Branch: refs/heads/master
  Home:   https://github.com/D-Programming-Language/druntime

  Commit: 850eaa6e06fc323ee8bd964f9c15be7ae1b02699
      https://github.com/D-Programming-Language/druntime/commit/850eaa6e06fc323ee8bd964f9c15be7ae1b02699
  Author: Daniel Murphy <yebblies at gmail.com>
  Date:   2011-07-12 (Tue, 12 Jul 2011)

  Changed paths:
    M src/rt/adi.d

  Log Message:
  -----------
  Do not sort arrays of immutable elements, use a cast instead to prevent a performance penalty.


  Commit: 54c5c78a60f1e6ee00fb53019eb3deaaed65ae39
      https://github.com/D-Programming-Language/druntime/commit/54c5c78a60f1e6ee00fb53019eb3deaaed65ae39
  Author: Steven Schveighoffer <schveiguy at yahoo.com>
  Date:   2011-07-12 (Tue, 12 Jul 2011)

  Changed paths:
    M src/rt/adi.d

  Log Message:
  -----------
  Merge pull request #37 from yebblies/issue1339

Do not sort arrays of immutable elements


Compare: https://github.com/D-Programming-Language/druntime/compare/eb8246c...54c5c78

Yay.. first post-release checkin, and the build breakage is in full swing. :)

On Tue, 12 Jul 2011, noreply at github.com wrote:

> Date: Tue, 12 Jul 2011 06:24:47 -0700
> From: noreply at github.com
> Reply-To: D's runtime library developers list <d-runtime at puremagic.com>
> To: d-runtime at puremagic.com
> Subject: [D-runtime] [D-Programming-Language/druntime] 850eaa: Do not sort
>     arrays of immutable elements, use a ca...
> 
>   Branch: refs/heads/master
>   Home:   https://github.com/D-Programming-Language/druntime
> 
>   Commit: 850eaa6e06fc323ee8bd964f9c15be7ae1b02699
>       https://github.com/D-Programming-Language/druntime/commit/850eaa6e06fc323ee8bd964f9c15be7ae1b02699
>   Author: Daniel Murphy <yebblies at gmail.com>
>   Date:   2011-07-12 (Tue, 12 Jul 2011)
> 
>   Changed paths:
>     M src/rt/adi.d
> 
>   Log Message:
>   -----------
>   Do not sort arrays of immutable elements, use a cast instead to prevent a performance penalty.
> 
> 
>   Commit: 54c5c78a60f1e6ee00fb53019eb3deaaed65ae39
>       https://github.com/D-Programming-Language/druntime/commit/54c5c78a60f1e6ee00fb53019eb3deaaed65ae39
>   Author: Steven Schveighoffer <schveiguy at yahoo.com>
>   Date:   2011-07-12 (Tue, 12 Jul 2011)
> 
>   Changed paths:
>     M src/rt/adi.d
> 
>   Log Message:
>   -----------
>   Merge pull request #37 from yebblies/issue1339
> 
> Do not sort arrays of immutable elements
> 
> 
> Compare: https://github.com/D-Programming-Language/druntime/compare/eb8246c...54c5c78
> _______________________________________________
> D-runtime mailing list
> D-runtime at puremagic.com
> http://lists.puremagic.com/mailman/listinfo/d-runtime
> 

On 2011-07-12 10:00, Brad Roberts wrote:
> Yay.. first post-release checkin, and the build breakage is in full swing.
> 
> :)

*Sigh* And it was so nice to see it all green. I was even considering commenting on how wonderful it was that the autotester was all green for once. We really should do a better job of avoiding breakages...

- Jonathan M Davis

bleh, that's junk.? You can't do a const-cast on tail-const stuff?

How is that supposed to work, do this?

cast()[]?

that doesn't work either.? Sorry Daniel, I guess I was wrong.? Or is it the compiler which is wrong?? It seems I should be able to do a const cast to remove const, seeing as how I can do:

struct S
{
?? dchar[] d;

}

const(S) s;

S s2 = cast()s;

I guess the correct thing to do now is switch it back to the way you were going to do it, then file a bug on how crappy cast() is.

For those of you wondering about the context, the fix was to cast a dstring to dchar[]:

dchar[] d = cast(dchar[])toUTF32(s);


so it could be sorted.? I suggested using cast(), since this doesn't risk changing the array type, just the constancy.? But apparently cast() does nothing to tail-const or tail-immutable arrays.

