SpiderMonkey 17 Standalone Release

classic Classic list List threaded Threaded
7 messages Options
Reply | Threaded
Open this post in threaded view
|

SpiderMonkey 17 Standalone Release

Sean Stangl
I am pleased to announce a release candidate for the standalone JS 17 source release, based off Mozilla ESR17:

http://people.mozilla.org/~sstangl/mozjs17.0.0.rc0.tar.gz

Release notes were also given a placeholder on MDN, currently in template form with no useful information:

https://developer.mozilla.org/en-US/docs/SpiderMonkey/17

For tracking, packaging work is happening in https://bugzilla.mozilla.org/show_bug.cgi?id=735599.

The plan is to wait roughly a week to gather feedback, with a final release planned for Thursday, March 7th if feedback is positive. Please give feedback either in reply to this post or in Bug 735599.

Sean
_______________________________________________
dev-tech-js-engine mailing list
[hidden email]
https://lists.mozilla.org/listinfo/dev-tech-js-engine
Reply | Threaded
Open this post in threaded view
|

Re: SpiderMonkey 17 Standalone Release

Jens Ayton-3
On Feb 27, 2013, at 23:45, Sean Stangl <[hidden email]> wrote:
>
> I am pleased to announce a release candidate for the standalone JS 17 source release, based off Mozilla ESR17:
...
> The plan is to wait roughly a week to gather feedback, with a final release planned for Thursday, March 7th if feedback is positive. Please give feedback either in reply to this post or in Bug 735599.

Under standard conditions on Mac OS X (Xcode 4.6 toolchain, Apple's clang build forked from 3.2, ./configure run without any tweaks) I get:

    In file included from .../mozjs17/js/src/jsalloc.cpp:9:
    In file included from ./jscntxt.h:17:
    ./jsapi.h:2585:1: error: 'JS_GetNaNValue' has C-linkage specified, but returns user-defined type 'jsval' (aka 'JS::Value') which is incompatible with C
          [-Werror,-Wreturn-type-c-linkage]
    JS_GetNaNValue(JSContext *cx);

This also applies to JS_GetNegativeInfinityValue(), JS_GetPositiveInfinityValue(), JS_GetEmptyStringValue(), JS_ComputeThis(), JS_GetReservedSlot() and JS_GetFrameReturnValue().

Suppressing this (-Wno-return-type-c-linkage) lets it build and apparently work.


I also get lots of these:
    In file included from .../mozjs17/js/src/jsalloc.cpp:9:
    ./jscntxt.h:1659:75: warning: deleted function definitions are a C++11 extension [-Wc++11-extensions]
        AutoReleaseNullablePtr operator=(const AutoReleaseNullablePtr &other) MOZ_DELETE;
                                                                              ^
    ./dist/include/mozilla/Attributes.h:196:35: note: expanded from macro 'MOZ_DELETE'
    #  define MOZ_DELETE            = delete

These can be suppressed by adding -Wno-c++11-extensions to OS_CXXFLAGS in autoconf.mk, or by adding -std=c++11 (or gnu++11) somewhere.


There area also lots of -Wcast-aligns ("cast from 'js::types::TypeObjectKey *' to 'js::types::TypeObjectKey **' increases required alignment from 1 to 8" etc).


--
Jens Ayton

_______________________________________________
dev-tech-js-engine mailing list
[hidden email]
https://lists.mozilla.org/listinfo/dev-tech-js-engine
Reply | Threaded
Open this post in threaded view
|

Re: SpiderMonkey 17 Standalone Release

Jeff Walden-2
In reply to this post by Sean Stangl
On 03/02/2013 04:51 AM, Jens Ayton wrote:> Under standard conditions on Mac OS X (Xcode 4.6 toolchain, Apple's clang build forked from 3.2, ./configure run without any tweaks) I get:

>
>     In file included from .../mozjs17/js/src/jsalloc.cpp:9:
>     In file included from ./jscntxt.h:17:
>     ./jsapi.h:2585:1: error: 'JS_GetNaNValue' has C-linkage specified, but returns user-defined type 'jsval' (aka 'JS::Value') which is incompatible with C
>           [-Werror,-Wreturn-type-c-linkage]
>     JS_GetNaNValue(JSContext *cx);
>
> This also applies to JS_GetNegativeInfinityValue(), JS_GetPositiveInfinityValue(), JS_GetEmptyStringValue(), JS_ComputeThis(), JS_GetReservedSlot() and JS_GetFrameReturnValue().
>
> Suppressing this (-Wno-return-type-c-linkage) lets it build and apparently work.

