On Tue, Jun 04, 2019 at 04:49:51PM -0700, Matthew DeVore wrote: > I tried to do it anyway :) I think this makes the strbuf API a bit easier to > reason about, and strbuf.h is a bit more self-documenting. WDYT? > > [...] > > +typedef int (*char_predicate)(char ch); > + > +int is_rfc3986_unreserved(char ch); > +int is_rfc3986_reserved_or_unreserved(char ch); > + > void strbuf_addstr_urlencode(struct strbuf *sb, const char *name, > - int reserved); > + char_predicate allow_unencoded_fn); Yeah, that seems reasonable. I worry slightly about adding function-call overhead to something that's processing a string character-by-character, but these strings tend to be short and infrequent. -Peff
