On Fri, 1 Mar 2019, Sven Schreiber wrote:
Hi,
here's a problem and --to whom it may concern-- a workaround:
<hansl>
trunc = 4
# set hac_lag trunc # fails
# workaround:
string st = sprintf("%d", trunc)
set hac_lag @st
eval lrcovar(mnormal(10,2))
</hansl>
Ah, the general idea is that predefined scalar variables should work
OK with "set" when a numerical value is wanted. In the case of hac_lag
this facility wasn't working, due to the complication that certain
codes could be given in place of a numerical value ("nw1", etc.).
That's now fixed in git.
Allin