On Wed, 03 Dec 2008 17:37:56 +0200
Martin Vermeer <[EMAIL PROTECTED]> wrote:

> On Wed, 03 Dec 2008 16:25:24 +0100
> Abdelrazak Younes <[EMAIL PROTECTED]> wrote:
> 
> > Martin Vermeer wrote:
> > > On Wed, 03 Dec 2008 13:49:44 +0100
> > > Abdelrazak Younes <[EMAIL PROTECTED]> wrote:
> > >
> > >   
> > >> Martin Vermeer wrote:
> > >>     
> > >>> Hi,
> > >>>
> > >>> I just was involuntarily forced to test Ubuntu 8.10 (my power supply
> > >>> burned out and probably something else too, because we couldn't get
> > >>> the box back to  life again with a new power supply), and having
> > >>> problems with XFig figures in LyX.
> > >>>
> > >>> They come out 90 degs rotated and with the wrong bounding box (?)
> > >>> i.e., cut off. If I export to LaTeX manually and re-generate the
> > >>> pdftex_t and pdf files from XFig manually, things work fine.
> > >>>
> > >>> I though the problem was with Angus' heuristics in fig2pdftex.py and
> > >>> fig2pstex.py in lib/scripts. However, from the cammand line these
> > >>> work OK too.
> > >>>
> > >>> Could somebody please have a look into this? Version is the official
> > >>> one, 1.5.6. XFig is 3.2 patch level 5,
> > >>>       
> > >> Hi Martin,
> > >>
> > >> I remember fixing in bug in the 1.6 version of the script in order to 
> > >> cleanup the external inset support. This was windows related AFAIR but 
> > >> you should have a look in there. I don't think this particular script 
> > >> has changed much anyway.
> > >>
> > >> Abdel.
> > >>     
> > >
> > > Abdel,
> > >
> > > it turns out that the "wrong" PDF is produced by ImageMagick, it says at 
> > > the start
> > >
> > > /Producer (ImageMagick 6.3.7 08/21/08 Q16 http://www.imagemagick.org)
> > >
> > > The "right" PDF has nothing like that.
> > >
> > > Does that ring a bell?
> > >   
> > Not really, sorry.
> > 
> > Abdel.
> 
> Well the good news is that in SVN the bug is not present.
> 
> BTW SVN prints in the minibuffer the python commands executed, like
> fig2pdftex. 1.5.6 doesn't do that. I suppose somehow it doesn't find
> the correct script and falls back to convertDefault.py, which uses
> Imagemagick.
> 
> But as to why it doesn't find the correct script?
> 
> - Martin


Another interesting detail: if you edit the figure and change it from
Landscape to Portrait, it comes out correctly -- at least for External
Material. This may be a workaround for now, and should be documented
somewhere.

Still having trouble with Graphics. Who's going to move subfigure over
to External material so we don't need Graphics anymore?

- Martin

Reply via email to