7:09 p.m.
Thanks, Juan Carlos. So the problem is confirmed, at least on Windows. I'll
investigate before long.
Allin
On Thu, Jun 23, 2022 at 3:06 PM ESTEVEZ NUÑEZ JUAN CARLOS <
ecadrian.estevez(a)usc.es> wrote:
Here, on Windows 10
gretl 2022b-git (x86_64, gtk2, 2022-06-22)
dbnomics 0.43 2022-01-12
Klaus script output:
------------------------------
_______________________________________________
Gretl-users mailing list -- gretl-users(a)gretlml.univpm.it
To unsubscribe send an email to gretl-users-leave(a)gretlml.univpm.it
Website:
https://gretlml.univpm.it/postorius/lists/gretl-users.gretlml.univpm.it/
_______________________________________________
Gretl-users mailing list -- gretl-users(a)gretlml.univpm.it
To unsubscribe send an email to gretl-users-leave(a)gretlml.univpm.it
Website:
https://gretlml.univpm.it/postorius/lists/gretl-users.gretlml.univpm.it/
8:55 a.m.
New subject: Bundl request from dbnomics errors
Am 23.06.2022 um 21:09 schrieb Cottrell, Allin:
Thanks, Juan Carlos. So the problem is confirmed, at least on
Windows.
I'll investigate before long.
Since it ran OK for me, some other ingredients are needed, I guess.
On this machine I am using the decimal point (i.e., _not_ using the
local German settings, which would be the comma), and my username does
not contain blanks or special characters.
Klaus and Juan Carlos, what about you?
thanks
sven
10:48 a.m.
New subject: Bundl request from dbnomics errors
My yesterday's test was made in a PC at University, but today we´ve a holiday in this
part of Spain,
so I can´t ensure (I believe was local Galician).
But now I tested at home with Ubuntu 20.04 and get the results of the attached image
(all ok, on Linux x86_64, gtk3, with locale Galician).
As at home I had also Windows 10, I tested again and have no problems with gretl
2022b-git
(x86_64, gtk2, with locale Galician) but with 2022/06/13 update.
So I installed the 2022/06/22 version and the problem appears again. So it seems there
was
a change between these updates that affects the way Klaus script works.
________________________________
De: Sven Schreiber <svetosch(a)gmx.net>
Enviado: venres, 24 de xuño de 2022 10:55
Para: gretl-users(a)gretlml.univpm.it <gretl-users(a)gretlml.univpm.it>
Asunto: [Gretl-users] Re: Bundl request from dbnomics errors
Am 23.06.2022 um 21:09 schrieb Cottrell, Allin:
Thanks, Juan Carlos. So the problem is confirmed, at least on
Windows.
I'll investigate before long.
Since it ran OK for me, some other ingredients are needed, I guess.
On this machine I am using the decimal point (i.e., _not_ using the
local German settings, which would be the comma), and my username does
not contain blanks or special characters.
Klaus and Juan Carlos, what about you?
thanks
sven
_______________________________________________
Gretl-users mailing list -- gretl-users(a)gretlml.univpm.it
To unsubscribe send an email to gretl-users-leave(a)gretlml.univpm.it
Website: https://gretlml.univpm.it/postorius/lists/gretl-users.gretlml.univpm.it/
3:48 p.m.
New subject: Bundl request from dbnomics errors
Am 24.06.2022 um 13:19 schrieb Klaus Hasenbach:
Thank you, Sven!
Actually I ran it on Windows 2022 Server with US-language (points
instead commas).
But I tried it also on my lokal machine with the same US-language
settings and it is just the same.
The error does not occure always.
Please try to run the script twice or three times than it is more
likeli to produce the error.
OK, I got the error now. Here's my preliminary theory, based on the
following error message from gretl together with the contents of my
dotdir (C:/...blabla..../AppData/Roaming/gretl):
File exists
*** error in function dbnomics_bundle_get_data, line 33
outfile --buffer=tracestr
Apparently on Windows the "outfile --buffer" variant is working via
tempfiles. I'm seeing a lot of files in the dotdir on the pattern of
"outfile.3419N1", and indeed most of them contain the text "trying
append..." which comes from the outfile block.
Then gretl complains "file exists", and so I'm guessing that the
mechanism for inventing unique temp file names is flawed and doesn't
really ensure unique names. Hence gretl chokes.
