On Thu, 2 Aug 2012 22:13:42 +0200 Roland Mainz wrote: > On Wed, Aug 1, 2012 at 4:21 PM, Glenn Fowler <g...@research.att.com> wrote: > [snip] > > I can't commit right now because all of the implications are not clear > > it feels like a longer term issue requiring more thought > > rather than lets start coding next week
> Erm... the DocBook/XML to _ast_getopts converter... is that a disputed area ? > BTW: Is it possible to use the |wc*()| API unconditionally in AST > applications or are there any platforms left which do not have the > POSIX |wc*()| APIs ? still some ancient platforms ... > > an intriguing thought from these posts though is preprocessing the option > > string > > into another string for faster option parsing -- this could maybe > > replace the option parse caching in optget() -- the issue not attacked > > by the current cache is --long-options > Ugrh... please not another string. My suggestion was to make it a > tree... hardcoded and calculated at build time so the consumer and > jump using the pointers in the tree directly. the string could imbed a jump table -- it just could not contain the nul char _______________________________________________ ast-developers mailing list ast-developers@research.att.com https://mailman.research.att.com/mailman/listinfo/ast-developers
