chicken-hackers
[Top][All Lists]
Advanced

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

Re: [Chicken-hackers] [PATCH] platform auto-detection mechanism


From: Mario Domenech Goulart
Subject: Re: [Chicken-hackers] [PATCH] platform auto-detection mechanism
Date: Fri, 01 Feb 2013 04:43:40 -0500
User-agent: Gnus/5.13 (Gnus v5.13) Emacs/24.2 (gnu/linux)

Hi Michele,

On Fri, 1 Feb 2013 01:23:22 +0100 Michele La Monaca <address@hidden> wrote:

>>> Maybe there are good reasons I overlooked not to use an auto-detection
>>> mechanism (maybe it's just not really feasible) but in case, here it
>>> is my naive attempt.  I wouldn't bet on it but I would give it a try.
>>
>> The idea is interesting, indeed.  What about cross-builds?
>
> The autodetection mechanism is "skipped" if you explicitly set the
> PLATFORM on the command line or in config.make. I hardly see any
> problem in this regard.  But if you do, let me know, I've never tried
> to cross-compile chicken and I have absolutely no intention to do so
> (life is too short!).  So I can't tell for sure.
>
> As a general remark, if I may be permitted this comment,

Sure!  Your opinion is very important.


> I would never ever sacrifice a feature for the sake of
> cross-compilation, I have seen other projects too encumbered by
> cross-compilation to have the time and freedom to deliver anything
> useful. But that's your project, of course.

Cross-compilation is critical, at least to me, since I need it to use
Chicken at work.  Anyway, I think cross-compilation support is an
important feature for any project.

You probably are thinking that Chicken is too conservative at accepting
patches for the core.  It may be true, depending on your point of view,
but there's a good reason for that: we have quite a few developers with
very limited time using only a small subset of the considerably large
number of supported platforms (hardware and software).  When something
breaks, there's only a small group of people to expect fixes from.

Some parts of the core are specially critical, and the build system is
one of them.  If something breaks on a platform which is not easily
available to developers, it can turn into a release blocker, for
example.

But, please, don't let all that put you down at contributing.
Contributions are welcome, specially in the form of patches, like you
are doing.  They are really appreciated, and we thank you.  But, for the
sake of not breaking what is not broken, we have to be picky with regard
to the critical core parts.


>>> My main concerns are about those mingw things... I'am not exactly an
>>> expert in the field.
>>
>> Yeah, we have only a few people testing Chicken on Windows, and it it
>> causes breakages when we least expect them.  Autodetection on Windows
>> sounds particularly hard.
>
> Same comment here. Autodetection is optional, if Windows and its
> Unix-like offprings are too tough to tame, well, let's continue to set
> PLATFORM on those.  Autodection would do no harm. I would not
> sacrifice this handy feature for just one nasty platform. Anyway, I
> wouldn't feel so pessimistic about that.

I see.  My main concern is not not autodetecting the platform, but
misdetecting it.  That can be tricky on Windows (and remember: we have
quite a few people testing on Windows).  The detected platform must
match, for example, the syntax for paths in PREFIX (see section 5
"Platform issues" in README).

Best wishes.
Mario
-- 
http://parenteses.org/mario



reply via email to

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