On 23 March 2018 at 02:37, Norm via Digitalmars-d <digitalmars-d@puremagic.com> wrote:
>
> Looking at bugzilla I see this is also now fixed but we were on 2.074 at the time. Sorry I don't have more specific details, it was hard enough just to get some devs to create bugzilla accounts let alone find existing tickets or raise new tickets.

I can echo this experience. I think only two colleagues (out of quite
a lot) of mine have ever created a bugzilla account.
Most of them get to the point where they see a website that looks like
it's from the 90's and it wants you to create
yet-another-internet-account™, they just close the page.
Nobody wants more internet accounts.

This happened to me again on Tuesday this week...

On 23 March 2018 at 03:16, Walter Bright via Digitalmars-d <digitalmars-d@puremagic.com> wrote:
> On 3/17/2018 9:25 PM, Manu wrote:
>>
>> What is so hard about implementing a pow intrinsic that CTFE can use?
>
>
> https://github.com/dlang/dmd/pull/8071

Best PR this year!

>> It's one of those blocker bugs that's been there almost 10 years.
>
>
> It's one that seems to engender lots of comments, but no action.
>
> (BTW, there's a way to do it without the CTFE fix. One can use the approach I've always used in the past for C, which is write a separate program to generate the tables. This was used in the DMD build, and was gradually replaced with CTFE. It still exists in the backend, which is still in C++.)

Right, but then there's no reason to use D. When D undermine's its own proposed usefulness, it loses against C++ every time, no competition.

In my experience, migration to D involves a series of case-studies, typically demonstrating points of weakness for C++ (ie, tables of data pasted in code), and how D seductively improves the situation. This is exactly one of those cases, except you're confronted with a very rude-awakening, and it's a powerful turn-off rather than a turn on.

On 3/23/2018 8:15 AM, Timon Gehr wrote:
> https://gcc.gnu.org/bugzilla/show_bug.cgi?id=51664

The money shot: https://gcc.gnu.org/bugzilla/show_bug.cgi?id=51664#c3

On 3/23/2018 11:14 AM, Manu wrote:
> I can echo this experience. I think only two colleagues (out of quite
> a lot) of mine have ever created a bugzilla account.
> Most of them get to the point where they see a website that looks like
> it's from the 90's and it wants you to create
> yet-another-internet-account™, they just close the page.
> Nobody wants more internet accounts.

I have no idea how they can use git, since that has a user interface from the 1970's :-) github itself may look modern, but it's a rube goldberg construction that is hardly user discoverable.


> This happened to me again on Tuesday this week...

All bugzilla requires is a name and a password. It does not do any verification. Heck, just type in xxx yyy and it'll work. This trivial bit of effort makes it effective in preventing troll posts :-)

On 23 March 2018 at 11:24, Walter Bright via Digitalmars-d <digitalmars-d@puremagic.com> wrote:
> On 3/23/2018 8:15 AM, Timon Gehr wrote:
>>
>> https://gcc.gnu.org/bugzilla/show_bug.cgi?id=51664
>
>
> The money shot: https://gcc.gnu.org/bugzilla/show_bug.cgi?id=51664#c3

Thing is, C++ has nothing to prove; it's already accepted as de-facto
standard. People have decades of familiarity and comfort with C++'s
edges. As I've said before, on first look at D, they need to see a
sparkling promised land; not a new set of edges they need to learn and
be aware of. People aren't interested in trading one edgy language for
another.
Trust me, these impressions are __super important__, and offering of
work-around's are unlikely to be considered satisfactory responses,
likewise pointing at things that suck about C++; it's not a defence,
they already know all too well!
When I do demo's, I tend to carefully guide the process such that we
avoid hitting a stream of edges I know about. People are already
sceptical and desperately looking for a reason to dismiss the whole
thing entirely. Most people seem to hate learning new stuff ;)
Key is to make a convincing sell consistently, and we're getting
dangerously close to that point I reckon.

Here's another one of these apparently trivial cases which is highly likely to emerge for a new user (ie, has, in my experience, and I have to 'explain' the situation, which is anti-merit): https://github.com/dlang/dmd/pull/8031

On 23 March 2018 at 12:02, Walter Bright via Digitalmars-d <digitalmars-d@puremagic.com> wrote:
> On 3/23/2018 11:14 AM, Manu wrote:
>>
>> I can echo this experience. I think only two colleagues (out of quite
>> a lot) of mine have ever created a bugzilla account.
>> Most of them get to the point where they see a website that looks like
>> it's from the 90's and it wants you to create
>> yet-another-internet-account™, they just close the page.
>> Nobody wants more internet accounts.
>
> I have no idea how they can use git, since that has a user interface from the 1970's :-) github itself may look modern, but it's a rube goldberg construction that is hardly user discoverable.

Oh, I've put a LOT of effort into trying to sell git (very
successfully) too!! ;)
Fortunately, git has loads of good clients now. Particularly on Windows.


>> This happened to me again on Tuesday this week...
>
> All bugzilla requires is a name and a password. It does not do any verification. Heck, just type in xxx yyy and it'll work. This trivial bit of effort makes it effective in preventing troll posts :-)

Well, my colleague isn't a troll. A genuinely interested party, but
he's not gonna go out of his way for it. I can't control the natural
reaction that most people have to being confronted with a registration
page.
I'd suggest openauth, and people using their github accounts; I think
that's what people expect. I mean, most people just expect the bug
tracker to BE on github ;)

On 3/23/2018 11:09 AM, Manu wrote:
> [...]

