update.sh

Nick Sabalausky a at a.a
Mon Jan 16 03:46:42 PST 2012


"Nick Sabalausky" <a at a.a> wrote in message 
news:jf0ois$1t01$1 at digitalmars.com...
> "Andrei Alexandrescu" <SeeWebsiteForEmail at erdani.org> wrote in message 
> news:jf0l94$1i6a$1 at digitalmars.com...
>> On 1/16/12 1:18 AM, Jacob Carlborg wrote:
>>> On 2012-01-16 06:20, Nick Sabalausky wrote:
>>>>
>>>> Soo...you made a tool to do what DVM already does? ;)
>>>
>>> And only works on Posix.
>>
>> It's a simple 200 lines script that does something that someone working 
>> on dmd etc. needs.
>
> Fair enough.
>
> And to be honest, DVM doesn't actually do the git checkout just yet 
> (that's planned). But it'll compile it once you have checked out from git 
> (Although I think the recent move of the DMD download page from 
> digitalmars.com to d-p-l.org broke part of it - we admittedly need to fix 
> that.)
>
>> I just looked over DVM (on https://bitbucket.org/doob/dvm/wiki/Home). 
>> It's a large program, has a large list of dependencies,
>
> It doesn't have any runtime dependencies.
>
> As far as compile-time dependencies, that's only relevent if you need to 
> compile it, which should be very rare since binaries are already provided 
> for Win, OSX, Lin32 and Lin64.
>
>> is largely undocumented (I'm unclear even to what extent it does what I 
>> need),
>
> Fair enough. That's something we could use improve.
>
>> it seems to do a ton of stuff I don't need,
>
> Uhh...so?
>
>> and there's no way it could be justified for inclusion in tools/.
>
> Are there any other reasons for this besides what you've outlined? (If so, 
> that would help us.)
>
>> It's great that DVM exists for people who need it, but I don't see a 
>> competition between it and update.sh.
>>
>
> I don't mean to say that your tool is bad or that you shouldn't have made 
> it, but the issues you addressed with DVM can be fixed, and really, I've 
> been detecting a bit of NIH coming from your general direction. I've 
> noticed that any time you want something, you go off and do it yourself, 
> and to hell with any part of the community that's been working on anything 
> similar. If you see a problem with the existing options, at the very least 
> you could say something instead of ignoring it and going cowboy-coder, 
> running off undermining anything that doesn't have the official 
> "Championed by Andrei" stamp.
>

I apologize for the tone, I overreacted. I'm sure you weren't trying to 
undermine anyone or anything, just posting a potentially helpful tool, like 
we all try to do. I do think it would be nice for you to give feedback if 
there's any D tools/libs that are similar to what you want but don't quite 
fit the bill for one reason or another. But I understand if it wasn't clear 
that DVM was trying to move in that direction.




More information about the Digitalmars-d-announce mailing list