8:06 a.m.
Hi Allen and others,
Item 1:
Today I found a repeatable crash.
Gretl crashed either
1) when I open a session file and tried to add more models to the model table and then
tried to open the table, or
2) After clearing the model table, putting a few model in it, then tried to reformat the
table after it was opened.
Attached is the crash report.
Item 2:
I have encountered a strange behavior of gretl.
Once I create a gnupot, I can zoom to the area that I wanted to see.
However, once I save the graph as an icon, any further zooming attempt will result in an
error:
"Failed to open file '/Users/fred/.gretl/gretltmp.png': No such file or
directory”.
so I closed the graph, and then open it by clicking on the icon, once more I can zoom.
Is this to be expected?
Fred
10:16 a.m.
Am 20.04.20 um 10:06 schrieb Fred Engst:
Hi Allen and others,
Item 1:
Today I found a repeatable crash.
Gretl crashed either
1) when I open a session file and tried to add more models to the model table and then
tried to open the table, or
2) After clearing the model table, putting a few model in it, then tried to reformat the
table after it was opened.
Attached is the crash report.
Is this with a very recent snapshot? Then it could be fallout from
another fix that Allin applied in that area because of another Mac problem.
Item 2:
I have encountered a strange behavior of gretl.
Once I create a gnupot, I can zoom to the area that I wanted to see.
However, once I save the graph as an icon, any further zooming attempt will result in an
error:
"Failed to open file '/Users/fred/.gretl/gretltmp.png': No such file or
directory”.
so I closed the graph, and then open it by clicking on the icon, once more I can zoom.
Is this to be expected?
Certainly not expected. Is it a new problem or have you just tried it
for the first time?
thanks
sven
2:24 p.m.
On Mon, 20 Apr 2020, Sven Schreiber wrote:
Am 20.04.20 um 10:06 schrieb Fred Engst:
> Hi Allen and others,
> Item 1:
> Today I found a repeatable crash.
> Gretl crashed either
> 1) when I open a session file and tried to add more models to the model
> table and then tried to open the table, or
> 2) After clearing the model table, putting a few model in it, then tried
> to reformat the table after it was opened.
>
> Attached is the crash report.
Is this with a very recent snapshot? Then it could be fallout from
another fix that Allin applied in that area because of another Mac problem.
No, it's a snapshot preceding the 2020b release.
Fred, thanks for the report but can you give us some more details,
please? I can see from the crash dump that there's a buffer overflow
in print_equation_stats() in the model table code, but on a simple
test I'm not able to replicate the problem and it's not obvious how
it's coming about. It would help to know what language you have
gretl running in, what's the estimator for the models in
question (OLS, or something else?), and what aspect of the model
table format you were trying to change.
If you're able to describe a walk-through example using one of the
built-in gretl datasets that would be great.
For reference, here's the test I tried:
Open Ramanathan data4-10
Via GUI estimate 3 models with ENROLL as dependent but with varying
numbers of regressors. In each case "save as icon".
Drag the 3 models onto the model table and open it for viewing.
Change various aspects of format (column headings, significant
digits, show p-values, ...). No crash.
Tried this in both English and German.
Allin
3:53 p.m.
On Mon, 20 Apr 2020, Allin Cottrell wrote:
On Mon, 20 Apr 2020, Sven Schreiber wrote:
> Am 20.04.20 um 10:06 schrieb Fred Engst:
>> Hi Allen and others,
>> Item 1:
>> Today I found a repeatable crash.
>> Gretl crashed either
>> 1) when I open a session file and tried to add more models to the model
>> table and then tried to open the table, or
>> 2) After clearing the model table, putting a few model in it, then tried
>> to reformat the table after it was opened.
>>
>> Attached is the crash report.
>
> Is this with a very recent snapshot? Then it could be fallout from
> another fix that Allin applied in that area because of another Mac problem.
[...]
Fred, thanks for the report but can you give us some more details, please? I
can see from the crash dump that there's a buffer overflow in
print_equation_stats() in the model table code, but on a simple test I'm not
able to replicate the problem and it's not obvious how it's coming about.
Actually I've spotted one possible failure mode in
print_equation_stats(), now fixed in git and snapshots. I'm not sure
if this is the active problem in your case, but it's worth trying
the current snapshot.
Allin
1832
days inactive
1832
days old
3 comments
3 participants
participants (3)
-
Allin Cottrell -
Fred Engst -
Sven Schreiber