In one of the recent threads, there was a mention of blockers.
One of these, I think, is the absence of any kind of a roadmap.
No great detail is required, just a crude travel plan along the lines of "we'll go to Switzerland, but not until after we've visited Portugal".

The most uncertain feature is also the most important:
When will implicit function template instantiation be introduced?

For example, it doesn't make sense to develop standard container classes without an answer to this question. Will it begin to happen in, say, the next ten releases? Or will the path to 1.0 be mainly bug fixes and minor tweaking?

It seems that most other major confirmed features have been categorised as "2.0" or "eventually, but not yet". But the timescale for IFTI is still very unclear.

"Don Clugston" <dac@nospam.com.au> wrote in message news:ds9qaa$1of9$1@digitaldaemon.com...
> In one of the recent threads, there was a mention of blockers.
> One of these, I think, is the absence of any kind of a roadmap.
> No great detail is required, just a crude travel plan along the lines of
> "we'll go to Switzerland, but not until after we've visited Portugal".
>
> The most uncertain feature is also the most important:
> When will implicit function template instantiation be introduced?
>
> For example, it doesn't make sense to develop standard container classes without an answer to this question. Will it begin to happen in, say, the next ten releases? Or will the path to 1.0 be mainly bug fixes and minor tweaking?
>
> It seems that most other major confirmed features have been categorised as "2.0" or "eventually, but not yet". But the timescale for IFTI is still very unclear.

The problem is these "must have" features keep coming up. I thought I'd have implicit function template instantiation by now, but haven't even started on it. I'd spent a lot of time working on the other template problems you uncovered. I'd say hold off on designing a template library without implicit function template instantiation.

Walter Bright wrote:
> "Don Clugston" <dac@nospam.com.au> wrote in message news:ds9qaa$1of9$1@digitaldaemon.com...
> 
>>In one of the recent threads, there was a mention of blockers.
>>One of these, I think, is the absence of any kind of a roadmap.
>>No great detail is required, just a crude travel plan along the lines of "we'll go to Switzerland, but not until after we've visited Portugal".
>>
>>The most uncertain feature is also the most important:
>>When will implicit function template instantiation be introduced?
>>
>>For example, it doesn't make sense to develop standard container classes without an answer to this question. Will it begin to happen in, say, the next ten releases? Or will the path to 1.0 be mainly bug fixes and minor tweaking?
>>
>>It seems that most other major confirmed features have been categorised as "2.0" or "eventually, but not yet". But the timescale for IFTI is still very unclear.
> 
> 
> The problem is these "must have" features keep coming up. I thought I'd have implicit function template instantiation by now, but haven't even started on it. I'd spent a lot of time working on the other template problems you uncovered. I'd say hold off on designing a template library without implicit function template instantiation. 

Fantastic! In that case, I withdraw all other proposals I've made regarding templates. The meta-programming template support seems to be really quite solid now, from my original list of workarounds, the only one that remains is
digitalmars.com digitalmars.D.bugs:6004
which I think you just missed, because you fixed the almost identical one digitalmars.com digitalmars.D.bugs:6003.

Now, almost everything I try works as I'd hoped (and I've tried some pretty wild stuff!). It tends to be unforgiving, in that some kinds of syntax errors segfault the compiler instead of giving error messages; but that does not prevent library development. It's possible to write really elegant metafunctions now.

I'm not sure how I managed to get your attention, but it's been an exciting thing to be involved with. Some of those quirks that were fixed seemed to be fundamental issues with scoping, I'm amazed that they were solved so quickly. Apparently the rumours that you never sleep are true <g>.

> I'd say hold off on designing a template library without implicit function template instantiation.

That scentence is a little tricky for me, does that mean IFTI is expected for 1.0 ?

Charlie

"Walter Bright" <newshound@digitalmars.com> wrote in message news:ds9tpg$1r87$1@digitaldaemon.com...
>
> "Don Clugston" <dac@nospam.com.au> wrote in message news:ds9qaa$1of9$1@digitaldaemon.com...
> > In one of the recent threads, there was a mention of blockers.
> > One of these, I think, is the absence of any kind of a roadmap.
> > No great detail is required, just a crude travel plan along the lines of
> > "we'll go to Switzerland, but not until after we've visited Portugal".
> >
> > The most uncertain feature is also the most important:
> > When will implicit function template instantiation be introduced?
> >
> > For example, it doesn't make sense to develop standard container classes without an answer to this question. Will it begin to happen in, say, the next ten releases? Or will the path to 1.0 be mainly bug fixes and minor tweaking?
> >
> > It seems that most other major confirmed features have been categorised
as
> > "2.0" or "eventually, but not yet". But the timescale for IFTI is still very unclear.
>
> The problem is these "must have" features keep coming up. I thought I'd
have
> implicit function template instantiation by now, but haven't even started
on
> it. I'd spent a lot of time working on the other template problems you uncovered. I'd say hold off on designing a template library without
implicit
> function template instantiation.
>
>

