, so it's not guaranteed to build.
 
 Last i remember, after you fix those issues you mention you're probably
 gonna come up against trickier ones in --web-audio, --geolocation,
 --netscape-plugin-api and --fullscreen-api (although i haven't checked
 recently).
 
 You can still run --minimal and enable the features that are giving you
 trouble (e.g. build-webkit --debug --minimal --javascript-debugger
 --geolocation ... etc.)
 
 Cheers,
 
 --
 pablo flouret
 motorola | webkit / browser team
 
 ___
 webkit-dev mailing list
 webkit-dev@lists.webkit.org
 http://lists.webkit.org/mailman/listinfo.cgi/webkit-dev
 ___
 webkit-dev mailing list
 webkit-dev@lists.webkit.org
 http://lists.webkit.org/mailman/listinfo.cgi/webkit-dev
 
 ___
 webkit-dev mailing list
 webkit-dev@lists.webkit.org
 http://lists.webkit.org/mailman/listinfo.cgi/webkit-dev

___
webkit-dev mailing list
webkit-dev@lists.webkit.org
http://lists.webkit.org/mailman/listinfo.cgi/webkit-dev