Carrying on with the speculation, why doesn't it happen on Linux? I see
two possibilities: Either it's not using tempfiles at all for the
"--buffer" option, (and is directly working via memory instead), or the
mechanism to produce unique filenames works better there.
The bug will surely be quickly fixed, but the only workaround I can see
right now is to retrieve less variables in one go, and/or manually clean
up the dotdir in between.
cheers
sven
7:40 a.m.
New subject: Bundl request from dbnomics errors
On Fri, 24 Jun 2022, Sven Schreiber wrote:
OK, I got the error now. Here's my preliminary theory, based on
the
following error message from gretl together with the contents of my
dotdir (C:/...blabla..../AppData/Roaming/gretl):
File exists
*** error in function dbnomics_bundle_get_data, line 33
> outfile --buffer=tracestr
Apparently on Windows the "outfile --buffer" variant is working via
tempfiles. I'm seeing a lot of files in the dotdir on the pattern of
"outfile.3419N1", and indeed most of them contain the text "trying
append..." which comes from the outfile block.
This happens on Linux too. The difference is that those temporary files
should, in theory, be deleted once you exit gretl. Apparently, for some
reason this is not happening on windows, so some sort of name collision
occurs.
I have no way of checking atm (and besides, I don't have a windows
machine), but if this conjecture is true, it follows that (a) cleaning
these files "by hand" should be an effective remedy for the time being and
(b) this bug will be easy to fix in time for the next release, which is
due soon.
-------------------------------------------------------
Riccardo (Jack) Lucchetti
Dipartimento di Scienze Economiche e Sociali (DiSES)
Università Politecnica delle Marche
(formerly known as Università di Ancona)
r.lucchetti(a)univpm.it
http://www2.econ.univpm.it/servizi/hpp/lucchetti
-------------------------------------------------------
12:06 p.m.
New subject: Bundl request from dbnomics errors
For test purposes I deleted all outfile.xxx as suggested (most of them
exept one file are deleted anyway when I exit gretl im MS-Windows).
Either I will get the "common outfile error" aigain ore following new
error occures:
sprintf: format string is missing
*** error in function sanitize_url, line 1
string dqu = sprintf("\"")
called by function build_parms
called by function dbnomics_get_multiple
Klaus
Am 25.06.2022 um 09:40 schrieb Riccardo (Jack) Lucchetti:
On Fri, 24 Jun 2022, Sven Schreiber wrote:
> OK, I got the error now. Here's my preliminary theory, based on the
> following error message from gretl together with the contents of my
> dotdir (C:/...blabla..../AppData/Roaming/gretl):
>
> File exists
> *** error in function dbnomics_bundle_get_data, line 33
>> outfile --buffer=tracestr
>
> Apparently on Windows the "outfile --buffer" variant is working via
> tempfiles. I'm seeing a lot of files in the dotdir on the pattern of
> "outfile.3419N1", and indeed most of them contain the text "trying
> append..." which comes from the outfile block.
This happens on Linux too. The difference is that those temporary
files should, in theory, be deleted once you exit gretl. Apparently,
for some reason this is not happening on windows, so some sort of name
collision occurs.
I have no way of checking atm (and besides, I don't have a windows
machine), but if this conjecture is true, it follows that (a) cleaning
these files "by hand" should be an effective remedy for the time being
and (b) this bug will be easy to fix in time for the next release,
which is due soon.
-------------------------------------------------------
Riccardo (Jack) Lucchetti
Dipartimento di Scienze Economiche e Sociali (DiSES)
Università Politecnica delle Marche
(formerly known as Università di Ancona)
r.lucchetti(a)univpm.it
http://www2.econ.univpm.it/servizi/hpp/lucchetti
-------------------------------------------------------
_______________________________________________
Gretl-users mailing list -- gretl-users(a)gretlml.univpm.it
To unsubscribe send an email to gretl-users-leave(a)gretlml.univpm.it
Website: https://gretlml.univpm.it/postorius/lists/gretl-users.gretlml.univpm.it/
2:23 p.m.
New subject: Bundl request from dbnomics errors
Am 25.06.2022 um 14:06 schrieb klaus.hasenbach(a)web.de:
For test purposes I deleted all outfile.xxx as suggested (most of
them
exept one file are deleted anyway when I exit gretl im MS-Windows).