We got rid of the |extern "C"| in whatever form from JSAPI on trunk.  We could conceivably do it for ESR17 as well, but that's not a patch we could backport.  (Some misguided souls link against the JSAPI symbols in Firefox in extensions, and we wouldn't want to break them in an ESR update.)  We'd have to fork ESR17 (not desirable) or start a new head (also not desirable) or something else.  Maybe adding in that flag is the right thing to do, not sure.

> I also get lots of these:
>     In file included from .../mozjs17/js/src/jsalloc.cpp:9:
>     ./jscntxt.h:1659:75: warning: deleted function definitions are a C++11 extension [-Wc++11-extensions]
>         AutoReleaseNullablePtr operator=(const AutoReleaseNullablePtr &other) MOZ_DELETE;
>                                                                               ^
>     ./dist/include/mozilla/Attributes.h:196:35: note: expanded from macro 'MOZ_DELETE'
>     #  define MOZ_DELETE            = delete
>
> These can be suppressed by adding -Wno-c++11-extensions to OS_CXXFLAGS in autoconf.mk, or by adding -std=c++11 (or gnu++11) somewhere.

Note that probably we'd want c++0x in both cases as the "11" bit is fairly new in compilers, while "0x" has been around a bit longer.

We have at least the former in Mozilla somewhere.  Possibly that showed up after ESR17?  Not sure offhand.  The latter might happen soon (I hope it will), but I don't think we want it for ESR17, or could land it there.  (We tried -std=c++0x at one point, but it interfered with C compatibility.  Now that C compat is no longer a concern, I'm not aware of any impediments to doing so on trunk.)

Jeff
_______________________________________________
dev-tech-js-engine mailing list
[hidden email]
https://lists.mozilla.org/listinfo/dev-tech-js-engine
Reply | Threaded
Open this post in threaded view
|

Re: SpiderMonkey 17 Standalone Release

Sean Stangl
In reply to this post by Jens Ayton-3
RC1: http://people.mozilla.org/~sstangl/mozjs17.0.0.rc1.tar.gz

These warnings should now be fixed via clang-specific pragmas. A slew of other clang-only warnings were also fixed.

Thanks for testing!
Sean

----- Original Message -----
From: "Jens Ayton" <[hidden email]>
To: [hidden email]
Cc: "Sean Stangl" <[hidden email]>
Sent: Saturday, March 2, 2013 4:51:08 AM
Subject: Re: SpiderMonkey 17 Standalone Release

On Feb 27, 2013, at 23:45, Sean Stangl <[hidden email]> wrote:
>
> I am pleased to announce a release candidate for the standalone JS 17 source release, based off Mozilla ESR17:
...
> The plan is to wait roughly a week to gather feedback, with a final release planned for Thursday, March 7th if feedback is positive. Please give feedback either in reply to this post or in Bug 735599.

Under standard conditions on Mac OS X (Xcode 4.6 toolchain, Apple's clang build forked from 3.2, ./configure run without any tweaks) I get:

    In file included from .../mozjs17/js/src/jsalloc.cpp:9:
    In file included from ./jscntxt.h:17:
    ./jsapi.h:2585:1: error: 'JS_GetNaNValue' has C-linkage specified, but returns user-defined type 'jsval' (aka 'JS::Value') which is incompatible with C
          [-Werror,-Wreturn-type-c-linkage]
    JS_GetNaNValue(JSContext *cx);

This also applies to JS_GetNegativeInfinityValue(), JS_GetPositiveInfinityValue(), JS_GetEmptyStringValue(), JS_ComputeThis(), JS_GetReservedSlot() and JS_GetFrameReturnValue().

Suppressing this (-Wno-return-type-c-linkage) lets it build and apparently work.


I also get lots of these:
    In file included from .../mozjs17/js/src/jsalloc.cpp:9:
    ./jscntxt.h:1659:75: warning: deleted function definitions are a C++11 extension [-Wc++11-extensions]
        AutoReleaseNullablePtr operator=(const AutoReleaseNullablePtr &other) MOZ_DELETE;
                                                                              ^
    ./dist/include/mozilla/Attributes.h:196:35: note: expanded from macro 'MOZ_DELETE'
    #  define MOZ_DELETE            = delete

These can be suppressed by adding -Wno-c++11-extensions to OS_CXXFLAGS in autoconf.mk, or by adding -std=c++11 (or gnu++11) somewhere.


