Re: [9fans] Re: Plan9 and Ada95?

Re: [9fans] Re: Plan9 and Ada95?

[forsyth]

[-- Attachment #1: Type: text/plain, Size: 166 bytes --]

the structure of the support for cross-compilation,
and arranging it for an arbitrary combination,
and the resulting file hierarchy, is also not trivial in gcc.


[-- Attachment #2: Type: message/rfc822, Size: 1819 bytes --]

To: 9fans@cse.psu.edu
Subject: Re: [9fans] Re: Plan9 and Ada95?
Date: Wed, 7 Nov 2001 09:44:41 GMT
Message-ID: <87u1w7z703.fsf@becket.becket.net>

dhog@plan9.bell-labs.com (David Gordon Hogan) writes:

> > Cross-compilation in GCC is a trivial matter.
>
> Having done a GCC port, I'd like to point out that with GCC,
> nothing is a trivial matter.

Cross-compilation is trivial; porting is not.

Re: [9fans] Re: Plan9 and Ada95?

[David Gordon Hogan]

> it[GCC] was awful.  pages of filth.
> and those directory structures!
> bleah.
> perhaps it's better now, but i somewhat cynically doubt it.

It's not.  If anything, it's worse.

Re: [9fans] Re: Plan9 and Ada95?

[forsyth]

[-- Attachment #1: Type: text/plain, Size: 542 bytes --]

absolutely.  that was all in my mind as i wrote that message.
also in my mind was that a few years ago,
i spent a miserable time finding all
the cross-portability problems in gcc cross-compiling
from (i think) a little endian processor to a big endian processor,
using gnu assembler and linker in a combination
that wasn't originally provided by ./configure (and
with good reason, it didn't work).  it was awful.  pages of filth.
and those directory structures!
bleah.
perhaps it's better now, but i somewhat cynically doubt it.


[-- Attachment #2: Type: message/rfc822, Size: 5467 bytes --]

[-- Attachment #2.1.1: Type: text/plain, Size: 467 bytes --]

I've never seen a system that supports cross-compilation
the way Plan 9 does, by calling it compilation and making
no bones about whether the compiling and executing
architectures differ.  The trick for doing this is trivial, yes,
but perhaps the observation that it should be done this way
is not, and should be more widely observed.  GCC, for
example, does not make it easy to walk into a directory
and perform the equivalent of

	mk installall

-rob


[-- Attachment #2.1.2: Type: message/rfc822, Size: 3389 bytes --]

[-- Attachment #2.1.2.1.1: Type: text/plain, Size: 166 bytes --]

the structure of the support for cross-compilation,
and arranging it for an arbitrary combination,
and the resulting file hierarchy, is also not trivial in gcc.


[-- Attachment #2.1.2.1.2: Type: message/rfc822, Size: 1819 bytes --]

To: 9fans@cse.psu.edu
Subject: Re: [9fans] Re: Plan9 and Ada95?
Date: Wed, 7 Nov 2001 09:44:41 GMT
Message-ID: <87u1w7z703.fsf@becket.becket.net>

dhog@plan9.bell-labs.com (David Gordon Hogan) writes:

> > Cross-compilation in GCC is a trivial matter.
>
> Having done a GCC port, I'd like to point out that with GCC,
> nothing is a trivial matter.

Cross-compilation is trivial; porting is not.

Re: [9fans] Re: Plan9 and Ada95?

[rob pike]

[-- Attachment #1: Type: text/plain, Size: 467 bytes --]

I've never seen a system that supports cross-compilation
the way Plan 9 does, by calling it compilation and making
no bones about whether the compiling and executing
architectures differ.  The trick for doing this is trivial, yes,
but perhaps the observation that it should be done this way
is not, and should be more widely observed.  GCC, for
example, does not make it easy to walk into a directory
and perform the equivalent of

	mk installall

-rob


[-- Attachment #2: Type: message/rfc822, Size: 3389 bytes --]

[-- Attachment #2.1.1: Type: text/plain, Size: 166 bytes --]

the structure of the support for cross-compilation,
and arranging it for an arbitrary combination,
and the resulting file hierarchy, is also not trivial in gcc.


[-- Attachment #2.1.2: Type: message/rfc822, Size: 1819 bytes --]

To: 9fans@cse.psu.edu
Subject: Re: [9fans] Re: Plan9 and Ada95?
Date: Wed, 7 Nov 2001 09:44:41 GMT
Message-ID: <87u1w7z703.fsf@becket.becket.net>

dhog@plan9.bell-labs.com (David Gordon Hogan) writes:

> > Cross-compilation in GCC is a trivial matter.
>
> Having done a GCC port, I'd like to point out that with GCC,
> nothing is a trivial matter.

Cross-compilation is trivial; porting is not.

Re: [9fans] Re: Plan9 and Ada95?

[Thomas Bushnell, BSG]

dhog@plan9.bell-labs.com (David Gordon Hogan) writes:

> > Cross-compilation in GCC is a trivial matter.
>
> Having done a GCC port, I'd like to point out that with GCC,
> nothing is a trivial matter.

Cross-compilation is trivial; porting is not.

Re: [9fans] Re: Plan9 and Ada95?

[David Gordon Hogan]

> Cross-compilation in GCC is a trivial matter.

Having done a GCC port, I'd like to point out that with GCC,
nothing is a trivial matter.

Re: [9fans] Re: Plan9 and Ada95?

