Re: [HACKERS] (: JDBC+(Sun ~3:pm MST) CVS :) -also question about regression tests

"Thomas G. Lockhart" wrote:
  >The other good possibility is for Oliver to develop a patch kit (he has done
      > so
  >already) which we can include in the v6.3 distribution to be applied only to
  >linux/glibc2. When the beta settles down perhaps Oliver can generate a new p
      >atch
  >kit which we can include?

I'm happy to remake the patch in a week or so.  However I'd like to be clear
what you think is 'Linux/glibc-specific' (and therefore undesirable); obviously
it is better to make changes as general as possible.

I'm attaching the previous patch I made (which, of course, may not
actually work with the current release - I haven't tried yet) so as to comment
on it:

1.  I use `#if __GLIBC__ < 2' to test for glibc, where the change is only
    for glibc.  So I use it to force HAVE_INT_TIMEZONE to be undefined
    and to redefine TMODULO(); incidentally, I used the hint in the existing
    source that modf() was sometimes broken and chose to use glibc's
    unbroken modf().

    [ Incidentally, is modf() _still_ broken on whatever platforms it was
    broken on? ]

2.  The second set of changes are to do with unnecessary printing of n.00
    instead of n, where the test was whether fsec is non-zero.  With the
    glibc version, fsec is often non-zero but integer, so I substituted
    the test (fsec != rint(fsec)) for (fsec != 0).  This seems to me to
    be a completely general change that will work correctly on all
    implementations.

Now I can understand that you may choose to ignore the existence of glibc
for the standard code (not that I agree, mind you) but I really don't see
how I could make the changes any more general than I have.
Oliver Elphick                                Oliver.Elphick@lfix.co.uk
Isle of Wight                              http://www.lfix.co.uk/oliver

PGP key from public servers; key ID 32B8FAA1