The v2.066.0-b4 binaries are now available. The review period for beta 4 will run until 0700 UTC (0000 PDT, 0300 EDT, 1600 JST) on 21 July 2014, at which time binaries for B5 will be produced and released. Due diligence in identifying regressions as early as possible is requested and appreciated. Issue 13137, [1], is provided for identifying any fixed regressions that needs to be picked and included in B5.

Outstanding regressions impeeding release are:

	13062 	"member x is not accessible" when passing field to template parameter
	13077 	[dmd 2.066-b2] std.range.array with shared InputRangeObject
	13084 	ModuleInfo.opApply delegate expects immutable parameter
	13117 	Executable size of hello world explodes from 472K to 2.7M
	13127 	DCannot deduce function with int[][] argument and "in" parameter
	11946 	"need 'this' to access member" when passing field to template parameter
	12242 	conflict error with public imports
	12453 	'ini' directory missing in ZIP release bundles
	13098 	std.path functions no longer works with DirEntry
	6329 	Out of range exceptions not thrown in certain cases

Binaries are located here:

	ALL
	downloads.dlang.org/pre-releases/dmd.2.066.0-b4.zip

	FREEBSD
	downloads.dlang.org/pre-releases/dmd.2.066.0-b4.freebsd-32.zip
	downloads.dlang.org/pre-releases/dmd.2.066.0-b4.freebsd-64.zip

	LINUX
	downloads.dlang.org/pre-releases/dmd.2.066.0-b4.linux.zip
	downloads.dlang.org/pre-releases/dmd_2.066.0~b4-0_i386.deb
	downloads.dlang.org/pre-releases/dmd_2.066.0~b4-0_amd64.deb
	downloads.dlang.org/pre-releases/libphobos2-66_2.066.0~b4-0_i386.deb
	downloads.dlang.org/pre-releases/libphobos2-66_2.066.0~b4-0_amd64.deb
	downloads.dlang.org/pre-releases/dmd-2.066.0~b4-0.fedora.i386.rpm
	downloads.dlang.org/pre-releases/dmd-2.066.0~b4-0.fedora.x86_64.rpm
	downloads.dlang.org/pre-releases/dmd-2.066.0~b4-0.openSUSE.i386.rpm
	downloads.dlang.org/pre-releases/dmd-2.066.0~b4-0.openSUSE.x86_64.rpm

	OSX
	downloads.dlang.org/pre-releases/dmd.2.066.0-b4.osx.zip
	downloads.dlang.org/pre-releases/dmd.2.066.0-b4.dmg

	WINDOWS
	downloads.dlang.org/pre-releases/2014/dmd.2.066.0-b4.windows.zip
	downloads.dlang.org/pre-releases/2014/dmd-2.066.0-b4.exe


Following is a list of regression fixes incorporated since the release of v2.066.0-b1:

DMD
0e90572 - fix Issue 13102 - Cannot parse 184467440737095516153.6L #3753
1efe1ad - fix Issue 13114 - old opCmp requirement for AA keys should be detected for classes #3757
ef56c27 - fix Issue 13132 - ICE on interface AA key #3762
d182caf - fix Issue 12989 - Wrong x86_64 code for delegate return when compiled as lib #3739
fe0a0dd - fix Issue 12859 - Read-modify-write operation for shared variable in Phobos #2281
1cccc74 - fix Issue 12485 - [REG2.065] DMD crashes when recursive template expansion #3683
994f4a4 - fix Issue 13027 - Assertion `ex->op == TOKblit || ex->op == TOKconstruct' failed. #3716
0eaf764 - fix Issue 13026 - object.get cannot be called with [] as "defaultValue" argument #3717
bcf55ca - fix Issue 12896 - ld.gold complains about bad relocations when building libphobos2.so #3715
5248be0 - fix Issue 12255 - opCmp requirement for AAs breaks code #3711
01e467d - fix Issue 13053 - Wrong warning on implicitly generated __xtoHash #3719
10f3bf1 - fix Issue 13021 - Constructing union with floating type and then accessing its field in one expression causes ICE #3723
86b9076 - fix Issue 13071 - [ICE] dmd 2.066.0-b1 assertion in nogc.c:73 #3730
e85556a - fix Issue 13081 - ICE with alias this and opSlice #3732
49f0db5 - Fix warning in cast.c #3736
49f0db5 - fix Issue 13088 - Compiler segfaults with trivial case code #3736
49f0db5 - fix Issue 13087 - Error: no property 'xyz' for type 'Vec!4' #3736
e0ca2d2 - fix Issue 13024 - [ICE](expression.c line 1172) with implicit supertype conversion of different enums in array literal #3733

DRUNTIME
c679110 - fix Issue 13111: GC.realloc returns invalid memory for large reallocation #884
7474e95 - fix Issue 13025 - Tools repository does not build on Ubuntu #889
7d2a2ef - fix Issue 12958 - core.checkedint.mulu is broken. #890
9bd5777 - fix Issue 13034 - [Reg] core.stdc.stdio - deprecation warning with dmd -inline
a13628a - fix Issue 13078 - [dmd 2.066-b2] AA rehash failed with shared #877

POUBOS
ced5598 - fix Issue 12455 [reg]Bad lowercase mapping for 'LATIN CAPITAL LETTER I WITH DOT ABOVE'
f70443e - fix Issue 13076 - [dmd 2.066-b2] DList clearing of empty list #2315
592eb10 - fix Issue 13056 - [2.066.0-b1] Regression: Error: template std.path.baseName cannot deduce function from argument types !()(DirEntry) #2306

INSTALLER
198f80d - fix Issue 13047 - cannot stat `./icons/16/dmd-source.png'

