Re: __plt_get_axis_arg__

[Rik]

On 07/22/2013 01:03 PM, John W. Eaton wrote:
> On 07/22/2013 03:12 PM, Dmitri A. Sergatskov wrote:
>> BTW,
>>
>> I tried to rebuild
>> changeset:   17034:8d84dc5f5b5d
>> tag:         tip
>> user:        Rik <address@hidden <mailto:address@hidden>>
>> date:        Sun Jul 21 20:16:03 2013 -0700
>> summary:     acinclude.m4: Use m4 quote '[' rather than '"' in
>> AC_MSG_ERROR.
>>
>> and I got error:
>>
>> mv testfun.texi-t testfun.texi
>> perl ./munge-texi.pl <http://munge-texi.pl> ../..
>> ../../scripts/DOCSTRINGS ../../libinterp/DOCSTRINGS < tips.txi >
>> tips.texi-t
>> mv tips.texi-t tips.texi
>> perl ./munge-texi.pl <http://munge-texi.pl> ../..
>> ../../scripts/DOCSTRINGS ../../libinterp/DOCSTRINGS < var.txi > var.texi-t
>> mv var.texi-t var.texi
>> perl ./munge-texi.pl <http://munge-texi.pl> ../..
>> ../../scripts/DOCSTRINGS ../../libinterp/DOCSTRINGS < vectorize.txi >
>> vectorize.texi-t
>> mv vectorize.texi-t vectorize.texi
>> ../../run-octave -f -q -H -p . --eval "geometryimages ('voronoi', 'eps');"
>> ../../run-octave -f -q -H -p . --eval "geometryimages ('triplot', 'eps');"
>> ../../run-octave -f -q -H -p . --eval "geometryimages ('griddata', 'eps');"
>> ../../run-octave -f -q -H -p . --eval "geometryimages ('convhull', 'eps');"
>> ../../run-octave -f -q -H -p . --eval "geometryimages ('delaunay', 'eps');"
>> ../../run-octave -f -q -H -p . --eval "geometryimages ('inpolygon',
>> 'eps');"
>> ../../run-octave -f -q -H -p . --eval "interpimages ('interpft', 'eps');"
>> ../../run-octave -f -q -H -p . --eval "interpimages ('interpn', 'eps');"
>> ../../run-octave -f -q -H -p . --eval "interpimages ('interpderiv1',
>> 'eps');"
>> ../../run-octave -f -q -H -p . --eval "interpimages ('interpderiv2',
>> 'eps');"
>> ../../run-octave -f -q -H -p . --eval "plotimages ('plot', 'eps');"
>> ../../run-octave -f -q -H -p . --eval "plotimages ('hist', 'eps');"
>> error: patch: expecting first argument to be axes handle
>> error: called from:
>> error:   /d2/home/dima/src/octave/scripts/plot/__plt_get_axis_arg__.m at
>> line 52, column 7
>> error:   /d2/home/dima/src/octave/scripts/plot/patch.m at line 48,
>> column 15
>> error:   /d2/home/dima/src/octave/scripts/plot/private/__bar__.m at line
>> 199, column 11
>> error:   /d2/home/dima/src/octave/scripts/plot/private/__bar__.m at line
>> 161, column 11
>> error:   /d2/home/dima/src/octave/scripts/plot/bar.m at line 95, column 16
>> error:   /d2/home/dima/src/octave/scripts/plot/hist.m at line 166, column 5
>> error:   /d2/home/dima/src/octave/doc/interpreter/plotimages.m at line
>> 40, column 5
>> make[3]: *** [hist.eps] Error 1
>> make[3]: Leaving directory `/d2/home/dima/src/octave/doc/interpreter'
>> make[2]: *** [all-recursive] Error 1
>> make[2]: Leaving directory `/d2/home/dima/src/octave/doc'
>> make[1]: *** [all-recursive] Error 1
>> make[1]: Leaving directory `/d2/home/dima/src/octave'
>> make: *** [all] Error 2
>
> Undoing this change
>
>   # HG changeset patch
>   # User Rik <address@hidden>
>   # Date 1374370945 25200
>   # Node ID 75d115e329b78b9daf029801011cd10c3521c1ff
>   # Parent  c358c6e7416bde5926851c14c9de7b22dae0e6ca
>   __plt_get_axis_arg__.m: Correctly handle 'parent' property when parent
> is hggroup.
>
>   * scripts/plot/__plt_get_axis_arg__.m: Don't delete 'parent'/value pair
>   if it represents an hggroup.  Downstream plot function will use this
>   property.
>
> allows the build to succeed for me.  In the case that fails with this
> change, the handle that is passed to __plt_get_axis_arg__ refers to an
> hggroup object, so I guess the logic is not quite right yet.
>
> I'm no longer sure what the intent is in __plt_get_axis_arg__ so I
> don't feel confident that I can fix this problem without breaking
> something else.
>
> Given the complexity of __plt_get_axis_arg__ now, I'm wondering
> whether we have a clear understanding of how creating new or using
> existing handles is supposed to work.
7/22/13

John et al.,

Just an FYI, Panxto and I are working together on completely overhauling
the __plt_get_axis_arg__ and newplot functionality.  This is something that
started a few days ago when we realized it wasn't doing the right thing in
several situations.  I've re-written __plt_get_axis_arg__ and newplot and
Panxto is working on transitioning the plot routines to use the new
functions.  I'd expect a few days before we have everything in place and
pushed to Mercurial.

--Rik