regenerating cgen sim sources for gdb 6.0

regenerating cgen sim sources for gdb 6.0

[Doug Evans]

It'd be great if cgen sim sources were regenerated for 6.0.
I'd be uncomfortable with 6.0 being released without up to date files.

Ok if I get someone at Redhat to regenerate them and check them in?

The sims in question are m32r and i960.
While the m32r is still, pedantically speaking, obsolete
and the i960 has been deleted, both targets still exist
in the gcc tree and having a sim to vet gcc changes is enormously useful.

I gather the i960 is on the way out even in gcc land though.
If Jim's ok with deleting src/sim/i960 fine.
Otherwise, I think the files should be updated.

I've already verified that there's no regressions in the
gcc testsuite from gcc top-of-tree sources (as of a few days ago)
with the before/after i960-coff-run simulator.

Re: regenerating cgen sim sources for gdb 6.0

[Frank Ch. Eigler]

dje wrote:

> It'd be great if cgen sim sources were regenerated for 6.0.
> I'd be uncomfortable with 6.0 being released without up to date files.
> [...]
> The sims in question are m32r and i960.
> [...]

... plus fr30 and perhaps sh64.

- FChE

Re: regenerating cgen sim sources for gdb 6.0

[Doug Evans]

Frank Ch. Eigler writes:
 > 
 > dje wrote:
 > 
 > > It'd be great if cgen sim sources were regenerated for 6.0.
 > > I'd be uncomfortable with 6.0 being released without up to date files.
 > > [...]
 > > The sims in question are m32r and i960.
 > > [...]
 > 
 > ... plus fr30 and perhaps sh64.

Good grief, right you are.  Mea culpa.

Re: regenerating cgen sim sources for gdb 6.0

[Andrew Cagney]

> It'd be great if cgen sim sources were regenerated for 6.0.
> I'd be uncomfortable with 6.0 being released without up to date files.
> 
> Ok if I get someone at Redhat to regenerate them and check them in?
> 
> The sims in question are m32r and i960.
> While the m32r is still, pedantically speaking, obsolete
> and the i960 has been deleted, both targets still exist
> in the gcc tree and having a sim to vet gcc changes is enormously useful.

Not deleting the i960 was an oversite, I'll delete it.

Please don't look to me for re-generating the m32r, I don't know how to 
use CGEN.

Andrew


> I gather the i960 is on the way out even in gcc land though.
> If Jim's ok with deleting src/sim/i960 fine.
> Otherwise, I think the files should be updated.
> 
> I've already verified that there's no regressions in the
> gcc testsuite from gcc top-of-tree sources (as of a few days ago)
> with the before/after i960-coff-run simulator.
> 

Re: regenerating cgen sim sources for gdb 6.0

[Doug Evans]

Andrew Cagney writes:
 > Not deleting the i960 was an oversite, I'll delete it.

IMO, checking with Jim first would be the thing to do.

 > Please don't look to me for re-generating the m32r, I don't know how to 
 > use CGEN.

I wasn't.  One would think that would be taken as a given.

If someone in a relevant position of authority authorizes the
work, I'll petition some folks at Redhat I know who are authorized
to check files into src/sim to do the work.
Some folks at Redhat are actually quite nice people to work with.

Alternatively, if someone in a relevant position of authority
authorizes me to do the work, I'll certainly volunteer my time.

Re: regenerating cgen sim sources for gdb 6.0

[Andrew Cagney]

> Andrew Cagney writes:
>  > Not deleting the i960 was an oversite, I'll delete it.
> 
> IMO, checking with Jim first would be the thing to do.

I did, something like a year ago.

>  > Please don't look to me for re-generating the m32r, I don't know how to 
>  > use CGEN.
> 
> I wasn't.  One would think that would be taken as a given.

Not really, I get to re-generate a few other files, and this task is 
easiest done by someone from Red Hat.  I guess Frank is the obvious person?

Andrew

Re: regenerating cgen sim sources for gdb 6.0

[Doug Evans]

Andrew Cagney writes:
 > I guess Frank is the obvious person?

Either Frank or Nick.

To regenerate the files

cvs co gdb
cvs co cgen
cd src/cgen
sh gen-all-sim

Figuratively speaking.  I haven't tried this precise sequence.
One runs gen-all-sim from the src/cgen directory in a gdb+cgen src tree.

Beyond that, what one does testing-wise before checkin is up to the person
doing the checkin.  S/he can certainly feel free to enlist my help.

Re: regenerating cgen sim sources for gdb 6.0

[Doug Evans]

Doug Evans writes:
 > Beyond that, what one does testing-wise before checkin is up to the person
 > doing the checkin.  S/he can certainly feel free to enlist my help.

Not that I'm suggesting Frank or Nick is a girl of course ... :-)

regenerating cgen sim sources for gdb 6.0

[Doug Evans]

I wrote:
 > I've already verified that there's no regressions in the
 > gcc testsuite from gcc top-of-tree sources (as of a few days ago)
 > with the before/after i960-coff-run simulator.

... and the m32r changes are minimal, they don't cause regressions either.