Re: :( (egcs tests)
On Fri, 5 Dec 1997, Nikita Schmidt wrote:
> Which version of dejagnu do you use? Is it from ftp.cygnus.com:/pub/egcs
> or some official version?
I used the one from cygnus. I believe it's required since they claim the
older one won't work (I guess expect has been hacked again :P).
> I attach my testsuite results for comparison:
> === gcc Summary ===
> # of expected passes 4874
> # of unexpected failures 9
> # of expected failures 5
> # of unsupported tests 7
> /usr/home/cetus/egcs-971122/gcc/xgcc version egcs-2.90.18 971122 (gcc2-970802 experimental)
> --- Note that these failures are known floating point related problems,
> and Richard Henderson is said to be working on them.
I just wiped my results, but gcc turned in much more failures for me. I
believe it might've been a misconfiguration on my part. I've rebuilt egcs
with no options now just to test it out.
> === g++ Summary ===
> # of expected passes 3330
> # of unexpected failures 66
> # of unexpected successes 2
> # of expected failures 86
> # of untested testcases 6
> /usr/home/cetus/egcs-971122/gcc/testsuite/../xgcc version egcs-2.90.18 971122 (gcc2-970802 experimental)
I didn't even pass 80 of the tests. It's halfway through the gcc tests
now, so I'll see how it does this time.
> --- Not that bad. Did you mean "_quite_ a few g++ tests" that worked?
No, quite a few meaning more than 50 :P
> === libio Summary ===
> # of expected passes 40
> === libstdc++ Summary ===
> # of expected passes 30
I passed this also before. The libs seem to be right on :)
TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to
Trouble? e-mail to firstname.lastname@example.org .