Walter Bright wrote:
> 
> The problem is these "must have" features keep coming up. I thought I'd have implicit function template instantiation by now, but haven't even started on it. I'd spent a lot of time working on the other template problems you uncovered. I'd say hold off on designing a template library without implicit function template instantiation.

Fantastic!  Speculation on this issue was killing me :-)


Sean

"Don Clugston" <dac@nospam.com.au> wrote in message news:dsa0p5$1tl7$1@digitaldaemon.com...
> Fantastic! In that case, I withdraw all other proposals I've made
> regarding templates. The meta-programming template support seems to be
> really quite solid now, from my original list of workarounds, the only one
> that remains is
> digitalmars.com digitalmars.D.bugs:6004
> which I think you just missed, because you fixed the almost identical one
> digitalmars.com digitalmars.D.bugs:6003.

The one about (a[x]) template arguments? That one turns out to be not so easy to solve, which is why I deferred it. The workaround (putting redundant parentheses around it) is ugly, but not onerous.

> Now, almost everything I try works as I'd hoped (and I've tried some pretty wild stuff!). It tends to be unforgiving, in that some kinds of syntax errors segfault the compiler instead of giving error messages; but that does not prevent library development. It's possible to write really elegant metafunctions now.

I still want any bug reports that produce seg faults.

> I'm not sure how I managed to get your attention, but it's been an exciting thing to be involved with.

It's because you and Eric offered a kick-ass response to those C++ people who denigrate D's templates as "not capable of metaprogramming." You and Eric have shown that D templates enable much more powerful uses than C++ ones do, and look a lot better to boot. The regex example blew my socks off, and I wanted to showcase it. That meant fixing D's templates so regex didn't need all those ugly workarounds.

> Some of those quirks that were fixed seemed to be fundamental issues with scoping, I'm amazed that they were solved so quickly. Apparently the rumours that you never sleep are true <g>.

I originally thought the scope thing wasn't solvable. But the answer one day just popped into my head, I have no idea where it came from <g>.

"Charles" <noone@nowhere.com> wrote in message news:dsag9h$2cm8$1@digitaldaemon.com...
>> I'd say hold off on designing a template library without implicit function template instantiation.
>
> That scentence is a little tricky for me, does that mean IFTI is expected for 1.0 ?

Yes.

Walter Bright wrote:
> "Charles" <noone@nowhere.com> wrote in message news:dsag9h$2cm8$1@digitaldaemon.com...
> 
>>>I'd say hold off on designing a template library without implicit
>>>function template instantiation.
>>
>>That scentence is a little tricky for me, does that mean IFTI is expected
>>for 1.0 ?
> 
> 
> Yes. 
> 

Hooray!

>> I'm not sure how I managed to get your attention, but it's been an exciting thing to be involved with.
>
>It's because you and Eric offered a kick-ass response to those C++ people who denigrate D's templates as "not capable of metaprogramming." You and Eric have shown that D templates enable much more powerful uses than C++ ones do, and look a lot better to boot. The regex example blew my socks off, and I wanted to showcase it. That meant fixing D's templates so regex didn't need all those ugly workarounds.

This might be a bit OT, but where may I find those examples?

--Sebastián.

"Sebastián E. Peyrott" <as7cf@yahoo.com> wrote in message news:dsb154$2ua8$1@digitaldaemon.com...
>>It's because you and Eric offered a kick-ass response to those C++ people
>>who denigrate D's templates as "not capable of metaprogramming." You and
>>Eric have shown that D templates enable much more powerful uses than C++
>>ones do, and look a lot better to boot. The regex example blew my socks
>>off,
>>and I wanted to showcase it. That meant fixing D's templates so regex
>>didn't
>>need all those ugly workarounds.
>
> This might be a bit OT, but where may I find those examples?

http://trac.dsource.org/projects/ddl/browser/trunk/meta/regex.d