[Thomas Bushnell, BSG]

arnold@skeeve.com (Aharon Robbins) writes:

> >GNAT would probably come fairly cheaply once we had a port of GCC 3.0.
>
> Actually not.  GNAT is written in Ada, and thus there would need to
> be a cross-compilation boostrap done at some point.

Cross-compilation in GCC is a trivial matter.

Re: [9fans] Re: Plan9 and Ada95?

[Theo Honohan]

Aharon Robbins wrote:
> >GNAT would probably come fairly cheaply once we had a port of GCC 3.0.
>
> Actually not.  GNAT is written in Ada, and thus there would need to
> be a cross-compilation boostrap done at some point.

Damn, I hate it when I accidentally say something that sounds upbeat
on 9fans!  I didn't mean to imply that either of these things
was trivial.

I was trying to point out that a GNAT port would rely heavily on
a GCC port, to the extent that it's not probably worth thinking about
porting GNAT until we have gcc.

In fact, from the point of view of getting things like GNAT working, even
a version of GCC that could cross-compile to Plan 9 would be very useful.
This might be easier than porting all that GNU C to compile with the Plan
9 compilers.  Maybe this is self-evidently the way to go, and I'm just
being thick.

Is there any progress with GCC to report, by the way?

Re: [9fans] Re: Plan9 and Ada95?

[Aharon Robbins]

>GNAT would probably come fairly cheaply once we had a port of GCC 3.0.

Actually not.  GNAT is written in Ada, and thus there would need to
be a cross-compilation boostrap done at some point.

And don't forget the Ada runtime (threading) stuff.  I've no idea
how well it does/doesn't map to Plan 9.

Arnold

Re: [9fans] Re: Plan9 and Ada95?

[forsyth]

[-- Attachment #1: Type: text/plain, Size: 106 bytes --]

See http://plan9.bell-labs.com/sys/man/ and http://plan9.bell-labs.com/sys/doc/
for more information.


[-- Attachment #2: Type: message/rfc822, Size: 2062 bytes --]

To: 9fans@cse.psu.edu
Subject: Re: [9fans] Re: Plan9 and Ada95?
Date: Mon, 5 Nov 2001 09:40:04 -0500 (EST)
Message-ID: <Pine.OSF.4.40.0111050939130.8136-100000@grover.WPI.EDU>

An additional question:  Does Plan9 come with many of the standard
Unix utilities (i.e. perl, c/c++, python, grep, *awk, etc)?

--
Ian Cooper
ian@wpi.edu

On Mon, 5 Nov 2001 forsyth@vitanuova.com wrote:

> Date: Mon, 5 Nov 2001 14:40:55 0000
> From: forsyth@vitanuova.com
> Reply-To: 9fans@cse.psu.edu
> To: 9fans@cse.psu.edu
> Subject: Re: [9fans] Re: Plan9 and Ada95?
>
> >>GNAT would probably come fairly cheaply once we had a port of GCC 3.0.
>
> i think that's the first time i've seen `cheaply' adjacent to `gcc'.
>

Re: [9fans] Re: Plan9 and Ada95?

[Theo Honohan]

forsyth@vitanuova.com wrote:
> >>GNAT would probably come fairly cheaply once we had a port of GCC 3.0.
>
> i think that's the first time i've seen `cheaply' adjacent to `gcc'.

Well, that's relativity for you.

Re: [9fans] Re: Plan9 and Ada95?

[forsyth]

>>GNAT would probably come fairly cheaply once we had a port of GCC 3.0.

i think that's the first time i've seen `cheaply' adjacent to `gcc'.

Re: [9fans] Re: Plan9 and Ada95?

[Ian Cooper]

An additional question:  Does Plan9 come with many of the standard
Unix utilities (i.e. perl, c/c++, python, grep, *awk, etc)?

--
Ian Cooper
ian@wpi.edu

On Mon, 5 Nov 2001 forsyth@vitanuova.com wrote:

> Date: Mon, 5 Nov 2001 14:40:55 0000
> From: forsyth@vitanuova.com
> Reply-To: 9fans@cse.psu.edu
> To: 9fans@cse.psu.edu
> Subject: Re: [9fans] Re: Plan9 and Ada95?
>
> >>GNAT would probably come fairly cheaply once we had a port of GCC 3.0.
>
> i think that's the first time i've seen `cheaply' adjacent to `gcc'.
>

Re: [9fans] Re: Plan9 and Ada95?

[Theo Honohan]

"martin.m.dowie" wrote:
> "Caffienator" <chris@dont.spam.me> wrote in message
> news:FxXC7.205641$K6.100988476@news2...
> > I know C and C++, but my preferred tool for building reliable software is
> > Ada95. Are there any Ada95 compilers available for the Plan9 platform?
> > If not, would a port of the GNAT compiler to Plan9 interest anyone?
>
> I'm interested! Though I'm still struggling to actually get my plan9 to
> boot... :-(

GNAT would probably come fairly cheaply once we had a port of GCC 3.0.
David Gordon Hogan is apparently working on porting GCC.

[9fans] Re: Plan9 and Ada95?

[martin.m.dowie]

"Caffienator" <chris@dont.spam.me> wrote in message
news:FxXC7.205641$K6.100988476@news2...
> I know C and C++, but my preferred tool for building reliable software is
> Ada95. Are there any Ada95 compilers available for the Plan9 platform?
> If not, would a port of the GNAT compiler to Plan9 interest anyone?

I'm interested! Though I'm still struggling to actually get my plan9 to
boot... :-(