On Sun, Sep 07, 2003 at 04:22:26PM +0200, Leopold Toetsch wrote:
> Nicholas Clark <[EMAIL PROTECTED]> wrote:
> > On Thu, May 15, 2003 at 08:35:39PM -0600, Luke Palmer wrote:
> >> Silly me.  Here's that patch without the C++ comments %-)
> 
> > This bug is still marked open:
> 
> > http://rt.perl.org/rt2/Ticket/Display.html?id=22214
> 
> > I don't see this test added, so should this patch be applied?
> 
> Yes. With the remarks WRT is_digit et al merged. The test can go in
> anyway.

Adding just the test gives:

t/op/number............NOK 37#     Failed test (t/op/number.t at line 1005)  
#          got: '1.000000
# 12.000000
# -2.450000
# 2500.000000
# 32.000000
# 6.000000
# 42.650000
# 12.000000
# '
#     expected: '1.000000
# 12.000000
# -2.450000
# 2500.000000
# 0.000000
# 0.000000
# 42.650000
# 12.000000
# '

Which is S4 and S5 from this set:

    set S0, "1"
    set S1, "12.0"
    set S2, "-2.45"
    set S3, "25e2"
    set S4, "Banana32"
    set S5, ".Monkey6"
    set S6, "   42.65"
    set S7, "12Chimp"

ie currently "Banana32" is converting to 32 and .Monkey6" to 6
Is this what we want?

Adding the string.c changes causes other tests to fail.

Nicholas Clark

Reply via email to