On 21/02/06, Harry Putnam <[EMAIL PROTECTED]> wrote: > "Luke Bakken" <[EMAIL PROTECTED]> writes: > > >> I've built it before from that site so am guessing it has grown out of > >> date or become neglected. > > > > Do you really need ksh93? pdksh should work just fine in 99.9% of your > > cases. > > No, in fact I can't site a single thing it can do that pdkish doesn't. > At least not within my usage. I am just used to using it so went > looking. > > A private poster sent a pointer to a message from the > ports-cvs developers list that says... > > Remove the ast-ksh port. > Restrictive, inscrutable license; weird build system; > code doesn't inspire confidence; mostly broken. > > I agree fully with the part about `weird build system' and will defer > to there notions about code. I'm no programmer. > >
I have looked at the ksh93 port and at the code delivered from AT&T for some time. I've worked with the AT&T guys (mainly Glenn Fowler) on tracking down a bug somewhere in either our gcc or in their IO library since the variadic function macros caused wierd things to happen on i386 (in one place, an argument "got lost" between the caller and the called function), but without success. Compiling the shell with gcc 4 on i386 seemed to work, but that's not a good solution for a port. There wasn't a problem on amd64 at that time, and neither on sparc64. I haven't looked at this for a year or so now, and I'm not planning to in the near future since, as was mentioned, pdksh is good enough for most purposes. The thing I sometimes miss (i.e. that would be useful to have from time to time) in pdksh are: 1. Floating point arithmetics 2. Structs 3. The automatic manual-generation stuff Cheers, Andreas -- Andreas Kahari