You make some good points. The CTFE one is kinda inexcusable on our part, because it was trivial to implement (just more or less add some table entries and some glue copying existing examples), and there were posts after posts here and on bugzilla talking about it and nobody doing anything about it.

Rvalue references are not trivial and can have major unintended consequences. They're a rather ugly feature in C++, with weirdities. I doubt D will ever have them.

Regarding ARC, we've tried on that front numerous times. Still we don't have a good RC solution, but it isn't for lack of trying.

But at some level, D cannot replace C++ on a line-by-line basis. There's always going to be something different. If not in the core language, in the way the standard library works. If you're heavily using templates and stuff in C++, you're likely going to have to rethink how the code works to get it in D (or any other language).

For example, in my efforts translating C to D, the clumsy part is the metaprogramming in the C preprocessor. There's nothing there D cannot do, but it has to be redesigned. The result is much better, but translating by rote is simply impossible.

Also, just try translating some of the code in Phobos to C++. It was tried to do ranges for C++, and the result was terrifying. (It worked, but that's about all that could be said for it.)

On Friday, 23 March 2018 at 18:09:01 UTC, Manu wrote:
> [snip]
>
> Like, in this particular project, being able to generate all tables at compile time is the thing that distinguishes the D code from the C++ code; it's the *whole point*... If I have to continue to generate tables offline and paste big tables of data in the source code (and then re-generate them manually when I change something); then situation is identical to C++, therefore, stick with C++.\
>

You can use import expressions, but then you have to parse the string at compile-time to turn it into something useful, I suppose.

On Friday, March 23, 2018 12:13:58 Manu via Digitalmars-d wrote:
> On 23 March 2018 at 12:02, Walter Bright via Digitalmars-d
>
> <digitalmars-d@puremagic.com> wrote:
> > On 3/23/2018 11:14 AM, Manu wrote:
> >> This happened to me again on Tuesday this week...
> >
> > All bugzilla requires is a name and a password. It does not do any verification. Heck, just type in xxx yyy and it'll work. This trivial bit of effort makes it effective in preventing troll posts :-)
>
> Well, my colleague isn't a troll. A genuinely interested party, but
> he's not gonna go out of his way for it. I can't control the natural
> reaction that most people have to being confronted with a registration
> page.
> I'd suggest openauth, and people using their github accounts; I think
> that's what people expect. I mean, most people just expect the bug
> tracker to BE on github ;)

Really? I've dealt with relatively few projects that use github as a bug tracker, and it's been my experience that most anything that's really serious has its own bugtracker (usually some form of bugzilla) - though most such projects predate github by a long shot. I'd think that signing up for a bugtracker would be par for the course and that if anything, the fact that a project was using github issues instead of its own bugtracker would imply that it was small, which doesn't necessarily give a good impression - especially for a compiler.

And with how simplistic github issues are in comparison to bugzilla, I don't know why you'd want to use it other than the fact that you don't have to go to the effort of setting up your own bugzilla. I'd certainly hate to see us switch to github issues just because a few folks weren't willing to sign up for a bugzilla account, though for whatever reason, some folks keep pushing for us to switch over.

- Jonathan M Davis

On 23 March 2018 at 12:25, Jonathan M Davis via Digitalmars-d <digitalmars-d@puremagic.com> wrote:
> On Friday, March 23, 2018 12:13:58 Manu via Digitalmars-d wrote:
>> On 23 March 2018 at 12:02, Walter Bright via Digitalmars-d
>>
>> <digitalmars-d@puremagic.com> wrote:
>> > On 3/23/2018 11:14 AM, Manu wrote:
>> >> This happened to me again on Tuesday this week...
>> >
>> > All bugzilla requires is a name and a password. It does not do any verification. Heck, just type in xxx yyy and it'll work. This trivial bit of effort makes it effective in preventing troll posts :-)
>>
>> Well, my colleague isn't a troll. A genuinely interested party, but
>> he's not gonna go out of his way for it. I can't control the natural
>> reaction that most people have to being confronted with a registration
>> page.
>> I'd suggest openauth, and people using their github accounts; I think
>> that's what people expect. I mean, most people just expect the bug
>> tracker to BE on github ;)
>
> Really? I've dealt with relatively few projects that use github as a bug tracker, and it's been my experience that most anything that's really serious has its own bugtracker (usually some form of bugzilla) - though most such projects predate github by a long shot. I'd think that signing up for a bugtracker would be par for the course and that if anything, the fact that a project was using github issues instead of its own bugtracker would imply that it was small, which doesn't necessarily give a good impression - especially for a compiler.
>
> And with how simplistic github issues are in comparison to bugzilla, I don't know why you'd want to use it other than the fact that you don't have to go to the effort of setting up your own bugzilla. I'd certainly hate to see us switch to github issues just because a few folks weren't willing to sign up for a bugzilla account, though for whatever reason, some folks keep pushing for us to switch over.

I'm not suggesting switch to github. I've never suggested that. I
understand it's inferior.
I'm suggesting supporting openauth.