Another big pile of bug fixes. More contributors than ever!

http://www.digitalmars.com/d/1.0/changelog.html
http://ftp.digitalmars.com/dmd.1.075.zip

http://www.digitalmars.com/d/2.0/changelog.html
https://github.com/downloads/D-Programming-Language/dmd/dmd.2.060.zip

Walter Bright:
> Another big pile of bug fixes. More contributors than ever!

And there is the first step of this change too:
http://d.puremagic.com/issues/show_bug.cgi?id=6652

Bye,
bearophile

Nice update, but broke Derelict2 :-(

Regression: delegates with default arguments are broken (worked in 2.059)


void foo(void delegate(int x = 0) fun)
{
	fun(); // Error: expected 1 function arguments, not 0
}

On 8/2/2012 12:40 PM, Peter Alexander wrote:
> Nice update, but broke Derelict2 :-(
>
> Regression: delegates with default arguments are broken (worked in 2.059)
>
>
> void foo(void delegate(int x = 0) fun)
> {
>      fun(); // Error: expected 1 function arguments, not 0
> }

Known issue, it's an inevitable result (it never worked right anyway):

http://d.puremagic.com/issues/show_bug.cgi?id=8454

P.S. You might want to monitor the beta releases.

On 8/2/2012 12:42 PM, Walter Bright wrote:
> http://d.puremagic.com/issues/show_bug.cgi?id=8454

corrected url:

http://d.puremagic.com/issues/show_bug.cgi?id=3866

On 02-08-2012 21:40, Peter Alexander wrote:
> Nice update, but broke Derelict2 :-(
>
> Regression: delegates with default arguments are broken (worked in 2.059)
>
>
> void foo(void delegate(int x = 0) fun)
> {
>      fun(); // Error: expected 1 function arguments, not 0
> }

I think it was decided that this was not a regression, if memory serves. I believe the reasoning was that default parameters on delegates just don't make sense in the general case, so it wasn't worth fixing.

I could be wrong. I can't seem to find the relevant bug. Anyone?

-- 
Alex Rønne Petersen
alex@lycus.org
http://lycus.org

On 02-08-2012 21:44, Alex Rønne Petersen wrote:
> On 02-08-2012 21:40, Peter Alexander wrote:
>> Nice update, but broke Derelict2 :-(
>>
>> Regression: delegates with default arguments are broken (worked in 2.059)
>>
>>
>> void foo(void delegate(int x = 0) fun)
>> {
>>      fun(); // Error: expected 1 function arguments, not 0
>> }
>
> I think it was decided that this was not a regression, if memory serves.
> I believe the reasoning was that default parameters on delegates just
> don't make sense in the general case, so it wasn't worth fixing.
>
> I could be wrong. I can't seem to find the relevant bug. Anyone?
>

Oh, Walter already replied with the relevant URL. :)

-- 
Alex Rønne Petersen
alex@lycus.org
http://lycus.org

On 02-08-2012 21:18, Walter Bright wrote:
> Another big pile of bug fixes. More contributors than ever!
>
> http://www.digitalmars.com/d/1.0/changelog.html
> http://ftp.digitalmars.com/dmd.1.075.zip
>
> http://www.digitalmars.com/d/2.0/changelog.html
> https://github.com/downloads/D-Programming-Language/dmd/dmd.2.060.zip

Changelog is not updated?

-- 
Alex Rønne Petersen
alex@lycus.org
http://lycus.org

On Thursday, 2 August 2012 at 19:19:04 UTC, Walter Bright wrote:
> Another big pile of bug fixes. More contributors than ever!
>
> http://www.digitalmars.com/d/1.0/changelog.html
> http://ftp.digitalmars.com/dmd.1.075.zip
>
> http://www.digitalmars.com/d/2.0/changelog.html
> https://github.com/downloads/D-Programming-Language/dmd/dmd.2.060.zip

The 1.0 changelog link correctly displays version 1.075, but the 2.0 changelog still displays version 2.059

Paul

On 8/2/12, Walter Bright <newshound2@digitalmars.com> wrote:
> Known issue, it's an inevitable result (it never worked right anyway):
>
> http://d.puremagic.com/issues/show_bug.cgi?id=8454
>
> P.S. You might want to monitor the beta releases.
>

I've posted about that exact Derelict case in Issue 3866 during the betas, but nobody replied.

But a library solution could be made if one wants to use an inner alias. Maybe we even have one in Phobos? Something like:

void foo(void delegate(string s, bool isTrue) dg)
{
    alias DefVal!(dg, true) deg;
    // use deg as if it were dg() with default for 'isTrue'
}

Not the most readable code.. but it's similar to curry (except
backwards) I guess.