3:25 p.m.
So far as I'm concerned, we're about ready to release. Anyone wish to
hold for some reason?
Allin
3:36 p.m.
Am 22.01.2019 um 16:25 schrieb Allin Cottrell:
So far as I'm concerned, we're about ready to release. Anyone
wish to
hold for some reason?
Is there a cutting-edge snapshot? If you want I could do some
rudimentary testing of the array-of-lists in system equations thing.
(I'm aware that you have some test scripts, but sometimes there's
something that one person does differently from another one.)
Apart from that, no reason to hold.
cheers,
sven
3:51 p.m.
On Tue, 22 Jan 2019, Sven Schreiber wrote:
Am 22.01.2019 um 16:25 schrieb Allin Cottrell:
> So far as I'm concerned, we're about ready to release. Anyone wish to hold
> for some reason?
Is there a cutting-edge snapshot? If you want I could do some rudimentary
testing of the array-of-lists in system equations thing. (I'm aware that you
have some test scripts, but sometimes there's something that one person does
differently from another one.)
I've just put up new snapshots. It probably would be good to give a
snapshot a work-out before we go ahead.
Allin
5:19 p.m.
Am 22.01.2019 um 16:51 schrieb Allin Cottrell:
I've just put up new snapshots. It probably would be good to give a
snapshot a work-out before we go ahead.
OK, here's what I think is a bug with
yesterday's snapshot (haven't
tried today's yet):
- I have a yearly dataset
- I do "save as" and enter in the dialog something like "hey" (no
extension"); the status line shows "hey"
- I go to increase/expand frequency and choose interpolate
- I get an error saying the file <path> + "hey" doesn't exist.
I haven't been able to reproduce this with an artificial dataset yet,
it's something more subtle apparently.
cheers,
sven
6:08 p.m.
On Tue, 22 Jan 2019, Sven Schreiber wrote:
Am 22.01.2019 um 16:51 schrieb Allin Cottrell:
>
> I've just put up new snapshots. It probably would be good to give a
> snapshot a work-out before we go ahead.
OK, here's what I think is a bug with yesterday's snapshot (haven't tried
today's yet):
- I have a yearly dataset
- I do "save as" and enter in the dialog something like "hey" (no
extension"); the status line shows "hey"
- I go to increase/expand frequency and choose interpolate
- I get an error saying the file <path> + "hey" doesn't exist.
I haven't been able to reproduce this with an artificial dataset yet, it's
something more subtle apparently.
I can reproduce this with the "np" sample dataset.
-------------------------------------------------------
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
-------------------------------------------------------
7:35 p.m.
On Tue, 22 Jan 2019, Sven Schreiber wrote:
Am 22.01.2019 um 16:51 schrieb Allin Cottrell:
>
> I've just put up new snapshots. It probably would be good to give a
> snapshot a work-out before we go ahead.
OK, here's what I think is a bug with yesterday's snapshot (haven't tried
today's yet):
- I have a yearly dataset
- I do "save as" and enter in the dialog something like "hey" (no
extension"); the status line shows "hey"
- I go to increase/expand frequency and choose interpolate
- I get an error saying the file <path> + "hey" doesn't exist.
I haven't been able to reproduce this with an artificial dataset yet, it's
something more subtle apparently.
I see one bug there: the filename was not getting the proper extension
added before its display in the status line. That's now fixed in git.
I'm not able to provoke an error from expand/interpolate.
Allin
8:28 p.m.
On Tue, 22 Jan 2019, Riccardo (Jack) Lucchetti wrote:
On Tue, 22 Jan 2019, Sven Schreiber wrote:
> Am 22.01.2019 um 16:51 schrieb Allin Cottrell:
>>
>> I've just put up new snapshots. It probably would be good to give a
>> snapshot a work-out before we go ahead.
> OK, here's what I think is a bug with yesterday's snapshot (haven't tried
> today's yet):
>
> - I have a yearly dataset
> - I do "save as" and enter in the dialog something like "hey" (no
> extension"); the status line shows "hey"
> - I go to increase/expand frequency and choose interpolate
> - I get an error saying the file <path> + "hey" doesn't exist.
>
> I haven't been able to reproduce this with an artificial dataset yet, it's
> something more subtle apparently.
I can reproduce this with the "np" sample dataset.
OK, I'll work on this before releasing. With the np dataset I'm
getting the (expected) error message "Missing values encountered", but
in the GUI I sometimes get "This function needs panel data" (!), which
kind of chimes with the seemingly irrelevant error message that Sven
got. Some kind of corruption seems to be going on; unfortunately
valgrind is not catching it.
Allin
2:04 p.m.
Am 22.01.2019 um 16:51 schrieb Allin Cottrell:
On Tue, 22 Jan 2019, Sven Schreiber wrote:
> Is there a cutting-edge snapshot? If you want I could do some
> rudimentary testing of the array-of-lists in system equations thing.
> (I'm aware that you have some test scripts, but sometimes there's
> something that one person does differently from another one.)
I've just put up new snapshots. It probably would be good to give a
snapshot a work-out before we go ahead.
Some quick testing of the array-of-lists way of dealing with 'equations'
went fine.
Something else popped up, though: The line "system method = sur" wasn't
accepted by the parser, it had to be "system method=sur", without blanks.
I (think I) know that a similar restriction happens with some
double-dash options like --display=output and similiar, but in general I
was a bit surprised that blanks are significant there.
Not a big deal though.
cheers,
sven
3:33 p.m.
On 22/01/2019 16:25, Allin Cottrell wrote:
So far as I'm concerned, we're about ready to release. Anyone
wish to
hold for some reason?
Hi all,
did anything change regarding the dependencies because of the libsvm
inclusion? The build fails on all fedora versions > 28
https://kojipkgs.fedoraproject.org//work/tasks/628/32230628/build.log
Thanks for any hints,
Johannes
Allin
_______________________________________________
Gretl-devel mailing list
Gretl-devel(a)lists.wfu.edu
http://lists.wfu.edu/mailman/listinfo/gretl-devel
3:34 p.m.
Am 25.01.19 um 16:33 schrieb Johannes Lips:
On 22/01/2019 16:25, Allin Cottrell wrote:
> So far as I'm concerned, we're about ready to release. Anyone wish to
> hold for some reason?
Hi all,
did anything change regarding the dependencies because of the libsvm
inclusion? The build fails on all fedora versions > 28
https://kojipkgs.fedoraproject.org//work/tasks/628/32230628/build.log
Hi Johannes,
I can just tell you that it worked fine on my ubuntu machine this morning.
Artur
7 p.m.
On Fri, 25 Jan 2019, Johannes Lips wrote:
On 22/01/2019 16:25, Allin Cottrell wrote:
> So far as I'm concerned, we're about ready to release. Anyone wish to hold
> for some reason?
Hi all,
did anything change regarding the dependencies because of the libsvm
inclusion? The build fails on all fedora versions > 28
https://kojipkgs.fedoraproject.org//work/tasks/628/32230628/build.log
Sorry, I should have mentioned: g++ is now needed for the build since
we now compile a .cpp file based on libsvm. Ideally one would
translate that file to C but it's not trivial, given the use of
classes.
Allin
1921
days inactive
1924
days old
10 comments
5 participants
participants (5)
-
Allin Cottrell -
Artur T. -
Johannes Lips -
Riccardo (Jack) Lucchetti -
Sven Schreiber