FYI: GCC dropped score support

Discussion:

(too old to reply)

Joel Brobecker

2014-10-06 23:36:49 UTC

While the deprecation theme is playing in the background ...
https://gcc.gnu.org/ml/gcc-patches/2014-10/msg00305.html
I'm looking at score-tdep.c/score-tdep.h and wondering if
it makes any sense to still maintain that...
http://en.wikipedia.org/wiki/S%2Bcore

Thinking out loud...

If the port isn't getting in the way, perhaps we could send out
a quick deprecation warning announce following our standard protocol;
but instead of actually removing support right away, we'd just
keep it on a best-effort basis? I'm not actually sure this is
a good idea at all, as not receiving any response today does not
mean that people might not be starting using it again tomorrow...

So, unless we think there is enough benefit in remove support for
this target sooner rather than later, the other option is to keep
the info you about GCC in mind if it ever starts getting in the way?

--
Joel

Stan Shebs

2014-10-07 00:10:19 UTC

Permalink

Post by Joel Brobecker

Thinking out loud...
If the port isn't getting in the way, perhaps we could send out
a quick deprecation warning announce following our standard protocol;
but instead of actually removing support right away, we'd just
keep it on a best-effort basis? I'm not actually sure this is
a good idea at all, as not receiving any response today does not
mean that people might not be starting using it again tomorrow...
So, unless we think there is enough benefit in remove support for
this target sooner rather than later, the other option is to keep
the info you about GCC in mind if it ever starts getting in the way?

It seems pretty dead, the company's pages are 404, and the only
product I can find that used it was the Mattel HyperScan, introduced in
2006 and discontinued in 2008 - apparently it made PC World's "10 Worst
Videogames of All Time" list.

Even if the code isn't being used, I think it still costs us, in the
sense that we have to think about whether it needs changing, it has
to be patchedwhen there are global changes, and so forth. So I
think we should whack it soon as decently possible. (Of course, the
code will actually live on forever, in old releases and repos. :-) )

Stan
***@codesourcery.com

Joel Brobecker

2014-10-07 00:18:05 UTC

Permalink

Post by Stan Shebs
It seems pretty dead, the company's pages are 404, and the only
product I can find that used it was the Mattel HyperScan, introduced in
2006 and discontinued in 2008 - apparently it made PC World's "10 Worst
Videogames of All Time" list.
Even if the code isn't being used, I think it still costs us, in the
sense that we have to think about whether it needs changing, it has
to be patchedwhen there are global changes, and so forth. So I
think we should whack it soon as decently possible. (Of course, the
code will actually live on forever, in old releases and repos. :-) )

Seen under this light, it does make sense!

--
Joel

Pedro Alves

2014-10-07 08:35:09 UTC

Permalink

Post by Joel Brobecker

Eh. :-)

Post by Joel Brobecker

Post by Stan Shebs
Even if the code isn't being used, I think it still costs us, in the
sense that we have to think about whether it needs changing, it has
to be patchedwhen there are global changes, and so forth. So I
think we should whack it soon as decently possible. (Of course, the
code will actually live on forever, in old releases and repos. :-) )

Seen under this light, it does make sense!

Agreed, that's my thinking too. For still-fresh examples, the software
single-step rework work had me stop, think and actually write patches
for IRIX (deprecated_insert_raw_breakpoint). And then just last week
I failed to notice some code was for Tru64 and wasted time writing
this patch:
https://sourceware.org/ml/gdb-patches/2014-10/msg00035.html

Thanks,
Pedro Alves