Igor Schein on Wed, 20 Oct 1999 13:45:22 -0400

[Date Prev] [Date Next] [Thread Prev] [Thread Next] [Date Index] [Thread Index]

Re: Bug in nfabsis/Round4

On Wed, Oct 20, 1999 at 01:39:29PM -0400, Ilya Zakharevich wrote:
> On Wed, Oct 20, 1999 at 12:11:15PM -0400, Xavier Roblot wrote:
> > > ? nfbasis(x^16-42*x^12+560*x^10-1645*x^8+2352*x^6+10290*x^4+10192*x^2+3969)
> > >   ***   impossible inverse modulo: Mod(0, 2).
> > 
> > The bug was due to a possible lost of precision in one of the modular reductions
> > performed in nilord2. It is corrected using the following patch (I have also
> > updated the CVS version).
> Are we going to have an actual test suite, or having the current joke
> (one test per function) looks enough?  *At least* it could have been
> populated with bug reports...
> Or I am missing something, and the bug reports *do* result in changes
> to test suite?
> Ilya

Well, I have my own test suite populated by bugs I've discovered over the
last year.  It has over 250 entries in it.  I run it once in a while, to make
sure old bugs haven't made their way into the code, and sometimes they do.

 If any1 has been doing the same, we could merge it, and have
something similar to GNU's testsuite.  Of course,  make dobench still
needs the existing short testsuite for the purpose of deriving