Nick has submitted a new pull request that fixes the rdmd issues.  It should be merged, and a new build of rdmd should be included in the next release.  The rdmd that shipped with 2.054 does *not* work correctly.

https://github.com/D-Programming-Language/tools/pull/11

-Lars

Eh, I forgot to push that one. Fixed.

On 9/2/2011 2:02 AM, Daniel Murphy wrote:
> Introduced in https://github.com/D-Programming-Language/dmd/commit/1193f7828b444056c943742daae0a5ccf262272e ....
>
> On Fri, Sep 2, 2011 at 6:59 PM, Daniel Murphy<yebblies at gmail.com>  wrote:
>> https://github.com/D-Programming-Language/dmd/blob/master/src/mars.c#L178
>>
>> On Fri, Sep 2, 2011 at 6:44 PM, Walter Bright<walter at digitalmars.com>  wrote:
>>>
>>> On 9/2/2011 1:20 AM, Daniel Murphy wrote:
>>>> I think this is just the halt Walter left in verror...
>>> Vat ist you talking abvat?

I'm generally concerned with our ctor/postblit/dtor situation.
If it's not fixed this release, mere mortals won't see foolproof RAII
for another two months.
I'm speaking of a bunch of quite similar stuff:
http://d.puremagic.com/issues/show_bug.cgi?id=6199
http://d.puremagic.com/issues/show_bug.cgi?id=5737
http://d.puremagic.com/issues/show_bug.cgi?id=6499
http://d.puremagic.com/issues/show_bug.cgi?id=6581

They are likely the reason for RefCounted in Phobos malfunctioning: http://d.puremagic.com/issues/show_bug.cgi?id=6437

-- 
Dmitry Olshansky

On 9/2/11, Dmitry Olshansky <dmitry.olsh at gmail.com> wrote:
> I'm generally concerned with our ctor/postblit/dtor situation.

Also I've filed this recently: http://d.puremagic.com/issues/show_bug.cgi?id=6364

This can't be by design, right? I mean it looks like a pretty big deal to me..

Since this thread turned into a show me yours, here is mine:

http://d.puremagic.com/issues/show_bug.cgi?id=6290

while a D1 bug, it makes it impossible for threaded applications on MacOSX it is fixed by:

https://github.com/D-Programming-Language/phobos/pull/141

Tho it would be nice if somebody with druntime knowledge would take a look at it, before it is pulled.

Cheers Jakob.

Hi,

It would probably be a good idea to move std.compiler in Phobos to rt.compiler in druntime before doing a beta release, so it can actually be used (now that it has useful values/checks). If the beta is released before doing this, code that starts using std.compiler would break when it's moved to druntime later.

Regards,
Alex

Which use cases do you have in mind for it? The rt package is druntime-internal, just as gc.* is.

David


On 9/2/11 7:12 PM, Alex wrote:
> Hi,
>
> It would probably be a good idea to move std.compiler in Phobos to rt.compiler in druntime before doing a beta release, so it can actually be used (now that it has useful values/checks). If the beta is released before doing this, code that starts using std.compiler would break when it's moved to druntime later.
>
> Regards,
> Alex
> _______________________________________________
> dmd-beta mailing list
> dmd-beta at puremagic.com
> http://lists.puremagic.com/mailman/listinfo/dmd-beta

On Fri, 2 Sep 2011, Dmitry Olshansky wrote:

> I'm generally concerned with our ctor/postblit/dtor situation.
> If it's not fixed this release, mere mortals won't see foolproof RAII for
> another two months.
> I'm speaking of a bunch of quite similar stuff:
> http://d.puremagic.com/issues/show_bug.cgi?id=6199
> http://d.puremagic.com/issues/show_bug.cgi?id=5737
> http://d.puremagic.com/issues/show_bug.cgi?id=6499
> http://d.puremagic.com/issues/show_bug.cgi?id=6581
> 
> They are likely the reason for RefCounted in Phobos malfunctioning: http://d.puremagic.com/issues/show_bug.cgi?id=6437
> 
> -- 
> Dmitry Olshansky

While I agree that fixing more bugs would be good, I also know that there will always be more bugs to fix in some area that's critical for someone. There's a large number of fixes that are already committed and ready to be released and having those out in an official release is also of high value.  It's all a bunch of tradeoffs and I think the every 4ish week cycle works pretty well.

Also, I'm not trying to single out this particular reply or set of bugs, just needed a post to reply to.  The comment applies to several in this thread.  Bugs which don't alraedy have fixes in hand are less likely to make it into the next release unless they're regressions introduced in this release cycle or have a fix already written and tested.

Make sense?

Thanks,
Brad

On Fri, 02 Sep 2011 12:41:31 +0300, Walter Bright <walter at digitalmars.com> wrote:

> Eh, I forgot to push that one. Fixed.

By the way, you can set up git to refuse to commit lines containing specific keywords. That way, you can just add a //NOCOMMIT comment to any debugging code and not worry about accidentally forgetting to remove it later.

Here's the pre-commit hook for one of my projects (a tweaked version of a public example), which also does some whitespace checks: https://gist.github.com/1189548

-- 
Your local git fanboy,
  Vladimir                            mailto:vladimir at thecybershadow.net

On 02.09.2011 23:00, Brad Roberts wrote:
> On Fri, 2 Sep 2011, Dmitry Olshansky wrote:
>
>> I'm generally concerned with our ctor/postblit/dtor situation.
>> If it's not fixed this release, mere mortals won't see foolproof RAII for
>> another two months.
>> I'm speaking of a bunch of quite similar stuff:
>> http://d.puremagic.com/issues/show_bug.cgi?id=6199
>> http://d.puremagic.com/issues/show_bug.cgi?id=5737
>> http://d.puremagic.com/issues/show_bug.cgi?id=6499
>> http://d.puremagic.com/issues/show_bug.cgi?id=6581
>>
>> They are likely the reason for RefCounted in Phobos malfunctioning: http://d.puremagic.com/issues/show_bug.cgi?id=6437
>>
>> -- 
>> Dmitry Olshansky
> While I agree that fixing more bugs would be good, I also know that there will always be more bugs to fix in some area that's critical for someone. There's a large number of fixes that are already committed and ready to be released and having those out in an official release is also of high value.  It's all a bunch of tradeoffs and I think the every 4ish week cycle works pretty well.

This is reasonable, I'm not pushing anyone to roll out a brand new fix
and get it into release.
(sorry if it did sound like that)
I just wanted to check if fixes available in this area (namely struct
ctor/dtor ) are getting merged before release. BTW at least one of them
was recently merged.

> Also, I'm not trying to single out this particular reply or set of bugs, just needed a post to reply to.  The comment applies to several in this thread.  Bugs which don't alraedy have fixes in hand are less likely to make it into the next release unless they're regressions introduced in this release cycle or have a fix already written and tested.
>
> Make sense?
Crystal clear.
Speaking to DMD compiler team (or band ;) ) in general:
I love the remarkable progress I see everyday, keep up the good work!

-- 
Dmitry Olshansky