There area also lots of -Wcast-aligns ("cast from 'js::types::TypeObjectKey *' to 'js::types::TypeObjectKey **' increases required alignment from 1 to 8" etc).


--
Jens Ayton

_______________________________________________
dev-tech-js-engine mailing list
[hidden email]
https://lists.mozilla.org/listinfo/dev-tech-js-engine
Reply | Threaded
Open this post in threaded view
|

Re: SpiderMonkey 17 Standalone Release

Sean Stangl
RC2: http://people.mozilla.org/~sstangl/mozjs17.0.0.rc2.tar.gz

Fixes OpenBSD issues from Bug 735599 and a JSContext destruction issue from Bug 852789.

Robin -- could you verify that Bug 852789 resolved your issue?

Sean

----- Original Message -----
From: "Sean Stangl" <[hidden email]>
To: [hidden email]
Sent: Monday, March 4, 2013 5:29:34 PM
Subject: Re: SpiderMonkey 17 Standalone Release

RC1: http://people.mozilla.org/~sstangl/mozjs17.0.0.rc1.tar.gz

These warnings should now be fixed via clang-specific pragmas. A slew of other clang-only warnings were also fixed.

Thanks for testing!
Sean

----- Original Message -----
From: "Jens Ayton" <[hidden email]>
To: [hidden email]
Cc: "Sean Stangl" <[hidden email]>
Sent: Saturday, March 2, 2013 4:51:08 AM
Subject: Re: SpiderMonkey 17 Standalone Release

On Feb 27, 2013, at 23:45, Sean Stangl <[hidden email]> wrote:
>
> I am pleased to announce a release candidate for the standalone JS 17 source release, based off Mozilla ESR17:
...
> The plan is to wait roughly a week to gather feedback, with a final release planned for Thursday, March 7th if feedback is positive. Please give feedback either in reply to this post or in Bug 735599.

Under standard conditions on Mac OS X (Xcode 4.6 toolchain, Apple's clang build forked from 3.2, ./configure run without any tweaks) I get:

    In file included from .../mozjs17/js/src/jsalloc.cpp:9:
    In file included from ./jscntxt.h:17:
    ./jsapi.h:2585:1: error: 'JS_GetNaNValue' has C-linkage specified, but returns user-defined type 'jsval' (aka 'JS::Value') which is incompatible with C
          [-Werror,-Wreturn-type-c-linkage]
    JS_GetNaNValue(JSContext *cx);

This also applies to JS_GetNegativeInfinityValue(), JS_GetPositiveInfinityValue(), JS_GetEmptyStringValue(), JS_ComputeThis(), JS_GetReservedSlot() and JS_GetFrameReturnValue().

Suppressing this (-Wno-return-type-c-linkage) lets it build and apparently work.


I also get lots of these:
    In file included from .../mozjs17/js/src/jsalloc.cpp:9:
    ./jscntxt.h:1659:75: warning: deleted function definitions are a C++11 extension [-Wc++11-extensions]
        AutoReleaseNullablePtr operator=(const AutoReleaseNullablePtr &other) MOZ_DELETE;
                                                                              ^
    ./dist/include/mozilla/Attributes.h:196:35: note: expanded from macro 'MOZ_DELETE'
    #  define MOZ_DELETE            = delete

These can be suppressed by adding -Wno-c++11-extensions to OS_CXXFLAGS in autoconf.mk, or by adding -std=c++11 (or gnu++11) somewhere.


There area also lots of -Wcast-aligns ("cast from 'js::types::TypeObjectKey *' to 'js::types::TypeObjectKey **' increases required alignment from 1 to 8" etc).


--
Jens Ayton

_______________________________________________
dev-tech-js-engine mailing list
[hidden email]
https://lists.mozilla.org/listinfo/dev-tech-js-engine
_______________________________________________
dev-tech-js-engine mailing list
[hidden email]
https://lists.mozilla.org/listinfo/dev-tech-js-engine
Reply | Threaded
Open this post in threaded view
|

Re: SpiderMonkey 17 Standalone Release

Robin Ehrlich
In reply to this post by Sean Stangl
Yes. The change corrected the problem I was encountering.

Thanks,
Robin

"Sean Stangl"  wrote in message
news:[hidden email]...

