In playing around with subversion, one of my first thoughts was to try and
setup a repository. So I ran a few of the KDE modules through cvs2svn
(kde-common, arts, kdelibs, kdemultimedia). All except for kdelibs worked
fine.
I'm using subversion from r3205, swig 1.3.15, python 2.2, FreeBSD-current
(more or less), and rcsparse.py from the viewcvs 0.9.2 tarball.
Traceback (most recent call last):
File "./cvs2svn.py", line 675, in ?
main()
File "./cvs2svn.py", line 672, in main
util.run_app(convert, ctx, start_pass=start_pass)
File "/usr/local/lib/python2.2/site-packages/svn/util.py", line 38, in run_app
return apply(func, (pool,) + args, kw)
File "./cvs2svn.py", line 623, in convert
_passes[i](ctx)
File "./cvs2svn.py", line 498, in pass1
os.path.walk(ctx.cvsroot, visit_file, (cd, p, stats))
File "/usr/local/lib/python2.2/posixpath.py", line 279, in walk
func(arg, top, names)
File "./cvs2svn.py", line 165, in visit_file
p.parse(open(pathname), cd)
File "./rcsparse.py", line 303, in parse
self.parse_rcs_tree()
File "./rcsparse.py", line 229, in parse_rcs_tree
timestamp = time.mktime(tuple(date_fields))
OverflowError: mktime argument out of range
With a few debuging prints I found that the first date that it barfs on
is: 2001.04.01.02.33.28. If I "fix" or remove this file, each subsequent
failure is on an April 1st date. Oy.
I've collected a few of the offending RCS files and put them up at
http://blarf.homeip.net:8080/rcs (if checking out a 185mb CVS repo from
cvsup.kde.org is a scary thought).
The RCS files look perfectly okay to me, but I'm not an RCS guru of any
sort. Any ideas on where this is comming from? Changing the date(s) by
hand to April 5th seems to fix everything. This is of course a less than
ideal solution.
- alex
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Sun Sep 22 05:50:20 2002