[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: About two open bug reports for the dataframe package
From: |
Julien Bect |
Subject: |
Re: About two open bug reports for the dataframe package |
Date: |
Tue, 24 Sep 2013 13:27:20 +0200 |
User-agent: |
Mozilla/5.0 (X11; Linux i686; rv:24.0) Gecko/20100101 Thunderbird/24.0 |
On 23/09/2013 09:17, CdeMills wrote:
Replied privatelly. The bug about '\b' in regexpes should be away;
Hello Pascal,
As I told you (privately), I am working on the tip of the default branch
and it's still there :
$ hg summary
parent: 183:a0d73e257b48
maint: fix hgignore regexp pattern typos
branch: default
commit: (clean)
update: 1 new changesets (update)
$ find . -name "*.m" -exec grep -Hn '\\b' {} \;
./inst/@dataframe/dataframe.m:306: %# content = cellfun(@(x)
regexp(x, '(\b|[-+\.''])[^,]*(''|\b)', 'match'),\
./inst/@dataframe/display.m:134: indi = regexp (dummy{3, 1},
'\b.*\b', 'match', 'dotexceptnewline');
./inst/@dataframe/display.m:146: indi = regexp (dummy{3,
1}, '\b.*\b', 'match', 'dotexceptnewline');
./inst/@dataframe/private/df_name2idx.m:79: subs{indi} = ['\b'
subs{indi} '\b'];
./inst/@dataframe/summary.m:44: strjust (char
(regexp (disp (sidxj.'), '\b.*', 'match', ...
Am i looking at the wrong repository ? I have submitted a patch for the
display function here : https://savannah.gnu.org/bugs/?38397, which also
improves (IMO) code clarity, will you consider applying it ?
For the rest it is a more genuine problem; fixing it internally means
that we'll need to stay in touch with upstreaml changes in functions
we have overriden.
It doesn't need to be so, as long as you rely internally on the base
skewness/kurtosis functions to do the job. I can work on a patch if
you're interested but, again, I need to know if you will consider
applying it.
Otherwise, yes, the solution that you propose here :
https://savannah.gnu.org/bugs/?32168 also does the job. But this patch
has been waiting since january 2011... Perhaps is it the opinion of
Octave's maintainers that base statistics functions (such as skewness
and kurtosis) should only care about dealing with basic numeric types
(double, single) ?
It is a fact that skewness and kurtosis are currently not working
properly with dataframes. There is a design decision to be made here,
about where the problem should be addressed : in Octave core or in the
dataframe package.
@++
Julien