-Steve




>________________________________
>From: Brad Roberts <braddr at puremagic.com>
>To: D's runtime library developers list <d-runtime at puremagic.com>
>Sent: Tuesday, July 12, 2011 1:00 PM
>Subject: Re: [D-runtime] [D-Programming-Language/druntime] 850eaa: Do not sort arrays of immutable elements, use a ca...
>
>Yay.. first post-release checkin, and the build breakage is in full swing. :)
>
>On Tue, 12 Jul 2011, noreply at github.com wrote:
>
>> Date: Tue, 12 Jul 2011 06:24:47 -0700
>> From: noreply at github.com
>> Reply-To: D's runtime library developers list <d-runtime at puremagic.com>
>> To: d-runtime at puremagic.com
>> Subject: [D-runtime] [D-Programming-Language/druntime] 850eaa: Do not sort
>>? ?  arrays of immutable elements, use a ca...
>> 
>>?  Branch: refs/heads/master
>>?  Home:? https://github.com/D-Programming-Language/druntime
>> 
>>?  Commit: 850eaa6e06fc323ee8bd964f9c15be7ae1b02699
>>? ? ? https://github.com/D-Programming-Language/druntime/commit/850eaa6e06fc323ee8bd964f9c15be7ae1b02699
>>?  Author: Daniel Murphy <yebblies at gmail.com>
>>?  Date:?  2011-07-12 (Tue, 12 Jul 2011)
>> 
>>?  Changed paths:
>>? ?  M src/rt/adi.d
>> 
>>?  Log Message:
>>?  -----------
>>?  Do not sort arrays of immutable elements, use a cast instead to prevent a performance penalty.
>> 
>> 
>>?  Commit: 54c5c78a60f1e6ee00fb53019eb3deaaed65ae39
>>? ? ? https://github.com/D-Programming-Language/druntime/commit/54c5c78a60f1e6ee00fb53019eb3deaaed65ae39
>>?  Author: Steven Schveighoffer <schveiguy at yahoo.com>
>>?  Date:?  2011-07-12 (Tue, 12 Jul 2011)
>> 
>>?  Changed paths:
>>? ?  M src/rt/adi.d
>> 
>>?  Log Message:
>>?  -----------
>>?  Merge pull request #37 from yebblies/issue1339
>> 
>> Do not sort arrays of immutable elements
>> 
>> 
>> Compare: https://github.com/D-Programming-Language/druntime/compare/eb8246c...54c5c78
>> _______________________________________________
>> D-runtime mailing list
>> D-runtime at puremagic.com
>> http://lists.puremagic.com/mailman/listinfo/d-runtime
>> 
>_______________________________________________
>D-runtime mailing list
>D-runtime at puremagic.com
>http://lists.puremagic.com/mailman/listinfo/d-runtime
>
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.puremagic.com/pipermail/d-runtime/attachments/20110712/7d3b35d4/attachment.html>

>From: Steve Schveighoffer <schveiguy at yahoo.com>
>I guess the correct thing to do now is switch it back to the way you were going to do it, then file a bug on how crappy cast() is.


http://d.puremagic.com/issues/show_bug.cgi?id=6297

On Tue, 12 Jul 2011, Steve Schveighoffer wrote:

> I guess the correct thing to do now is switch it back to the way you were going to do it, then file a bug on how crappy cast() is.

Yes, please revert if it can't be fixed right away.  I'm going to be louder this dev cycle about build and test breakages being reverted faster.  The tree being red hurts a lot (hides failures, keeps people from being able to sync without suffering failures, etc) and so please try to keep it green.

Thanks,
Brad
-------------- next part --------------
_______________________________________________
D-runtime mailing list
D-runtime at puremagic.com
http://lists.puremagic.com/mailman/listinfo/d-runtime

I'll check in a tested fix momentarily, since it was my fault.

-Steve


>________________________________
>From: Brad Roberts <braddr at puremagic.com>
>To: Steve Schveighoffer <schveiguy at yahoo.com>; D's runtime library developers list <d-runtime at puremagic.com>
>Sent: Tuesday, July 12, 2011 2:54 PM
>Subject: Re: [D-runtime] [D-Programming-Language/druntime] 850eaa: Do not sort arrays of immutable elements, use a ca...
>
>On Tue, 12 Jul 2011, Steve Schveighoffer wrote:
>
>> I guess the correct thing to do now is switch it back to the way you were going to do it, then file a bug on how crappy cast() is.
>
>Yes, please revert if it can't be fixed right away.? I'm going to be louder this dev cycle about build and test breakages being reverted faster.? The tree being red hurts a lot (hides failures, keeps people from being able to sync without suffering failures, etc) and so please try to keep it green.
>
>Thanks,
>Brad
>
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.puremagic.com/pipermail/d-runtime/attachments/20110712/8096affd/attachment.html>

Le 2011-07-12 ? 14:26, Jonathan M Davis a ?crit :

> On 2011-07-12 10:00, Brad Roberts wrote:
>> Yay.. first post-release checkin, and the build breakage is in full swing.
>> 
>> :)
> 
> *Sigh* And it was so nice to see it all green. I was even considering commenting on how wonderful it was that the autotester was all green for once. We really should do a better job of avoiding breakages...

I think it'd help a lot if it was possible to request a specific commit to be tested by the autotester. I develop on OS X and it's hard for me to check that something works on other platforms. In my latest DMD pull request (Type-checked arrays) I asked if other people could test on other platforms... which isn't really nice at the stage of a pull request. Being able to request a test run on this (and having the error log) would make me able to fix problems for all platforms by myself before even creating a pull request.

