Re: [Gretl-devel] Can't open datafile
by Allin Cottrell
On Tue, 18 Jan 2011, Alexander B. Gedranovich wrote:
> After I compiled latest CVS gretl crashes at opening datafiles (either
> via GUI or console).
>
> Example:
> =============================================
> ales@opensuse:~> gretl arma
> get_gretl_charset: using UTF-8
>
> Read datafile /usr/local/share/gretl/data/misc/arma.gdt
> periodicity: 12, maxobs: 624
> observations range: 1950:01-2001:12
>
> Illegal instruction
What type of CPU, and what version of gcc? It looks as if
something must be going wrong with detection of SSE2 at configure
time, but this puzzling. (SSE2 support was recently added to
support a speed-up of the gretl RNG, and I can't think of any
reason why you'd be getting an "Illegal instruction" message other
than some screw-up related to this.)
Could you show the output from ./configure when you're building
gretl? Output from "make" might be helpful too.
Allin Cottrell
14 years, 10 months
Re: [Gretl-devel] Bug in error message output
by Allin Cottrell
On Sun, 16 Jan 2011, Hélio Guilherme wrote:
> When I was generating a Range-Mean graph for a small sample variable
> (size=10), I got garbage text after the error message. This happens in
> non-English versions, like both Portuguese variants...
> I have commented a line of code and problem was solved:
> ~/gretl$ vi +137 plugin/range-mean.c
>
> if (nsamp < 16) {
> pputs(prn, _("Sample is too small for range-mean graph\n"));
> /* errmsg(err, prn); */
> return E_DATA;
> }
Thanks, you're right, that line you commented out should not be
there. Now fixed in CVS.
Allin
14 years, 10 months
Bug in error message output
by Hélio Guilherme
Hi,
When I was generating a Range-Mean graph for a small sample variable
(size=10), I got garbage text after the error message. This happens in
non-English versions, like both Portuguese variants.
PT:
A amostra é demasiado pequena para o gráfico amplitude-média
Project-Id-Version: pt
Report-Msgid-Bugs-To:
POT-Creation-Date: 2010-12-03 09:35-0500
PO-Revision-Date: 2010-12-06 21:35+0000
Last-Translator: Hélio Guilherme
Language-Team:
Language:
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
X-Generator: KBabel 1.11.4
X-Poedit-Language: Portuguese
X-Poedit-Country: PORTUGAL
X-Poedit-SourceCharset: iso-8859-15
Plural-Forms: nplurals=2; plural=(n != 1);
X-Language: pt_PT
---
I have commented a line of code and problem was solved:
~/gretl$ vi +137 plugin/range-mean.c
if (nsamp < 16) {
pputs(prn, _("Sample is too small for range-mean graph\n"));
/* errmsg(err, prn); */
return E_DATA;
}
Bye,
Hélio
14 years, 10 months
gretl conference submission question
by Sven Schreiber
Hi,
it just occurred to me that I might want to send a paper to the gretl
conference :-)
However, after visiting the page http://gretl.konferencja.org/ I feel a
little confused. Do I have to register for the conference to submit a
paper? (Normally one first submits and later decides if one wants to go.)
Any advice is appreciated.
thanks,
sven
14 years, 10 months
Re: [Gretl-devel] Bootstrapped IRFs for restricted VECMs
by Allin Cottrell
On Thu, 13 Jan 2011, artur tarassow wrote:
> 2011/1/13 Allin Cottrell <cottrell(a)wfu.edu>:
> > On Thu, 13 Jan 2011, artur tarassow wrote:
> >
> >> I just wanted to compute the IRF and corresponding confidence
> >> intervals but I am getting the error message "out of memory".
> >
> > What does your model look like?
> >
> It is quite a big system - I admit... ;-)
>
> Please see the attached .doc file.
Thanks. Size of the model is unlikely to be an issue. More likely,
the "out of memory" message is just a wrong guess as to
what went wrong -- possibly some quirk of the restrictions leading
gretl to try to allocate a matrix with zero rows or columns.
I've made a few modifications in CVS that might help. But if the
problem persists, perhaps you could send me data and script
off-list?
Allin
14 years, 10 months
Re: [Gretl-devel] Bootstrapped IRFs for restricted VECMs
by Allin Cottrell
On Thu, 13 Jan 2011, artur tarassow wrote:
> I just wanted to compute the IRF and corresponding confidence
> intervals but I am getting the error message "out of memory".
What does your model look like?
Allin
14 years, 10 months
Bootstrapped IRFs for restricted VECMs
by Allin Cottrell
In CVS you're now able to produce bootstrapped confidence
intervals around the impulse response functions for restricted
VECMs.
If you're interested, please test and report any bad stuff!
Allin
14 years, 10 months
Re: [Gretl-devel] Error saving specific variables
by Allin Cottrell
On Tue, 4 Jan 2011, Henrique Andrade wrote:
> I need to save only two variables of my data set (PAU and PUS) but I'm
> getting a problem. Please take a look in the following script:
> *open* australia.gdt
>
> *store *"C:\Users\henrique\Desktop\dados1.gdt" PAU PUS *--gzipped *#
> Works ok!
> *store *"C:\Users\henrique\Desktop\dados 1.gdt" PAU PUS *--gzipped *#
> Works ok!
> *store *"C:\Users\henrique\Desktop\dados (1.gdt" PAU PUS *--gzipped *#
> Works ok!
>
> *store *"C:\Users\henrique\Desktop\dados 1).gdt" PAU PUS *--gzipped*
> *store *"C:\Users\henrique\Desktop\dados (1).gdt" PAU PUS *--gzipped
This all works OK in current CVS. I've recently updated the
Windows snapshot on sourceforge to reflect that.
Allin Cottrell
14 years, 11 months
Error saving specific variables
by Henrique Andrade
Dear Gretl Developers,
I need to save only two variables of my data set (PAU and PUS) but I'm
getting a problem. Please take a look in the following script:
<script>
*open* australia.gdt
*store *"C:\Users\henrique\Desktop\dados1.gdt" PAU PUS *--gzipped *#
Works ok!
*store *"C:\Users\henrique\Desktop\dados 1.gdt" PAU PUS *--gzipped *#
Works ok!
*store *"C:\Users\henrique\Desktop\dados (1.gdt" PAU PUS *--gzipped *#
Works ok!
*store *"C:\Users\henrique\Desktop\dados 1).gdt" PAU PUS *--gzipped*
*store *"C:\Users\henrique\Desktop\dados (1).gdt" PAU PUS *--gzipped
*
</script>
When I use the file names "dados1.gdt", "dados 1.gdt", and "dados (1.gdt"
I've got no problems, but if use a ")" I can't save just my two series (PAU
and PUS), instead, the entire australia.gdt file is saved.
Is this the normal behaviour?
Best,
Henrique
14 years, 11 months
