Could the following text and screenshot be added as a new section in
Appendix C in the Gretl User's Guide?
Headline: Getting the source from CVS in Microsoft Windows using
In Windows, the easiest way to get the source is by using a CVS client such
as TortoiseCVS, which can be downloaded from http://www.tortoisecvs.org/.
After installing TortoiseCVS, open Windows Explorer and choose the File
menu entry or click on the right mouse button. You will then see two new
menu entries: CVS Checkout... and CVS. Choose CVS Checkout... and enter as
Protocol: Password server
Repository folder: /cvsroot/gretl
User name: anonymous
Then, just click on Fetch list... and you should got something like the
attached screen shot. Choose gretl to get the source code.
(See attached file: TortoiseCVS - Checkout Module.jpg)
I've uploaded most files for the 1.8.0 release to
However, I realize that the OS X build is not quite in sync
(doesn't have the --quiet option to quantreg). So I'll do that
build again later today.
I plan to make the release on sourceforge tonight or tomorrow
unless I hear of any last-minute issues.
I've had a report of problems (won't run!) with the current
gretl build (and I think this applies to the 1.7.9 release too)
on OS X 10.5 Leopard. But I'm not sure how general the
problems are; I suspect idiosyncratic breakage.
The current OS X snapshot
was built on OS X 10.4. It's not quite in sync with CVS but
pretty close. I'm not able to test on 10.5: could anyone who's
running 10.5 please confirm/deny that they're able to run the
snapshot? Thanks very much. (If nobody can run it we need to try
for a fix before the 1.8.0 release.)
I should perhaps mention: the gretl laucher incorporates code
which is supposed to handle the different behaviour of X11 on OS X
10.5 versus 10.4 -- but, as I say, I'm not in a position to test
Department of Economics
Wake Forest University
I was teaching someone how to use gretl for research and while showing
the various plot options, the first thing she noticed and asked about
was whether it is possible to have "bar graphs". I think it would
really be useful if we could select "bars" among the various line
types such as points, impulses and steps. This would really open up
for some very nice looking plots involving combinations of lines and
vertical bars especially useful for time-series analysis.
"Remember not only to say the right thing in the right place, but far
more difficult still, to leave unsaid the wrong thing at the tempting
moment." - Benjamin Franklin (1706-1790)
good afternoon everybody,
I'm using the new R integration and have managed to get R to write out
matrices as text files as described in the "Gretl and R" chapter. I
think it's worth repeating myself that the R integration is a very cool
Just a small thing that I've noticed, I guess it doesn't even qualify as
a bug: when I estimate a spectrum in R with something like:
my.spc <- spec.pgram(...)
and then I want to export the estimated spectra with:
it works, but the saved matrix file is named '$.mat'. The obvious
workaround is to do:
myspec = my.spc$spec
I cannot open the attached csv file which has semicolons as field
separators. (I get errors saying it's not rectangular.) However, if I
replace all semicolons with commas and do nothing else, everything works
fine, so I think it's a bug and not a malformed file.
This is with yesterday's cvs checkout.
I seem to have introduced an encoding issue with the latest de.po
update, causing problems with all non-ASCII characters. I guess it has
to do with the collision that happened (I tried to commit de.po when in
the meantime all .po (or is .pot in this case?) files were updated by
you; the changes were somehow merged together automatically, I edited
the file again and then committed.)
It would be nice if you could wait with the release until this is fixed.
I hope to get it done today, but I may have more technical questions
(Sorry for the cross-posting to the two lists, but I wanted to
make sure I got everyone).
We're now close to the point of releasing gretl 1.8.0, having
assembled quite a few new features and a lot of bug fixes.
See http://gretl.sourceforge.net/ChangeLog.html .
But the most exasperating thing in this business is when you make
a new release and immediately get a report of a serious bug and so
have to make a further release right away.
In this regard, everyone can help. The Windows snapshot, at
and the OS X (Intel) snapshot, at
are both up to date with gretl CVS as of January 12.
If you can find 15 minutes to run your favourite gretl scripts, or
try your favourute gretl GUI operations, then please do, and
report any problems on one or other of the mailing lists.
If any errors come up, we'll fix them tout de suite, then make
gretl 1.8.0 available. Also, if any translators have things they
want to update for the forthcoming release, please do that in the
next day or two. Thanks!
Department of Economics
Wake Forest University
I'm not sure if this is supposed to work or not, but in light of the
stress-testing of the release candidate I report the following:
Gretl failed to recognize the dates in a csv file with monthly data in
the format 'yyyy-mm' (from the US Fed), where the first observation is
in April 1953; here's an excerpt of gretl's output:
erster Zeilenkopf "1953-04", letzter Kopf "2008-12"
versuche Zeilenbezeichner als Daten zu parsen...
1953: wahrscheinlich ein Jahr... Behandlung als undatierte Daten
So gretl correctly finds the date labels, but then it still treats the
data as undated. (Of course it's easy to fix, so at most a minor bug.)
BTW, just a practical hint about a loosely related issue: quarterly data
files from the US BLS often have additional lines with the annual
aggregates which complicates the import -- I think we even discussed
this here on this list. Now I found out that the jEdit editor has the
cool feature of "cut lines containing..." so with just two clicks you
can get rid of those 'annual' lines. Normally I don't like Java programs
that much, but jEdit also has rectangular selection, so I like it better
a new feature is that gretl prints "using p lags of (1-L)x" in the ADF
test output. However, if the option "test down from maximum lag order"
is used, it seems that the new information line is not printed in the
output (today's cvs).