Perhaps we should have a web form to enqueue a test run like that:

User: michelf
DMD Commit: 000000
Druntime Commit: 000000
Phobos Commit: 000000

This would enqueue a request to fetch michelf's github repository for dmd, druntime, and phobos, checkout specific commit for each repo and then run the tests.

-- 
Michel Fortin
michel.fortin at michelf.com
http://michelf.com/

On Tue, 12 Jul 2011, Michel Fortin wrote:

> Le 2011-07-12 ? 14:26, Jonathan M Davis a ?crit :
> 
> > On 2011-07-12 10:00, Brad Roberts wrote:
> >> Yay.. first post-release checkin, and the build breakage is in full swing.
> >> 
> >> :)
> > 
> > *Sigh* And it was so nice to see it all green. I was even considering commenting on how wonderful it was that the autotester was all green for once. We really should do a better job of avoiding breakages...
> 
> I think it'd help a lot if it was possible to request a specific commit to be tested by the autotester. I develop on OS X and it's hard for me to check that something works on other platforms. In my latest DMD pull request (Type-checked arrays) I asked if other people could test on other platforms... which isn't really nice at the stage of a pull request. Being able to request a test run on this (and having the error log) would make me able to fix problems for all platforms by myself before even creating a pull request.
> 
> Perhaps we should have a web form to enqueue a test run like that:
> 
> User: michelf
> DMD Commit: 000000
> Druntime Commit: 000000
> Phobos Commit: 000000
> 
> This would enqueue a request to fetch michelf's github repository for dmd, druntime, and phobos, checkout specific commit for each repo and then run the tests.

There's value in it, yes.  There's also significant security risk in it that I'm not comfortable with nor eager to invest the time to reduce. Yes, I recognize that there's risk even with the current system, but less as there's far more eyeballs on the main source tree.  I'm ok with things breaking from time to time, it happens.  I'm mostly interested in reacting appropriately when it does.

If anyone else wants to volunteer their time and systems to running an adhoc tester like the above, please do so.  Keep in mind that the systems I'm running are my donated time and money (with a nod to Sean for the contribution of an account on an osx box).

Also, not that I'm picking on today's commit, since it's not the first time it's happened, but commits that fail on _every_ platform just shouldn't happen.

My 2 cents,
Brad

Oops, my bad.  Serves me right for deciding to skip running the tests 'just this once'.

On Wed, Jul 13, 2011 at 3:00 AM, Brad Roberts <braddr at puremagic.com> wrote:
> Yay.. first post-release checkin, and the build breakage is in full swing. :)
>
> On Tue, 12 Jul 2011, noreply at github.com wrote:
>
>> Date: Tue, 12 Jul 2011 06:24:47 -0700
>> From: noreply at github.com
>> Reply-To: D's runtime library developers list <d-runtime at puremagic.com>
>> To: d-runtime at puremagic.com
>> Subject: [D-runtime] [D-Programming-Language/druntime] 850eaa: Do not sort
>> ? ? arrays of immutable elements, use a ca...
>>
>> ? Branch: refs/heads/master
>> ? Home: ? https://github.com/D-Programming-Language/druntime
>>
>> ? Commit: 850eaa6e06fc323ee8bd964f9c15be7ae1b02699
>> ? ? ? https://github.com/D-Programming-Language/druntime/commit/850eaa6e06fc323ee8bd964f9c15be7ae1b02699
>> ? Author: Daniel Murphy <yebblies at gmail.com>
>> ? Date: ? 2011-07-12 (Tue, 12 Jul 2011)
>>
>> ? Changed paths:
>> ? ? M src/rt/adi.d
>>
>> ? Log Message:
>> ? -----------
>> ? Do not sort arrays of immutable elements, use a cast instead to prevent a performance penalty.
>>
>>
>> ? Commit: 54c5c78a60f1e6ee00fb53019eb3deaaed65ae39
>> ? ? ? https://github.com/D-Programming-Language/druntime/commit/54c5c78a60f1e6ee00fb53019eb3deaaed65ae39
>> ? Author: Steven Schveighoffer <schveiguy at yahoo.com>
>> ? Date: ? 2011-07-12 (Tue, 12 Jul 2011)
>>
>> ? Changed paths:
>> ? ? M src/rt/adi.d
>>
>> ? Log Message:
>> ? -----------
>> ? Merge pull request #37 from yebblies/issue1339
>>
>> Do not sort arrays of immutable elements
>>
>>
>> Compare: https://github.com/D-Programming-Language/druntime/compare/eb8246c...54c5c78
>> _______________________________________________
>> D-runtime mailing list
>> D-runtime at puremagic.com
>> http://lists.puremagic.com/mailman/listinfo/d-runtime
>>
> _______________________________________________
> D-runtime mailing list
> D-runtime at puremagic.com
> http://lists.puremagic.com/mailman/listinfo/d-runtime
>