Commit Graph

11 Commits

Author SHA1 Message Date
shivers 7e1ad158c7 Added "See file COPYING." to all the copyright notices, so we can
do the open-source thing w/o larding up the source code.
1999-08-06 13:28:52 +00:00
shivers 96d9b0e068 Checking in a slew of systemic changes:
- Switching scsh over to the SRE regexp system.
    - Upgraded string & character handling
    - Switching code over to list-lib SRFI names. Some switchover to
      probably new string-lib SRFI names, though that SRFI hasn't happened,
      so it's probably a waste of time at the moment to stress out about it.
    - General switching from "reduce-" lexeme to "fold-" lexeme.
	-Olin

A per-file breakdown of changes:
    ccp.scm		New
    char-set.scm	New features
    stringlib.scm	New
    stringpack.scm	New

    awk.scm		SRE
    filemtch.scm	SRE string-lib
    fname.scm		string-lib
    fr.scm		SRE string-lib
    glob.scm		SRE
    procobj.scm		list-lib
    rdelim.scm		SRE
    re.c		SRE killed
    re.scm		SRE killed
    re1.c		SRE moved to rx/ dir
    re1.h		SRE moved to rx/ dir
    rx/...		SRE General fitting-it-in to the upgraded scsh.
    scsh-interfaces.scm
    scsh-package.scm
    scsh.scm		type-check loophole added; string-lib; list-lib
		    	reduce-port -> foldl-port

    syscalls.scm	string-lib; diked out unused MATCH-FILES code
    syscalls.c

    utilities.scm	string-lib; {any?,every?} -> {any,every}
    procobj		list-lib
1999-07-11 20:38:42 +00:00
shivers c23ba5b0cb Fixed a minor bug; made a small change using FOLDR. 1999-07-10 05:47:48 +00:00
shivers 4703ce142a There was an atomicity problem/race condition -- if a child process died
after it was forked, but before the scsh fork procedure could register
the child's procobj in the pid/procobj table, then when the SIGCHLD
signal-handler reaped the process, there would be no procobj for it.
We now lock out interrupts across the fork and register operations.
1997-04-30 20:33:26 +00:00
shivers 417b6046f8 Early autoreaping now done via SIGCHLD signal handler.
Ignored signals were done by putting a noop handler on the signal.
This could screw up passing ignores across execs.
Now, when you set an S48 interrupt handler to #f, we pass this through
to Unix.
1997-03-09 07:28:20 +00:00
shivers b64ae78af7 AUTOREAP-POLICY's optional argument wasn't. 1996-09-13 01:28:10 +00:00
shivers 1df0338fb0 I removed DEFINE-SIMPLE-ERRNO-SYSCALL because it is simply a special
case of DEFINE-ERRNO-SYSCALL. However, this change causes simple errno
syscalls to return 0 values, which blows up I/O methods used in S48's
extensible port system, so I had to hack three or four of those methods
to return a random value (#F).
1996-08-23 23:29:51 +00:00
shivers ea45fca8ef Hacked the system to use the new LET-OPT optional argument machinery,
LET-OPTIONAL, LET-OPTIONAL*, and :OPTIONAL. The new macros are
faster and easier to read than the old PARSE-OPTIONALS and OPTIONAL-ARG
procedures.
1996-04-19 18:39:14 +00:00
shivers 2ccc95ff78 filter-weak-ptr-list had args to reduce in wrong order. 1995-10-25 07:44:25 +00:00
shivers 3e52c5100b - Fixed errno-error calls to include the syscall's args. This should greatly
improve the quality of the error messages.

- Fixed file-match so that if a filter procedure raises an error condition,
  it is caught and treated as a match failure (as if the procedure returned
  #f). This means you no longer get blown out of the water by
    (file-match "." #f file-directory?)
  if the cwd contains a dangling symlink, for example.

- Added set-file-times (utime).

- Caught a bug in an unused arm of the define-errno-syscall macros (rest arg
  case).

- Perhaps one or two other minor tweaks.
1995-10-21 12:07:25 +00:00
bdc 2302efe24e *** empty log message *** 1995-10-14 03:34:21 +00:00