On 2013-08-30 10:26, Ramon wrote:

> My point of view, however, was that of a newbie. As such I won't
> make my learning curve steeper by learning two "standard
> libraries", which one to use when, etc.

Oh, and BTW, if you want to code on DStep you need to learn libclang even more than you need to learn Tango.

-- 
/Jacob Carlborg

On 2013-08-30 09:54, Artur Skawina wrote:

> and it seems to be alive:
>
> https://bitbucket.org/larsivi/amber/commits

That is not Tango. That is a completely separate language. Although is very similar to D1 and they do use Tango as their runtime and standard library.

Tango for D2 is located here:

https://github.com/SiegeLord/Tango-D2

-- 
/Jacob Carlborg

On Thursday, 29 August 2013 at 19:18:48 UTC, H. S. Teoh wrote:
>
> Right now, having no way to actually update that site to add a
> notice to this effect

On this point, when's the last time someone tried pinging him via email?  Is the whois for the domain not current?

-Wyatt

On 08/30/13 13:45, Jacob Carlborg wrote:
> On 2013-08-30 09:54, Artur Skawina wrote:
> 
>> and it seems to be alive:
>>
>> https://bitbucket.org/larsivi/amber/commits
> 
> That is not Tango. That is a completely separate language. Although is very similar to D1 and they do use Tango as their runtime and standard library.

The question was not about Tango.

artur

On Friday, 30 August 2013 at 11:42:00 UTC, Jacob Carlborg wrote:
>> - why tango? What's better there than in phobos?
>
> In my opinion:
>
> * Argument parsing
> * XML package
> * Net package (not any external dependencies)
> * Zip package
> * I think it has more collections
> * Logging
>
>> - what's the offering? (Is it a complete replacement for phobos?)
>
> Yes, but you can use just what you want. As I said, I'm only using the argument parsing from Tango.

Thanks for the information and your thoughts. I'd like to ask one
more question on that, as I happened to work a little with
phobos' getopt and looked at your code and happened to notice
(your use of) tangos arg parsing/handling:
Why do you consider tangos arg parsing being better? Could you
kindly elaborate that somewhat? (Because I, possibly due to my
newbness, failed to see striking advantages in tangos arg parsing
when looking at Dstep).


On Friday, 30 August 2013 at 11:43:15 UTC, Jacob Carlborg wrote:
> Oh, and BTW, if you want to code on DStep you need to learn libclang even more than you need to learn Tango.

libclang is a very interesting matter, anyway, and indeed one of
the more reasonable ways to attack the C->D problem. On a
sidenote, I also keep a very interested eye on the D frontend for
llvm.

Please, be assured that I have no interest whatsoever to dig in
old wounds or to talk against tango which sure enough is a major
piece of work.

Thanks & A+ -R

On 2013-08-30 20:19, Ramon wrote:

> Thanks for the information and your thoughts. I'd like to ask one
> more question on that, as I happened to work a little with
> phobos' getopt and looked at your code and happened to notice
> (your use of) tangos arg parsing/handling:
> Why do you consider tangos arg parsing being better? Could you
> kindly elaborate that somewhat? (Because I, possibly due to my
> newbness, failed to see striking advantages in tangos arg parsing
> when looking at Dstep).

I think the Tango version is more flexible and has more features.

With the Tango argument parser you can:

* Restrict values
* Specify the min and max number of arguments a given flag need to have
* Aliases
* Default value
* Conflict handling

The Phobos version accepts weird syntax's like:

--t 4

-- 
/Jacob Carlborg

On 2013-08-30 18:55, Artur Skawina wrote:

> The question was not about Tango.

Oh, my bad.

-- 
/Jacob Carlborg

On 8/31/13 4:21 AM, Jacob Carlborg wrote:
> On 2013-08-30 20:19, Ramon wrote:
>
>> Thanks for the information and your thoughts. I'd like to ask one
>> more question on that, as I happened to work a little with
>> phobos' getopt and looked at your code and happened to notice
>> (your use of) tangos arg parsing/handling:
>> Why do you consider tangos arg parsing being better? Could you
>> kindly elaborate that somewhat? (Because I, possibly due to my
>> newbness, failed to see striking advantages in tangos arg parsing
>> when looking at Dstep).
>
> I think the Tango version is more flexible and has more features.
>
> With the Tango argument parser you can:
>
> * Restrict values
> * Specify the min and max number of arguments a given flag need to have
> * Aliases
> * Default value
> * Conflict handling

Haven't seen Tango's arguments parser, but it's a given getopt can be improved in any number of ways. Yet the way I see it, with command line parsing, the margin between a good enough argument parser and a terrific one is razor thin. One parses arguments by definition once in every program, and things like checking against limits and constraints across multiple arguments can be easily done after basic parsing.

> The Phobos version accepts weird syntax's like:
>
> --t 4

Only if you tell it to!


Andrei

On 2013-08-31 19:11, Andrei Alexandrescu wrote:

> Haven't seen Tango's arguments parser, but it's a given getopt can be
> improved in any number of ways. Yet the way I see it, with command line
> parsing, the margin between a good enough argument parser and a terrific
> one is razor thin. One parses arguments by definition once in every
> program, and things like checking against limits and constraints across
> multiple arguments can be easily done after basic parsing.

I don't want this to start a huge argument about Tango versus Phobos. But Tango supports this _now_, out of the box, Phobos doesn't. No need to do any post processing. It's that simple. Again I don't understand why some people having so much trouble that some developers here are using Tango.

-- 
/Jacob Carlborg

On 8/31/13 11:12 AM, Jacob Carlborg wrote:
> On 2013-08-31 19:11, Andrei Alexandrescu wrote:
>
>> Haven't seen Tango's arguments parser, but it's a given getopt can be
>> improved in any number of ways. Yet the way I see it, with command line
>> parsing, the margin between a good enough argument parser and a terrific
>> one is razor thin. One parses arguments by definition once in every
>> program, and things like checking against limits and constraints across
>> multiple arguments can be easily done after basic parsing.
>
> I don't want this to start a huge argument about Tango versus Phobos.
> But Tango supports this _now_, out of the box, Phobos doesn't. No need
> to do any post processing. It's that simple.

I agree with the sentiment, but not with the example. It's just petty. There's no out-of-the-box vs it-could-be-built, it's just a difference without a distinction. We're talking about a few lines per application.

> Again I don't understand why some people having so much trouble that
> some developers here are using Tango.

In this case you're seeing things where they aren't. Speaking only for myself I think Tango is a fine library and I'm glad it's keeping traction in D.


Andrei