Hi Vadim- On Tue, Jan 06, 2009 at 12:53:06AM +0300, Vadim Zhukov wrote: > Tested on i386. Regression fails due to error in test itself, and > I'm not a Python expert - can anyone look into it? In any case I > didn't find any regressions when using the program itself, except > GUI not very suitable for small displays.
First, I'd suggest creating a simple setup.py file (and submitting it upstream) so you can use the usual python.port.mk build/install targets. There are a few cases where we've created setup.py files for ports that don't have them; they make integration with our existing Python stuff much easier. The normal Python route will correctly take care of the #! munging so you can drop your post-patch target. If you define data_files in setup.py, you can drop your post-install target, too. I think the failing regress test is due to assumptions about locale. I haven't played much with non-ascii data, but I have seen problems with non-ascii data and locale support in Python on OpenBSD. Explicitly encoding the names makes regress run, though the output isn't very pretty. For example, the following line: print "%d %s (%s)" % (fr['number'], fr['name'], \ MetroMap.Lines[fr['line']]['name']) ...would become: print "%d %s (%s)" % (fr['number'].encode('ascii', 'replace'), fr['name'], MetroMap.Lines[fr['line']]['name'].encode('ascii', 'replace') That's not a real solution, but I don't do much non-ascii stuff. Upstream might have ideas about fixing regress output. Note that locale.getlocale() returns (None, None) on OpenBSD. Thanks! -- o--------------------------{ Will Maier }--------------------------o | web:.......http://www.lfod.us/ | email.........willma...@ml1.net | *---------------------[ BSD: Live Free or Die ]--------------------*