Le jeudi 02 novembre 2006 à 00:08 +0100, Waldek Hebisch a écrit : [...] > No very strong reason. However: > > 1) 'axiom-probe-file' is intended to be good replacement for 'probe-file'. > More precisely, Axiom uses directories to implement what Microsoft > would call "structured storage" and old IBM systems called "partitioned > data sets". Such directories from Axiom point of view are equivalent > to files.
I have no knowledge on this. > 2) I did not check all uses of 'probe-name', so it was not clear if > change in 'probe-name' is apropriate. As far as I know probe-name is only used in nlib.lisp.pamphlet. > 3) 'probe-name' is doing a lot of extra work (calls 'probe-file' first, > then calls 'namestring' Yes, and namestring is not necessary. > Concerning erase: I would prefer to limit use of 'directoryp': > something which has name endig in 'p' should return a boolean. > What 'directoryp' returns is unexepected, so wide use of 'directoryp' > would make interpreter code hareder to read. Yes, not the lisp convention. Anyway these things need to be fixed, no matter for me how this will be done. Thank you so much (as Martin implicitly says :-). I am, as Martin, a Hyperdoc adept. Greg _______________________________________________ Axiom-developer mailing list Axiom-developer@nongnu.org http://lists.nongnu.org/mailman/listinfo/axiom-developer
