GMP 5.1.1 make check fail on Solaris 11
marc.glisse at inria.fr
Fri Feb 15 14:57:45 CET 2013
On Fri, 15 Feb 2013, bodrato at mail.dm.unipi.it wrote:
> Il Ven, 15 Febbraio 2013 10:55 am, Torbjorn Granlund ha scritto:
>> If we reorder the test, so that he last 4 tests that are run are
>> t-does-exception-work-at-all, t-assign, t-constr, and t-rand are run
>> last, it might give users a clue. If these are the only failing tests,
>> and users feel like using the C++ with potentially dangerouse exception
>> behaviour ("derail execution to some arbitrary position in memory"),
>> then that's their decision.
> Or we can move all exception checking from t-assign, t-constr, and t-rand
> to a single t-exception, so that the user knows if only exceptions fail or
> if there are potentially some problems also in rand, assign...
It makes sense, but I don't think we are making the users a favor by
letting them believe it is ok to use a broken compiler.
More information about the gmp-bugs