Re:"Is it dead yet?"

Re:"Is it dead yet?"

>For these ports that have clearly not compiled
> in a long time, is it
>really worth keeping them alive?  I'm not 
>suggesting we drop anything
>we've seen any activity in, but I've watched 
>these lists for a long time
>and don't recall a single problem or success 
>report on, say, clipper-*,
>tahoe-*, pyr-*, or 1750-* in years.  If the 
>code won't compile and
>nobody has even noticed, it sounds like dead 
>wood.

>A quick check on some of those targets diffing 
>against the initial "1.1"
>version when EGCS was created shows nothing 
>but bulk search and replace
>kinds of fixes.

>The GDB project has instated a plan to try to >trim some of these out of
>the code.   Does GCC have a similar plan?  
>Does it need one?

Within the last year I used the 1750 files
to create a compiler.  The company I work
for is building an updated version of the
1750 and we needed tools for it.

From the start I knew that the 1750 wasn't
actively supported and 1750 expertise was
not likely to be available.  However, I did
have the source code and was able to take
advantage of the gcc expertise that is
out there to get a working tool.

The fact that there were existing 1750 files 
saved me a great deal of effort.  Creating
them from scratch would have been much
harder than just modifying them, epecially
since I started with no knowledge of .md 
files and the like.

I would prefer that these files be kept 
around.  It is possible that the next
processor we are asked to build will be
based on the tahoe-* and the existing 
files are certainly better than nothing.

Jeff


--== Sent via Deja.com http://www.deja.com/ ==--
Share what you know. Learn what you don't.

Re: "Is it dead yet?"

[Robert Lipe]

> >For these ports that have clearly not compiled in a long time, is it
> >really worth keeping them alive?  I'm not

> Within the last year I used the 1750 files
> to create a compiler.  The company I work

(I knew I shouldn't have included specific examples...)

I was asking less about moving specific ports into "archival-only" mode
than about the policy of doing it in general.  It's surely a good idea
to keep these old ports publicly available.  I was wondering if it was
worthwhile to just pull them out of the main tree.

> The fact that there were existing 1750 files 
> saved me a great deal of effort.  Creating

Did you submit your work back so the next person doesn't have to
duplicate it?

RJL

Re:"Is it dead yet?"

[Gerald Pfeifer]

On Thu, 13 Jan 2000,     wrote:
> Within the last year I used the 1750 files to create a compiler.  The
> company I work for is building an updated version of the 1750 and we
> needed tools for it.

Have you already submitted your changes for inclusion into the official
GCC sources? Respectively, are you planning to do so? When?

> From the start I knew that the 1750 wasn't actively supported and 1750
> expertise was not likely to be available.  However, I did have the
> source code and was able to take advantage of the gcc expertise that
> is out there to get a working tool.

Great.

> The fact that there were existing 1750 files saved me a great deal of
> effort.  Creating them from scratch would have been much harder [...]

Well, I just checked our ChangeLogs and it seems that you have not
submitted back your changes, which is a real pitty. Perhaps you could
talk your company into submitting your changes for inclusion into our
official tree?

Please have a look at < http://gcc.gnu.org/contributewhy.html >, which
has some thoughts on such issues.

Gerald
-- 
Gerald "Jerry" pfeifer@dbai.tuwien.ac.at http://www.dbai.tuwien.ac.at/~pfeifer/

Re: Re:"Is it dead yet?"

[Gerald Pfeifer]

On Fri, 28 Jan 2000,     wrote:
> The 1750A files that are with the GCC releases conform to this
> standard. You definately want to keep these files as they are.

Perhaps you made some improvements or fixes to these files that you
could submit? If we really keep that port, that definitely would be
nice. (If you look at the ChangeLogs you'll notice that some cleanups
have been made last year by Jeff Law and others, when changing some
basic GCC structures.)


If the scope of your new development really is that small as to be only
interesting to a single customer, I agree that it's probably not the case
to integrate that.

However, it still might be worthwhile to put the changes on your web
server and submit it for inclusion at < http://gcc.gnu.org/extensions.html >


Of course, if you made improvements to generic parts of the compiler
to accomodate the new design, these would be interesting in any case.

Gerald
-- 
Gerald "Jerry" pfeifer@dbai.tuwien.ac.at http://www.dbai.tuwien.ac.at/~pfeifer/

Re: Re:"Is it dead yet?"

On Fri, 28 Jan 2000 21:50:01   Gerald Pfeifer wrote:
>On Thu, 13 Jan 2000,     wrote:
>> Within the last year I used the 1750 files
>> to create a compiler.  The
>> company I work for is building an updated 
>> version of the 1750 and we
>> needed tools for it.
>
> Have you already submitted your changes for
> inclusion into the official
> GCC sources? Respectively, are you planning
> to do so? When?
>

The changes haven't been sent in and I
doubt that they will be.

The short term answer is that we haven't
tested it enough to release it outside the
company.  The company won't export a buggy
release under any circumstances.

The long term answer is more complicated but
I'll explain.  Almost 20 years ago the USAF
created the Instruction Set Achitecture for
the MIL-STD-1750A.  This described to the
hardware manufacturers what the ISA of the
1750A was supposed to be.  Many versions of
this CPU were made but they all conformed to
the standard.  The 1750A files that are with
the GCC releases conform to this standard.
You definately want to keep these files as
they are.

They processor we are developing also conforms
to the standard but adds some features.  These
include 32 bit addr and data busses, rather
than the standard 16 bit widths.  To accomodate
this we had to add instructions which are not
part of the standard.  Our updates should not
be considered replacements for what already
exists, they really describe a new processor
and won't work with any other 1750 in existance.

The only people who might care about our source
files are our customers, who will get them 
with their purchase whether they want them 
or not.  I did not offer to send in these files
because I estimate zero demand for them through
your site, not worth the effort it would take
to add them.

If you still want the files I'd be happy to 
send them in when we release them to our
customers.

Regards,

Jeff


--== Sent via Deja.com http://www.deja.com/ ==--
Share what you know. Learn what you don't.