Either I will get the "common outfile error" aigain
Yes, I'm afraid Jack's assessment isn't entirely correct. The problem is
not that something is wrong when gretl exits, but something doesn't work
as expected during the ongoing gretl session.
ore following new
error occures:
sprintf: format string is missing
*** error in function sanitize_url, line 1
string dqu = sprintf("\"")
OK, this is completely weird (and not reproducible for me when I try
these lines in isolation). My best bet is that after the other error
gretl is deeply confused until it is restarted. Of course this shouldn't
happen, either, but currently isn't so important, perhaps.
Am 25.06.2022 um 09:40 schrieb Riccardo (Jack) Lucchetti:
> This happens on Linux too. The difference is that those
temporary
> files should, in theory, be deleted once you exit gretl. Apparently,
> for some reason this is not happening on windows, so some sort of name
> collision occurs.
See above, IMHO the collision doesn't have to do with what happens when
gretl exits.
I did some research on what internal functions are called, and it seems
to go down via gretl's gretl_mktemp (in lib/src/gretl_paths.c) to Glib's
g_mkstemp on Windows, and apparently then to Glib's get_tmp_file.
(For example here:
https://github.com/GNOME/glib/blob/glib-2-72/glib/gfileutils.c starting
at line 1462, assuming this is the right branch for the glib 2.72
version for gretl on Windows?)
OK, but what does it all mean? It's hard to believe that a pretty core
glib function is the root of the problem, isn't it?
cheers
sven
8:20 p.m.
New subject: Bundl request from dbnomics errors
On Sat, 25 Jun 2022, Sven Schreiber wrote:
I did some research on what internal functions are called, and it
seems
to go down via gretl's gretl_mktemp (in lib/src/gretl_paths.c) to Glib's
g_mkstemp on Windows, and apparently then to Glib's get_tmp_file.
Yes, there appears to be a race condition in g_mkstemp on Windows.
I'll see what I can find out about that.
Allin
8:24 p.m.
New subject: Bundl request from dbnomics errors
On Sat, 25 Jun 2022, Sven Schreiber wrote:
Am 25.06.2022 um 14:06 schrieb klaus.hasenbach(a)web.de:
>
> sprintf: format string is missing
> *** error in function sanitize_url, line 1
>
> string dqu = sprintf("\"")
OK, this is completely weird (and not reproducible for me when I try
these lines in isolation). My best bet is that after the other error
gretl is deeply confused until it is restarted.
I cannot make sense of the failing sprintf() call. I could be wrong,
but I don't think the "deeply confused until restarted" hypothesis
make sense here. However, right now I don't have an alternative
hypothesis.
Allin
7:46 p.m.
New subject: Bundl request from dbnomics errors
On Sat, Jun 25, 2022 at 8:07 AM klaus.hasenbach(a)web.de
<klaus.hasenbach(a)web.de> wrote:
For test purposes I deleted all outfile.xxx as suggested (most of them except one file
are deleted anyway when I exit gretl im MS-Windows). Either I will get the "common
outfile error" again or following new error occurs:
sprintf: format string is missing
*** error in function sanitize_url, line 1
string dqu = sprintf("\"")
called by function build_parms
called by function dbnomics_get_multiple
In my testing on Windows 10 I could reproduce the "common outfile
error" (as you say, not every time but sooner or later). After the
changes I've made today, I'm not seeing errors on Windows any more, so
I think the issue may be fixed. But since I'm not 100% sure that I
found the effective cause I'd welcome confirmation or disconfirmation!
(Please use the latest snapshot.)
Allin
9:06 p.m.
New subject: Bundl request from dbnomics errors
Am 26.06.2022 um 21:46 schrieb Cottrell, Allin:
In my testing on Windows 10 I could reproduce the "common
outfile
error" (as you say, not every time but sooner or later). After the
changes I've made today, I'm not seeing errors on Windows any more, so
I think the issue may be fixed. But since I'm not 100% sure that I
found the effective cause I'd welcome confirmation or disconfirmation!
(Please use the latest snapshot.)
I ran Klaus's example a couple of times with the new snapshot and never
got an error.
I'm still puzzled by your suggestion of a "race condition", though -
isn't that only a problem for parallel code execution?
thanks
sven
9:43 p.m.
