different eigenvalue error handling for non-numeric input
by Sven Schreiber
Hi,
consider this with gretl 2025a:
<console>
? matrix hello = {1/0} # to get invalid input
? = eigensym(hello)
Not a Number geschah bei Berechnung
? = eigen(hello)
1.#INF
? = svd(hello)
nan
</console>
So we have everything: a hard-stopping error, some IEEE code (?), and a
gretl code. Plus, with a 6x6 matrix input I also saw that eigen() spat
out a "Data error".
I think it would be good if this could be a little harmonized, no?
thanks
sven
6 days, 4 hours
plots, mouse zooming
by Sven Schreiber
Hi,
I was wondering if it could be possible without too much effort to
enable mouse-wheel-based scrolling or zooming as described here:
http://gnuplot.info/docs_6.1/loc14336.html
I know that in a gretl-controlled plot one can already select "zoom..."
and then define a rectangle with the mouse. This works nicely.
I also know that after saving the plot as a session icon and then
executing the raw gnuplot code, one gets a native gnuplot window (not
gretl-controlled), and there the mouse-wheel thing works as well. (This
is on Windows 10.) So as an alternative route it might also be a
possibility to make it easier to send the plot directly to native
gnuplot, instead of having to go via the raw plotting code first.
Thanks
sven
2 weeks, 4 days
problem with omit --auto=BIC
by Sven Schreiber
Hi,
with a recent snapshot I'm seeing weird behavior when playing around
with the relatively new auto-omit option through BIC. Sorry, I cannot
provide a self-contained example with public data yet. This is with
7-days-daily data.
First, what I'm seeing is that with --auto=BIC the constant term is
eliminated. It doesn't do that with --auto=0.05 and in this data it
doesn't seem to make sense. I thought that the constant is always kept,
at least per default?
Along the elimination sequence I'm seeing: " Drop const BIC = 1.#QNAN"
... this already looks fishy, something going wrong here? There are more
lines with the BIC = 1.#QNAN value.
And then at the end gretl tells me:
"Test statistic: F(30, 128) = 157.789, p-value 5.72395e-087
Omitting variables improved 0 of 3 information criteria."
Needless to say, the resulting model is very dubious.
Maybe this information already helps a little to diagnose the problem.
thanks
sven
1 month
dbnomics w/o open dataset
by Sven Schreiber
Hi all,
just a quick note: running an dbnomics query ("data ...") without an
active dataset doesn't result in an error (should it?), but it seems to
confuse gretl afterwards.
cheers
sven
1 month, 1 week
saving and loading user-defined functions
by Allin Cottrell
A small novelty in git which may be of use to developers: you can
now use the "store" command, with a new --functions option, to save
function definitions as XML. Material saved in this way can then be
loaded using "include" on the XML file.
What's the point? The short story is that it should be a bit faster
to load functions from XML rather than "raw hansl", if you have a
set of several, possibly complicated, functions. And there's also a
potential benefit for anyone writing a C program which uses libgretl
independently of the gretl CLI and GUI clients.
Why faster? When gretl is handling a raw hansl function definition,
we run the full parser for hansl statements, and carry out a number
of checks (e.g. to see if there's a mismatch in the use of if, elif
and endif). But if a function definition has already been processed
into XML by libgretl (as happens with gfn packages) we assume it's
pre-checked, and can simply use libxml2 to zap the function into
memory.
What about the C angle? Up till now there has been no
straightforward way to load hansl functions in the context of a
"libgretl C program", but now you can do
<C>
int err;
err = load_XML_functions_file(<XML filename>, OPT_NONE, NULL);
</C>
and the functions will be pulled into memory.
Here's a little illustration of usage in gretl or gretlcli:
initial script:
<hansl>
function scalar trivial (scalar a)
return (a+3)^2
end function
store trivial.xml --functions
</hansl>
and later:
<hansl>
include trivial.xml
eval trivial(2)
</hansl>
An XML file produced in this way will contain all the functions that
are currently defined, other than those that belong to gfn packages
-- since there should already be an external XML representation of
the latter.
Allin
1 month, 1 week
windows dark theme: problem with icon view
by Sven Schreiber
Hi,
here's a theming problem with Windows-10-dark: The window with the icon
view keeps its white background (I guess it shouldn't), but the font is
then also white (as it should), so the icon names are invisible.
Not sure whether it's something that would have to be adjusted in the
gtk.css stylesheet or whether it's something for the C level.
This is on Windows 11.
thanks
sven
1 month, 2 weeks
