Anyone have reason why we should wait?
Dear Allin,
The package editor seems to not know yet
about new feature of matrices as a function
argument
When I try to package something like that
function void ma (scalar x , matrix m[null])
isnm = isnull(m)
if isnm
print "null matrix"
else
print m
endif
end function
After saving the ma.gfn
The function becomes
function void ma (scalar x , matrix m)
i. e. "[null]" is autodeleted
Oleh
7 квітня 2016, 00:17:28, від "Allin Cottrell" <cottrell(a)wfu.edu>:
On Wed, 6 Apr 2016, Sven Schreiber wrote:
> Ok, I have found the (or at least one) reason for the crash, and it has
> nothing to do with the whole R-business. Here's a minimal (?) example that
> crashes for me every time:
>
>
> matrix m = ones(3,1)
> ix = 1
> loop 3 # must be bigger than the 2 in the if-clause
> if (ix <= 2) || (m[ix] > 0.5)
> print "hey"
> endif
> ix++
> endloop
>
>
> Why do I always have to get these things? Sigh.
I can see why you're sighing. But anyway, thanks for bringing this
bug to light; I think it must have been lurking for quite a while.
it was to do with a subtle type-mixup between the two expressions
joined by "||" (OR). (I mean an internal mixup, nothing wrong with
your hansl code.) It's now fixed in git and snapshots.
BTW, I think we have by now accumulated more than enough bug-fixes
to justify a new release. Anyone have reason why we should wait?
Allin
_______________________________________________
Gretl-devel mailing list
Gretl-devel(a)lists.wfu.edu
http://lists.wfu.edu/mailman/listinfo/gretl-devel