IRC channel logs
2024-03-10.log
back to list of logs
<lilyp>rlb: I think Python had enough troubles with the string vs unicode-string distinction to make mimicking that a somewhat foolish idea <rlb>lilyp: the problems were mostly *before* they did that, or rather, doing that was part of backing out of the mess. <rlb>And while it definitely raised my eyebrows, I've begun to wonder if it's not plausible as part of one flavor of solution. <rlb>(to the general problem) <rlb>...unless you're going to have a full set of similar algorithms to handle #u8(), i.e. the bytevector equivalent of string-split, regexps, etc., then being able to treat paths, usernames, group names, etc. as a string flavor may be handy. Not convinced, just pondering. And of course you *can* just handle them as latin-1, but that seems potentially unfriendly to non-western languages... <old>is is possible to use case-lambda with define-inlinable? <old>Ehh I see that define-inlinabled is simply a clever way to either expand or reference the implementation <old>well after an hour trying to implement it, I failed. although very close <old>sneek: later tell dthompson how would you eliminate the allocations in look-at! in (chickadee math matrix); I wonder why Guile can not do scratch allocation onto the stack?