RC2: http://people.mozilla.org/~sstangl/mozjs17.0.0.rc2.tar.gz

Fixes OpenBSD issues from Bug 735599 and a JSContext destruction issue from
Bug 852789.

Robin -- could you verify that Bug 852789 resolved your issue?

Sean

----- Original Message -----
From: "Sean Stangl" <[hidden email]>
To: [hidden email]
Sent: Monday, March 4, 2013 5:29:34 PM
Subject: Re: SpiderMonkey 17 Standalone Release

RC1: http://people.mozilla.org/~sstangl/mozjs17.0.0.rc1.tar.gz

These warnings should now be fixed via clang-specific pragmas. A slew of
other clang-only warnings were also fixed.

Thanks for testing!
Sean

----- Original Message -----
From: "Jens Ayton" <[hidden email]>
To: [hidden email]
Cc: "Sean Stangl" <[hidden email]>
Sent: Saturday, March 2, 2013 4:51:08 AM
Subject: Re: SpiderMonkey 17 Standalone Release

On Feb 27, 2013, at 23:45, Sean Stangl <[hidden email]> wrote:
>
> I am pleased to announce a release candidate for the standalone JS 17
> source release, based off Mozilla ESR17:
...
> The plan is to wait roughly a week to gather feedback, with a final
> release planned for Thursday, March 7th if feedback is positive. Please
> give feedback either in reply to this post or in Bug 735599.

Under standard conditions on Mac OS X (Xcode 4.6 toolchain, Apple's clang
build forked from 3.2, ./configure run without any tweaks) I get:

    In file included from .../mozjs17/js/src/jsalloc.cpp:9:
    In file included from ./jscntxt.h:17:
    ./jsapi.h:2585:1: error: 'JS_GetNaNValue' has C-linkage specified, but
returns user-defined type 'jsval' (aka 'JS::Value') which is incompatible
with C
          [-Werror,-Wreturn-type-c-linkage]
    JS_GetNaNValue(JSContext *cx);

This also applies to JS_GetNegativeInfinityValue(),
JS_GetPositiveInfinityValue(), JS_GetEmptyStringValue(), JS_ComputeThis(),
JS_GetReservedSlot() and JS_GetFrameReturnValue().

Suppressing this (-Wno-return-type-c-linkage) lets it build and apparently
work.


I also get lots of these:
    In file included from .../mozjs17/js/src/jsalloc.cpp:9:
    ./jscntxt.h:1659:75: warning: deleted function definitions are a C++11
extension [-Wc++11-extensions]
        AutoReleaseNullablePtr operator=(const AutoReleaseNullablePtr
&other) MOZ_DELETE;
                                                                             
  ^
    ./dist/include/mozilla/Attributes.h:196:35: note: expanded from macro
'MOZ_DELETE'
    #  define MOZ_DELETE            = delete

These can be suppressed by adding -Wno-c++11-extensions to OS_CXXFLAGS in
autoconf.mk, or by adding -std=c++11 (or gnu++11) somewhere.


There area also lots of -Wcast-aligns ("cast from 'js::types::TypeObjectKey
*' to 'js::types::TypeObjectKey **' increases required alignment from 1 to
8" etc).


--
Jens Ayton

_______________________________________________
dev-tech-js-engine mailing list
[hidden email]
https://lists.mozilla.org/listinfo/dev-tech-js-engine 

_______________________________________________
dev-tech-js-engine mailing list
[hidden email]
https://lists.mozilla.org/listinfo/dev-tech-js-engine
Reply | Threaded
Open this post in threaded view
|

Re: SpiderMonkey 17 Standalone Release

Sean Stangl
In reply to this post by Sean Stangl
I am pleased to announce the release of SpiderMonkey (MozJS) 17.0.0:

http://ftp.mozilla.org/pub/mozilla.org/js/mozjs17.0.0.tar.gz
MD5: 20b6f8f1140ef6e47daa3b16965c9202

Thanks to everyone involved in finally bringing this release to fruition, especially darkxst, glandium, Ted, Waldo, and dmandelin. Release notes exist at https://developer.mozilla.org/en-US/docs/SpiderMonkey/17 and will be updated in short order.

The next JS major release will be SpiderMonkey 24, around November of this year, tracking the Firefox ESR schedule.

Sean
_______________________________________________
dev-tech-js-engine mailing list
[hidden email]
https://lists.mozilla.org/listinfo/dev-tech-js-engine