GNU bug report logs - #15160
Is --disable-posix excluding too much?

Previous Next

To add a comment to this bug, you must first unarchive it, by sending
a message to control AT debbugs.gnu.org, with unarchive 15160 in the body.
You can then email your comments to 15160 AT debbugs.gnu.org in the normal way.

Toggle the display of automated, internal messages from the tracker.

Message #5 received at submit <at> debbugs.gnu.org (full text, mbox):

From: Jan Schukat <shookie <at> email.de>
To: bug-guile <at> gnu.org
Subject: Is --disable-posix excluding too much?
Date: Thu, 22 Aug 2013 13:27:10 +0200

Hello, it's been a while since I wrote here. Took me a while to figure 
out how to best tackle the native type vector alignment thing properly. 
Ended up writing my own binary data formats and the smaller dynamically 
created ones get their own SCM_DEFINE c constructors. The compiled guile 
byte-vector format isn't too efficient anyway atm. Once true 
elf-binaries are generated I guess that is different, since they can be 
directly and natively embedded. When that comes along I might look into 
it again. It's always good to consolidate and unify code that can be, to 
reduce redundancy.

Anyway, new issues arise. Using a scripting language somewhat portably 
for non-performance critical management tasks is pretty normal right? 
One of the prime uses.
So, I'm walking directory trees now, and occasionally need to copy a 
file. But copy-file is not available when --disable-posix is configured. 
There are lot's of possible workarounds. Copying files in different 
languages is first semester programming assignments. The point is: a 
high level language shouldn't make you do this. File systems and paths 
are pretty much the same on all platforms guile runs on. And that's the 
the interface to the function: two paths. There is no reason to to not 
have that function everywhere.

Similarly for chdir. You can work around it, and the file tree walking 
functions make it mostly unnecessary. But I see no reason to leave it 
out. Any system guile runs on knows paths.

(oh, and --disable-posix is necessary to compile on/for windows/mingw, 
the cost of portability)

Regards

Jan Schukat

Message #10 received at 15160-done <at> debbugs.gnu.org (full text, mbox):

From: Andy Wingo <wingo <at> pobox.com>
To: Jan Schukat <shookie <at> email.de>
Cc: 15160-done <at> debbugs.gnu.org
Subject: Re: bug#15160: Is --disable-posix excluding too much?
Date: Tue, 21 Jun 2016 11:41:06 +0200

Hi!

--disable-posix is intended to be a size reduction thing, not
necessarily a portability hack.  Guile should build on MinGW with POSIX
enabled; if it doesn't that's a different bug :) When POSIX support is
enabled we should simply compile out features that aren't detected on
the platform.  For that reason it doesn't make sense to take more things
out from under --disable-posix.

Regards,

Andy

On Thu 22 Aug 2013 13:27, Jan Schukat <shookie <at> email.de> writes:

> Hello, it's been a while since I wrote here. Took me a while to figure
> out how to best tackle the native type vector alignment thing
> properly. Ended up writing my own binary data formats and the smaller
> dynamically created ones get their own SCM_DEFINE c constructors. The
> compiled guile byte-vector format isn't too efficient anyway atm. Once
> true elf-binaries are generated I guess that is different, since they
> can be directly and natively embedded. When that comes along I might
> look into it again. It's always good to consolidate and unify code
> that can be, to reduce redundancy.
>
> Anyway, new issues arise. Using a scripting language somewhat portably
> for non-performance critical management tasks is pretty normal right?
> One of the prime uses.
> So, I'm walking directory trees now, and occasionally need to copy a
> file. But copy-file is not available when --disable-posix is
> configured. There are lot's of possible workarounds. Copying files in
> different languages is first semester programming assignments. The
> point is: a high level language shouldn't make you do this. File
> systems and paths are pretty much the same on all platforms guile runs
> on. And that's the the interface to the function: two paths. There is
> no reason to to not have that function everywhere.
>
> Similarly for chdir. You can work around it, and the file tree walking
> functions make it mostly unnecessary. But I see no reason to leave it
> out. Any system guile runs on knows paths.
>
> (oh, and --disable-posix is necessary to compile on/for windows/mingw,
> the cost of portability)
>
> Regards
>
> Jan Schukat

Previous Next

GNU bug tracking system
Copyright (C) 1999 Darren O. Benham, 1997,2003 nCipher Corporation Ltd, 1994-97 Ian Jackson.