> This is precisely why we should never release 1.0 with all these > problems.
1.0 will be released when the 1.0 nominated bug count is close to zero. However, there are some things that we just won't do to fix a bug during this feature-freeze period. Namely: 1) Rewrite a substantial portion of the underlying architecture 2) Introduce new dependencies 3) Break other platforms in order to fix one 4) Introduce new bugs in place of an old bug Now, we plan on using freetype2 to fix the problem mentioned in bug 1030. Using freetype2 will require us to rewrite a large portion of our font handling code and drawing code, introduce new dependencies, and potentially will break other platforms. Am I willing to do this after 1.0 is released and we start hacking on some new cool stuff? Definitely. Software will always have bugs. AbiWord 1.0 will have bugs. Should abi 1.0 have this bug? Hopefully someone will think of ways to work around this bug. I'm optimistic. Abi's current font engine on Unix works with TTFs, CJK fonts, BiDirectional glyphs, and a slew of other neat and exotic things. These bug and features are all well documented and well tested. Am I willing to introduce potential new bugs, dependencies, and potentially break other platforms to fix this bug and then perform another year of testing on this new engine before we release 1.0? Quite simply, No. Dom
signature.asc
Description: This is a digitally signed message part
