Why do ^C tests fail on linux?

Why do ^C tests fail on linux?

[Michael Snyder]

There is a test in pthreads.exp which sends "continue", then
sends "^C", to see if it can interrupt the child process.
I discovered that it is only passing thru an expect failure, 
and if you make expect/dejagnu do the right thing, the test fails.

Since I know that there is another test in the testsuite that
tests the same thing (interrupt.exp), I checked to see if that
was also failing, and discovered that it has been xfailed for
i*86-pc-linux-gnu since 1996, with no comment to explain why.

Does anybody have any idea why this would not be working on linux?

Michael

Re: Why do ^C tests fail on linux?

[Jim Blandy]

Michael Snyder <msnyder@cygnus.com> writes:
> Since I know that there is another test in the testsuite that
> tests the same thing (interrupt.exp), I checked to see if that
> was also failing, and discovered that it has been xfailed for
> i*86-pc-linux-gnu since 1996, with no comment to explain why.

I can't think of any reason this shouldn't work normally.  Perhaps in
1996 Linux's signal handling or job control wasn't up to par, but
I think we can raise our expectations a bit nowadays.