some package using data-dependent default values for non-int scalars?
by Sven Schreiber
Hi all,
referring to the function package guide section 4.3.2, "Run-time limits
and defaults". This should apply to int(egers) and scalar(s), and I
think for integers I have used that successfully at least for the upper
and lower limits (not sure about the defaults).
But for general scalars, I'm having problems cooking up a working
example - so is there a package out there that uses this feature?
Or going back to integers, a package that does run-time defaults, not
just limits?
thanks
sven
4 days, 3 hours
ui-maker and missing warning messages
by Sven Schreiber
Good morning,
this is again about authoring a function package and using ui-maker.
Take conditional sensitivity, for example. I'm often making the mistake
that I don't use a string there, i.e. I don't put the name of the
argument in quotes. Example: _(depends=hey) instead of _(depends="hey").
When building the package (via gretlcli -m), gretl doesn't complain
about this, it just doesn't work (understandably). It would be helpful
if gretl complained if the type is wrong in this context, i.e. not a string.
thanks
sven
5 days
string-valued series and string functions
by Sven Schreiber
Hi,
I know this is very much in the nice-to-have-but-not-strictly-necessary
category, but working with a cross-country panel where the country codes
are in a string-valued series, I must say it would be nice if functions
like instring() worked for those series. Instead I'm using a chain of
direct equality comparisons, which works fine, but gets a bit tedious.
Probably material for a request ticket.
cheers
sven
1 week, 1 day
Failure to build the latex files in fedora 44 and 45
by Johannes Lips
Dear all,
I am having issues building the latest release for fedora 44 and fedora 45.
--------------------------------------------------------
kpathsea: Appending font creation commands to missfont.log.
! Font U/rsfs/m/n/12=rsfs10 at 12.0pt not loadable: Metric (TFM) file not
found
.
<to be read again>
relax
l.108
?
! Emergency stop.
<to be read again>
relax
l.108
! ==> Fatal error occurred, no output PDF file produced!
Transcript written on gig.log.
make[3]: *** [Makefile:10: gig.pdf] Error 1
make[3]: Leaving directory
'/builddir/build/BUILD/gretl-2026a-build/gretl-2026a/addons/gig/doc'
make[2]: *** [Makefile:42: doc/gig.pdf] Error 2
make[2]: Leaving directory
'/builddir/build/BUILD/gretl-2026a-build/gretl-2026a/addons/gig'
make[1]: *** [Makefile:30: gig] Error 2
make[1]: Leaving directory
'/builddir/build/BUILD/gretl-2026a-build/gretl-2026a/addons'
make: *** [Makefile:73: addons] Error 2
--------------------------------------------------------
The main build fails in the creation of pdfs for addons.
https://kojipkgs.fedoraproject.org//work/tasks/7561/142227561/build.log
Thanks a lot!
All the best,
Johannes
1 week, 1 day
"empty" return value
by Sven Schreiber
Hi,
is the following supposed to work?:
<hansl>
function matrices Mhey(void)
return empty # gives generic "data error", works with defarray()
end function
Mhey()
</hansl>
thanks
sven
1 week, 1 day
spurious description string in pointerized series arg
by Sven Schreiber
Hi,
the attached script files demonstrate a bug. Please run the
bugtest_driver.inp script which includes the other script
port_bugtest.inp. The example uses the built-in example dataset abdata.
The point is that the top-level series "e" which is passed as a pointer
arg to the function (factcoint_test) gets a spurious and more or less
arbitrary description string, sometimes including non-printable
characters. Actually sometimes the correct description happens; in that
case please just run the script again. To trigger the bug it seems
necessary to have an enclosing loop.
In itself it's not a big deal, but since apparently chunks of memory end
up where they do not belong, maybe some related unknown bugs exist and
may be fixed by getting this straight.
This is a fairly recent snapshot.
thanks
sven
1 week, 2 days
about genr markers (from "Panel data marker strings")
by Sven Schreiber
Hi again,
as you will have guessed, I'm juggling around some panel datasets at the
moment. (BTW, the handling overall has become very nice in recent years,
thanks to string-valued series, refinements in panel time setting, and
join, etc.)
I tried the recipe from the guide described under "Panel data marker
strings" in chapter 4. Instead of the given line <genr markers =
"%s:%d", marker, year> I tried a more compact thing, replacing the
"year" series with the formula "1989+time" (where "time" exists
already). This produced no error, but gave rubbish output in the new
markers. I guess a formula expression is not allowed there, but then I
guess it should be intercepted somehow.
(The recipe as given worked OK.)
thanks
sven
1 week, 4 days
crash looping through strings array
by Sven Schreiber
Hi, here's a crash that I guess should be not too difficult to fix. It
only happens (apparently) when using wrong syntax, as I found out by
mistake:
<hansl>
strings A = defarray("a")
# correct:
loop foreach which A
string renamed = "_" ~ "$which"
print "$which"
endloop
# wrong / crash:
loop foreach which A
string renamed = "_" ~ which
endloop
</hansl>
This happens with 2026a, but I guess it's older than that.
thanks
sven
1 week, 5 days
dbnomics, no sync of main window in loop
by Sven Schreiber
Hi,
and here's another little problem, but far less serious.
<hansl>
clear
open dbnomics
country = "ARG"
whichpol = "LEV1_INT"
loop 1
data
OECD/DSD_CAPMF@DF_CAPMF/@country.A.POL_STRINGENCY.@whichpol.0_TO_10
endloop
</hansl>
The thing is that there are no series visible in the main window, as if
no dataset existed. But it's there; when doing, for example, "varlist"
(or "ls") in the console, then variables and data are printed. The
enclosing loop seems to be necessary to trigger the problem - obviously
I originally had many more iterations there, but that's not even
necessary, it seems.
Sometimes the main window is updated after "ls", but right now I also
saw a case where this didn't happen. A workaround is to use "store" and
then reload the result.
This is 2026a on Windows.
thanks
sven
1 week, 6 days