picrin
picrin copied to clipboard
picrin-scheme
(scheme char) is missing
@KeenS
To cover the all the parts (scheme char) specifies, quite a few static tables with a large number of initial elements (as big as unicode, in short) are necessary. I do think it is too big to be a part of benz core. Instead I will propose that we will implement this module as a contribution library just like (picrin *) libraries. You may feel the asymmetry that other (scheme *) libraries are all included in benz and not to (scheme char), but the huge block of (scheme char) is not one which benz intends to have.
I'd love to work on this unless anyone has started to do yet.
My current design based on the comment of @wasabiz:
char.chandles ASCII only, so(picrin base)also does.r7rsnitro contains an implementation of character-related procedures in(scheme base)and(scheme char)that is, at least, UTF-8 aware.- r7rs nitro adds
full-unicodefeature. (unless we decide not to use Unicode version 6.0 or later :dizzy_face:) - Tools to convert from Unicode Character Database to internal tables will be added to
etc/.
What do you think?
Great! I can't wait for it :) I love that idea to implement unicode features as an pluggable nitro. Picrin's major policy is to achieve high compatibility with various kinds of environment even without libc, and I think UTF-8/Unicode is too big to be part of the core (only supporting utf-8 should be small enough, though). BTW, I guess you probably must be annoyed using both your repository and picrin's. I have missed chances so far but I give you permission to get access to picrin repo. Is it ok for you? Even after you got the right development flow would not change. Every big modification against the core should be requested via pull request (@keens has followed that rules, too). For small bugfixes you can push directly to master anytime!
2015年6月26日(金) 14:41 OGINO Masanori [email protected]:
I'd love to work on this unless anyone has started to do yet.
My current design based on the comment of @wasabiz https://github.com/wasabiz:
- char.c handles ASCII only, so (picrin base) also does.
- r7rs nitro contains an implementation of character-related procedures in (scheme base) and (scheme char) that is, at least, UTF-8 aware.
- r7rs nitro adds full-unicode feature. (unless we decide not to use Unicode version 6.0 or later [image: :dizzy_face:])
- Tools to convert from Unicode Character Database to internal tables will be added to etc/.
What do you think?
— Reply to this email directly or view it on GitHub https://github.com/picrin-scheme/picrin/issues/211#issuecomment-115526610 .
Of course. I'm glad to become a collaborator.
BTW, when my modification is related both benz and picrin, what should I do?
Sending a PR to benz and another to picrin? Sending a PR including diffs on extlib/benz/ to picrin?
@omasanori Since benz is a subtree of picrin, PRing only to picrin will do enough. The one who merged will push the patch to benz. As such, be careful not to mix commit to benz and commit to picrin. Keep commit log clean.