guile-devel
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: AARRRRGGH! Die Libtool, die!


From: tomas
Subject: Re: AARRRRGGH! Die Libtool, die!
Date: Tue, 18 Feb 2003 12:24:31 +0100
User-agent: Mutt/1.3.28i

On Sat, Feb 15, 2003 at 04:56:12PM -0800, Tom Lord wrote:
> 

[perl or python instead of sh]

> Bah.
> 
> Both perl and python are not stable. [...]

Whereas I consider Perl or Python more appropriate than sh for such
things, this is a real problem. To watch this in action see what
happens to the Debian packaging system when Perl `grows' a bit ;-)

> The problems with the auto* family of tools include:
> 
>       1) reliance on separate distribution
[...]
Definitely. I'd dream of a very small embeddable interpreter
with stable semantics and a tight C interface, very portable,
uh, wait a minute.

>       2) wrong approach to application portability
[...]
Many things have changed from the beginnings of Auto*. That's true.

>       3) wrong approach to /bin/sh portability (m4)
[...]
I don't really think that /bin/sh is the right tool for this job.
Especially not the mythical monster called ``portable shell''.

>       4) wrong approach to makefile portability
[...]
I have more issues with /bin/sh than with M4, actually. OK, M4 may
be obscure, but it has at least semantics ;-)

>       5) wrong approach to makefile automation
[...]
If everyone had GNU Make... (but you're quite right: the only widespread
non-Unixoid platform out there probably will have GNU Make whenever it
has some form of /bin/sh: via Cygwin. Times have changed).

>       6) lack of consideration for package management
[...]
Uh, oh. Do you mean offering `hooks' to package management systems to
let them gather whatever bits of information they need? Or yet-another-
package-management-system? Tightly coupled to an application build
framework, at that.

>       7) excessive complexity
[...]
Absolutely. I think we agree that this stems mainly from 3), 4) and 5).

I'll definitely look into your project.

Thanks
-- tomas




reply via email to

[Prev in Thread] Current Thread [Next in Thread]