[1] https://issues.dlang.org/show_bug.cgi?id=13137
[2] https://issues.dlang.org/buglist.cgi?bug_severity=regression&resolution=---

On Tuesday, 15 July 2014 at 22:40:02 UTC, Andrew Edwards wrote:
>
> Outstanding regressions impeding release are:
>

It would be nice if back-end  issues got some attention prior to releases.
For example:
https://issues.dlang.org/show_bug.cgi?id=9465
https://issues.dlang.org/show_bug.cgi?id=11435
https://issues.dlang.org/show_bug.cgi?id=12164
https://issues.dlang.org/show_bug.cgi?id=12503
https://issues.dlang.org/show_bug.cgi?id=12833
https://issues.dlang.org/show_bug.cgi?id=13023

Code gen bugs can cause "dynamic regressions" in user code if a change causes a different compiler path to be used.

I have a kludge / patch for #11435, but I get the impression of apathy towards back end issues so I don't feel motivated to contribute.

bug tracker is just a thing to collecting dust. you can write your report there, or to /dev/null, or not write it at all -- the result will be the same.

i know at least 3 bugs in phobos and at least one very nasty bug in compiler (which causes UB, so-called heisenbug), but have no motivation to report. did i mention that i have fixes too?

but it's ok, spice must flow, new releases must be done.

On 7/15/14, 9:43 PM, ketmar via Digitalmars-d-announce wrote:
> bug tracker is just a thing to collecting dust. you can write your
> report there, or to /dev/null, or not write it at all -- the result
> will be the same.

You mean http://issues.dlang.org? That's used regularly.

> i know at least 3 bugs in phobos and at least one very nasty bug in
> compiler (which causes UB, so-called heisenbug), but have no motivation
> to report. did i mention that i have fixes too?

Whatcha waiting for? reports + pull requests please. Thanks! -- Andrei

"safety0ff"  wrote in message news:xfceasqsqxxygwzscuxy@forum.dlang.org... 

> I have a kludge / patch for #11435, but I get the impression of apathy towards back end issues so I don't feel motivated to contribute.

Are you joking?  Do a pull request.

On Wednesday, 16 July 2014 at 04:56:14 UTC, ketmar via Digitalmars-d-announce wrote:
> bug tracker is just a thing to collecting dust. you can write your
> report there, or to /dev/null, or not write it at all -- the result
> will be the same.
>
> i know at least 3 bugs in phobos and at least one very nasty bug in
> compiler (which causes UB, so-called heisenbug), but have no motivation
> to report. did i mention that i have fixes too?
>
> but it's ok, spice must flow, new releases must be done.


You spent the effort implementing a fix, the time talking about your fix but cannot be buggered submitting a PR for the fix?

D has a problem ATM of too many talkers, not enough doers. Please submit a PR if you have a fix for anything in the bug tracker.

bye,
uri

On Wed, 16 Jul 2014 06:39:56 +0000
uri via Digitalmars-d-announce <digitalmars-d-announce@puremagic.com>
wrote:

> You spent the effort implementing a fix, the time talking about your fix but cannot be buggered submitting a PR for the fix?
one fix at a time, not more. the first one is still sitting in bugtracker, collecting dust. maybe after it will be reviewed i'll find some motivation to fill another ticket... in a month... or two... or six...

anyway, sorry for the noise. i'm really annoyed and just wanted to tell the world about it. sorry again.

On Tue, 15 Jul 2014 23:03:38 -0700
Andrei Alexandrescu via Digitalmars-d-announce
<digitalmars-d-announce@puremagic.com> wrote:

> You mean http://issues.dlang.org? That's used regularly.
oh, really? https://issues.dlang.org/show_bug.cgi?id=12853

> Whatcha waiting for? reports + pull requests please. Thanks! -- Andrei
have no motivation. and no, i'm neither using github, nor planning to use it in the current millenia. that's why i'm saying that issues.dlang.org is just a dust collector: no guthub pull request? your bug will sit in issues.dlang.org for indefinite time.

so what is the sense in issues.dlang.org existence? it should be closed to stop frustrating people.

On Wednesday, 16 July 2014 at 06:56:24 UTC, ketmar via Digitalmars-d-announce wrote:
> On Tue, 15 Jul 2014 23:03:38 -0700
> Andrei Alexandrescu via Digitalmars-d-announce
> <digitalmars-d-announce@puremagic.com> wrote:
>
>> You mean http://issues.dlang.org? That's used regularly.
> oh, really? https://issues.dlang.org/show_bug.cgi?id=12853

Yes. Issues can sit for years before getting fixed, it is normal workflow for situation when there are much more issue reporters than developers.

> so what is the sense in issues.dlang.org existence? it should be closed
> to stop frustrating people.

close all bug trackers -> no more bugs -> declare perfect D stability
sounds like a plan :D

On 16/07/14 08:49, ketmar via Digitalmars-d-announce wrote:
> On Wed, 16 Jul 2014 06:39:56 +0000
> uri via Digitalmars-d-announce <digitalmars-d-announce@puremagic.com>
> wrote:
>
>> You spent the effort implementing a fix, the time talking about
>> your fix but cannot be buggered submitting a PR for the fix?
> one fix at a time, not more. the first one is still sitting in
> bugtracker, collecting dust. maybe after it will be reviewed i'll find
> some motivation to fill another ticket... in a month... or two... or
> six...
>
> anyway, sorry for the noise. i'm really annoyed and just wanted to tell
> the world about it. sorry again.
>

Then paste the fix here and allow somebody else to do the boring PR stuff for you.

Just saying "I've got a fix but not sharing" sounds childish.