jas@ernie.Berkeley.EDU (Jim Shankland) (02/04/89)
In article <28200266@mcdurb> aglew@mcdurb.Urbana.Gould.COM writes: > >>(Quick: who's run into Unix's 10K command-line limit?) > >I `have` - ie. I have produced overlong commands using backquote >expansion. Although I seem to remember that the limit was considerably >smaller at the time (on a V7 hybrid running csh). > I ran into so many of csh's built-in limits that I almost >completely abandoned it for serious programming - although I still >used in interactively. Even BSD's 10K limit (N.B.: it's a kernel limit, not a csh limit) is nowhere near enough for me. I know it's a pain to implement, but for my money, the limit on environment size and total argv length ought to be on the order of the limit on a process's virtual address space size. (Hint to implementors: *don't* shrink the max. address space size to make this true!) Mr. Glew also mentions the 14-character filename length limit as an annoyance. True enough; but I find other limits worse. Countless utilities have line length limits, and many silently truncate lines that are too long. And recently, I was unable to run vi on a ca. 390K file on an AT&T 3B2/600 running SVR3: "sorry, file too big". AT&T is positively quaint in its notion of what constitutes a "big" file. I know it's a pain to find and root out all these limits. But they've got to go. They don't belong to 1989. (This is no longer an architecture topic. Followups, if any, to comp.unix.wizards.) Jim Shankland jas@ernie.berkeley.edu "I've been walking in a river all my life, and now my feet are wet"