> Per Andersson escreveu isso aí: >> Please see the attached patch and try if it works.
> Heh, except that you missed other usages of the same pattern. Andreas, > please try the attached patch. > As Per explained, the problem is that in Ruby 1.8, a string is an array > of bytes, so str[0] returns a number (the byte at position 0). Ruby 1.9 > is fully encoding-aware, so a string is an array of characters and > str[0] returns a string that is the first *character* in str. > What I did was replacing the occurrences of str[0] with str.bytes.first > to explicitly request the first *byte* in str. > The resulting output when run with Ruby 1.9 is pretty much the same as > the original version generates when run by Ruby 1.8, *except* for the > ordering between lowercase and lowercase letters. Maybe that is due to > some other detail, but understanding that script completely is too much > for me. :) > If that's not acceptable, you can also (at least for Wheezy) run the > script from debian/rules with `ruby1.8` instead of `ruby` and build > depend on ruby1.8 explicitly. Hi, Andreas and the other people! I see you have already patched the script. Sorry for my later response: I'm very busy these days: I'm trying to change my job. I will able to do something in week or two. Please don't drop packages in the time :) -- . ''`. Dmitry E. Oboukhov : :’ : email: un...@debian.org jabber://un...@uvw.ru `. `~’ GPGKey: 1024D / F8E26537 2006-11-21 `- 1B23 D4F8 8EC0 D902 0555 E438 AB8C 00CF F8E2 6537
signature.asc
Description: Digital signature
