Re: [Sbcl-devel] Patch for run-program problem in Intel Mac OS 10.4

[Juho S. <js...@ik...>]

Nikodemus Siivola <nik...@ra...> writes:

> Sidney Markowitz <si...@si...> writes:
> 
> > I tried to minimize the amount of change from the original code. I leave it up
> > to someone who knows why the test case was written as it was to decide if the
> > use of synonym-streams is supposed to be part of the test, or if it can be
> > removed as superfluous now that the test can work without it.
> 
> Many thanks for puzzling the issues here out!
> 
> I'll merge this in a while, plus add a few comments on the motivation to
> help future explorers.

Note that :PTY T won't currently work on systems that use unix98 ptys
(for example many Linux distributions). There SB-IMPL:FIND-A-PTY needs
to look something like this:

(define-alien-routine ptsname c-string (fd int))
(define-alien-routine grantpt boolean (fd int))
(define-alien-routine unlockpt boolean (fd int))

(defun find-a-pty ()
  (let* ((master-name (coerce (format nil "/dev/ptmx") 'base-string))
         (master-fd (sb-unix:unix-open master-name
                                       sb-unix:o_rdwr
                                       #o666)))
    (when master-fd
      (grantpt master-fd)
      (unlockpt master-fd)
      (let* ((slave-name (ptsname master-fd))
             (slave-fd (sb-unix:unix-open slave-name
                                          sb-unix:o_rdwr
                                          #o666)))
        (when slave-fd
          (return-from find-a-pty
            (values master-fd
                    slave-fd
                    slave-name)))
        (sb-unix:unix-close master-fd))))
  (error "could not find a pty"))

-- 
Juho Snellman