Gabriel Dos Reis <[EMAIL PROTECTED]> writes: > On Sun, 1 Jul 2007, Stephen Wilson wrote: > > | Am I correct in assuming the generated .lisp files are still easily > | inspected and modified, and share their own component of the > | dependency graph for make? > > If I understand your question correctly, the answer is yes.
Ok. Just for the sake of clarity, I meant that one can modify the intermediate lisp, and have those changes picked up by the build process without them being overwritten by a fesh tangling of a pamphlet. > > | I guess I would like a tad bit more info than saying the current > | scheme is bad. It tells me nothing about the benifits of your change. > > In my experience, the current scheme has led to more problems, complications > than any perceived benefits. It has many pitfalls when trying to > port to "non-conventional" filesystems, e.g. what you'd find on windows > and unixes with mounted filesystems. I believe wh-sandbox was > branched before I finished making build-improvements fully workable > on windows. I believe people have ecncountered amny variations of the > problems there. On balance, I see the current steup having more > problems than benefits it gives us on daily basis. Sure. Still unclear about how that relates to getting rid of int/, as its just another directory. Of course, Im the first to admit I have near-zero experience with "non-conventional" file systems, particularly windows. Regardless, my main concern is the current property of int/ w.r.t re-make, as mentioned above. Many thanks for taking the time to respond to my questions! Take care, Steve _______________________________________________ Axiom-developer mailing list [email protected] http://lists.nongnu.org/mailman/listinfo/axiom-developer