New subject: Bundl request from dbnomics errors
Great !! With the latest snapshot gretl and dbnomics works in MS Windows
perfekt. I can confirm the test of Sven. I testet it for even more then
3k time- series in one run, again and again.
(Just a an other question of a dum user. I fetch monthly data into a
monthly workfile und quarterly data into a quarterly workfile. Otherwise
I will get a frequency error.
Afterwards I put all together into one monthly workfile with the
command "join" . Unfortunately the description labels from the quarterly
workfile will be lost. Is there a way to come araound this?)
Thanks
Klaus
Am 26.06.2022 um 23:06 schrieb Sven Schreiber:
Am 26.06.2022 um 21:46 schrieb Cottrell, Allin:
> In my testing on Windows 10 I could reproduce the "common outfile
> error" (as you say, not every time but sooner or later). After the
> changes I've made today, I'm not seeing errors on Windows any more, so
> I think the issue may be fixed. But since I'm not 100% sure that I
> found the effective cause I'd welcome confirmation or disconfirmation!
> (Please use the latest snapshot.)
I ran Klaus's example a couple of times with the new snapshot and never
got an error.
I'm still puzzled by your suggestion of a "race condition", though -
isn't that only a problem for parallel code execution?
thanks
sven
_______________________________________________
Gretl-users mailing list -- gretl-users(a)gretlml.univpm.it
To unsubscribe send an email to gretl-users-leave(a)gretlml.univpm.it
Website:
https://gretlml.univpm.it/postorius/lists/gretl-users.gretlml.univpm.it/
11:28 p.m.
New subject: Bundl request from dbnomics errors
On Sun, Jun 26, 2022 at 5:06 PM Sven Schreiber <svetosch(a)gmx.net> wrote:
Am 26.06.2022 um 21:46 schrieb Cottrell, Allin:
> In my testing on Windows 10 I could reproduce the "common outfile
> error" (as you say, not every time but sooner or later). After the
> changes I've made today, I'm not seeing errors on Windows any more, so
> I think the issue may be fixed. But since I'm not 100% sure that I
> found the effective cause I'd welcome confirmation or disconfirmation!
> (Please use the latest snapshot.)
I ran Klaus's example a couple of times with the new snapshot and never
got an error.
I'm still puzzled by your suggestion of a "race condition", though -
isn't that only a problem for parallel code execution?
Maybe I was wrong about a "race condition", but it's not exclusively
related to multi-threading.
https://en.wikipedia.org/wiki/Race_condition
Allin
9:19 p.m.
New subject: Bundl request from dbnomics errors
I tested Klaus script on Win10, with gretl_install-64.exe and the gretl-win64.zip
installations
and both go fine now.
________________________________
De: Cottrell, Allin <cottrell(a)wfu.edu>
Enviado: domingo, 26 de xuño de 2022 21:46
Para: Gretl list <gretl-users(a)gretlml.univpm.it>
Asunto: [Gretl-users] Re: Bundl request from dbnomics errors
On Sat, Jun 25, 2022 at 8:07 AM klaus.hasenbach(a)web.de
<klaus.hasenbach(a)web.de> wrote:
For test purposes I deleted all outfile.xxx as suggested (most of them except one file
are deleted anyway when I exit gretl im MS-Windows). Either I will get the "common
outfile error" again or following new error occurs:
sprintf: format string is missing
*** error in function sanitize_url, line 1
string dqu = sprintf("\"")
called by function build_parms
called by function dbnomics_get_multiple
In my testing on Windows 10 I could reproduce the "common outfile
error" (as you say, not every time but sooner or later). After the
changes I've made today, I'm not seeing errors on Windows any more, so
I think the issue may be fixed. But since I'm not 100% sure that I
found the effective cause I'd welcome confirmation or disconfirmation!
(Please use the latest snapshot.)
Allin
_______________________________________________
Gretl-users mailing list -- gretl-users(a)gretlml.univpm.it
To unsubscribe send an email to gretl-users-leave(a)gretlml.univpm.it
Website: https://gretlml.univpm.it/postorius/lists/gretl-users.gretlml.univpm.it/
663
days inactive
666
days old
14 comments
7 participants
participants (7)
-
Allin Cottrell -
Cottrell, Allin
-
ESTEVEZ NUÑEZ JUAN CARLOS -
Klaus Hasenbach -
klaus.hasenbach＠web.de
-
Riccardo (Jack) Lucchetti -
Sven Schreiber