for some reason we didn't talk at the conference about the possibility
of using Twitter to spread information about gretl, in parallel to other
Suppose I were to open a quasi-"official" twitter account for gretl,
what would be your arguments against it? The tweets would contain mainly
links to news about gretl I guess (features, tutorials, conferences,
Not going to happen in the next days (weeks?) though...
as we know MPI block has an option "--send-functions" for sharing any
functions defined in the invoking script with MPI nodes. But what about
Suppose I have a package consists of three functions: foo1, foo2 and
foo3 and inside the foo3 I have MPI block in which I want to use foo1
and foo2. Will the "--send-functions" option then work? Because I
suppose for now it doesn't.
Hi everybody and especially Hélio,
one (Greek, I think) participant at the Gretl conference in Athens said
that Gretl needs a suite of unit tests. We replied that Hélio had done
some work on that, but nobody seemed to know more details.
Hélio, could you bring us up to date on that? Is it working? If not,
what needs to be done?
anonymous use inside dummify fails:
series Y = $obsmajor
list dY = dummify(Y) # works
list dY2 = dummify($obsmajor) # doesn't work
This is the June-17th snapshot.
I frequently encounter problems in pasting text into the gretl script window. It happens that the text from the clipboard cannot be pasted. I am using the clipboarder tool, which allows to have a long history of snippets in the clipboard and switch through them fast, but this problem seems not to be related to this problem. In specific, if I copy text from any application (like this outlook window I am currently writing in) I do not encounter any problems even moving back and forth in this history. Except, if there happens to be a snippet from a gretl script itself, sometimes. If I paste this snippet, it simply seems not to work. If I paste it in some other non gretl text-window and copy it again, I can paste it to gretl. This makes me think it could have to do something with how the text is formatted (but what I cannot see). Anyway, I cannot pin down this problem any further. I do not have this problem with no other application, why it seems to be gretl specific.
I use the current snapshot of gretl on windows, but I had this problem as long as I remember using gretl...
I'm trying to make some plots (a lot of plots!) using gnuplot command
but Gretl gives me a warning message telling "File exists".
This behavior occurs after the 26th plot. I would like to know if this
is a feature or a bug ;-)
Please take a look athe following script (running on Windows 7 / Gretl
2017c-git build 2017-06-09):
scalar max_plot = 30
nulldata 100 --preserve
setobs 12 1980:01
loop i = 1..max_plot
series s$i = normal(0, 1)
gnuplot s$i --with-lines --time-series --output=display
at the Gretl conference it seemed that sometimes function package
authors think they would have to write a big amount of documentation
(help file) for their package. So I want to repeat that this is not the
case (especially not for the first two or three releases)!
Yes, a help file is mandatory. Yes, every argument of each public
function must be briefly explained, in addition to the example script
which is also needed. But that's about it. I think everybody agrees that
it is better to make new packages available instead of postponing them.
(Of course, if the package is already some years old then authors might
be asked to provide additional documentation if they update their
package. But that's a different situation.)
So I encourage everybody who has some package sitting on their machines,
please submit/upload them with minimal documentation. (Including those
who were at the conference!)
this is with 2017b on Windows.
- Trying to grab $model where no model was estimated I get a "data
error". The message could be more specific.
- The keyword "const" is not syntax colored if it appears as in "(const"
-- this is something I've noticed for a long time, but obviously not
important. But suppose I master the handling of git some day, I guess
there is some kind of regexp in the source where one could correct that?
Where would that be, or how could it be found without knowing?
- I like the auto-completion-suggestion feature in the script editor,
but sometimes it drives me crazy:
I type the last word/token in the line and simply hit return to go to
the next line. But the auto-suggestion suggested something in the
meantime and automatically replaces my word with the first suggestion in
its choice list, even though I did not actively accept it (other than
hitting return, that is). Any ideas how to avoid that? (If I made myself