kosaki.motohiro / gmail.com wrote:
> >We should expect users of this to be able to read and follow
> >documentation. We already have a warning about it. I prefer we
> >allow it to improve, and allow users to shoot themselves in the
> >foot if necessary. 
> 
> This is unrelated what I said. I said current interface (both C level and Ruby level) is not designed well and then,
> user have _no way_ to write correct code. Ruby code have no way to care about memory alignment. I didn't only talk 
> about just careless user.

We can support String#pack for those cases, I think (or add support).
I haven't tried the incompatible functions/arch, yet.

> >As I've said before: I don't want Ruby to be
> >a nanny scripting language.
> 
> I agree. But I don't think this patch is a right direction.
> I'm curious. Why you don't like to make proper new C extension?
> C program have a _way_ to treat syscall(2) interface correctly.

C extensions require user to either have a compiler, or install
a pre-built binary.  Both have extra distribution and
installation costs which are high for small (old i686) systems
and users with limited bandwidth/storage.  For those reasons, I
prefer to use scripting as much as possible.   Over the past
year or so, I've been trying to avoid programming in any
compiled languages.

> >Anyways, I plan on having this release GVL for slow syscalls
> >and maybe other small improvements.
> 
> Wait. This? Which patch do you talk about? As far as I can see, current attached patch only remove a warning. Doesn't it?

This patch to undeprecate, first.  I have not implemented GVL
release, yet; I will if I can get this undeprecated.

Unsubscribe: <mailto:ruby-core-request / ruby-lang.org?subject=unsubscribe>
<http://lists.ruby-lang.org/cgi-bin/mailman/options/ruby-core>