This is the mail archive of the binutils@sources.redhat.com mailing list for the binutils project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: PATCH: Set xfail on empic for Linux/mips


Last post on this thread, it sounds like others have actually
addressed the underlying issues!  Yay.

At Tue, 1 Oct 2002 02:45:52 +0000 (UTC), "H. J. Lu" wrote:
> If noone is going to support embedded PIC, it may not work with the
> future binutils. I know nothing about embedded PIC nor use it. Someone
> else has to spend time on verifying and fixing it if it is broken by
> accident.

For the record: Helping to provide "support" and "testing" for empic
is _exactly_ why a bunch have empic test cases and bug fixes have been
added (by me) over the last few years.  The code now works reliably,
with full functionality, for MIPS ELF, when originally it barely
worked, only for small test cases, and wasn't complete.  So, I think I
pretty much _have_ been supporting embedded PIC for MIPS for the last
bit of time...

I expect that if i were to _stop_ actively doing so, at least test
case failures wouldn't occur.  Otherwise, what's the point of having
test cases?  That doesn't say the code would continue working -- the
binutils test suite is sparse, and there are a lot of corner cases to
be had even yet, i'm sure -- but at least the stuff covered by the
existing, working test cases shouldn't break and stay broken.

A few of us who hack MIPS binutils, if we took what I interpret as the
attitude stated above, could say we don't care about Linux test
failures, because we don't know much if anything about it and we
certainly don't care for it as our free OS of choice.  If that end up
costing _you_ time and effort down the road, well, I doubt you'd like
it.  But in any case, that's not the attitude we're inclined to take,
I think.  It may save us a little effort in the short run, but it's
not productive for us in the longer term, or for the community at
large.  This is, after all, supposed to be cooperative development...


Note that I'm not claiming that you did anything that caused immediate
breakage here, AFAIK you didn't.  But once things which 'should work'
get XFAILed or otherwise ignored, they really do bit rot quickly over
time, to the point where it takes a lot of extra effort to remember
where they went wrong.  (I can point to examples.)


cgd
-- 
Chris Demetriou                                            Broadcom Corporation
Principal Design Engineer                     Broadband Processor Business Unit
  Any opinions expressed in this message are mine, not necessarily Broadcom's.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]