From: Gary P. <pa...@in...> - 2004-08-27 13:14:50
|
----- Original Message ----- From: "John Hunter" <jdh...@ac...> > >>>>> "Gary" == Gary Pajer <pa...@in...> writes: > > Gary> I've poked around the docs and archives, for a clue to this, > Gary> but if it's there I didn't recognize it. > > Gary> version: 0.62 WinXP default backend: TkAgg interactive (with > Gary> ipython) > > Hi Gary, just to make sure we're clear. The last version of > matplotlib released was 0.61.0. Is this what you mean, or are you > using CVS? 0.61.0 (my typo) > > Gary> On computer A I have version 0.54 installed. There I can > Gary> create a plot and say savefig('plot.eps') and get a good eps > Gary> file. > > Gary> On computer B I have version 0.62 installed. There I can > Gary> create a plot and say savefig('plot.eps') with different > Gary> results: ghostscript chokes. > > There is a known bug in the ps backend on win32 that was discussed > here a couple of weeks ago. Fortunately, it has a trivial fix. In > site-packages/matplotlib/backends/backend_ps.py, in the encodeTTFasPS > function, replace the line > > font = file(fontfile) > > with > > font = file(fontfile, 'rb') > > Windows cares a lot about that binary flag. > > Please let me know if this cures what ails you, because we are getting > ready to release the next matplotlib version and I hate to release > code with known bugs. That was it. Problem solved. I remember that thread now. It didn't sink in at the time. Thanks, Gary > > Thanks! > JDH |