Usenet 1994 January/usenetsourcesnewsgroupsinfomagicjanuary1994.iso/sources/misc/volume29/parseargs/part03/patchlevel.h

C/C++ Source or Header | 1992-05-19 | 7.5 KB | 166 lines

/*************************************************************************
** ^FILE: patchlevel.h - current patchlevel for parseargs
**
** ^HISTORY:
**
**    01/02/92    Brad Appleton    <[e-mail]>
**    Patch11
**    - Added strndup() to strfuncs.[ch]
**    - vms_args.c:vms_parse was not setting avstr if ps_NOTCMDLINE was set.
**    - made a few other fixes for VMS. Gratuitous "thank you"s go out to
**      Mike Levin for helping me find and test these fixes (thanx Mike!).
**    - added VMSbuild.com (contributed by Luke Brennan) to build parseargs
**      on VAX/VMS.
**    - fixed unix_man.c to call indent_para() properly.
**    - added support for TCL to parseargs(1).
**    - changed the way that parseargs(1) handles escape-sequences.
**
**    12/05/91    Brad Appleton    <[e-mail]>
**    Patch10
**    - Fix bug that was introduced in last patch. Pgopen() was not
**      using $USAGE_PAGER, it was only using $PAGER.
**    - When Earl added the extra argument to indent_para(), he forgot to
**      include the argument when printing the command description. I added
**      this into *_args.c
**    - Added #ifdef POSIX_SOURCE stuff to unix_args.c to allow the use
**      of '--' instead of '+' as the long option prefix.
**
**    08/27/91    Brad Appleton    <[e-mail]>
**    08/27/91    Earl Chew    <[e-mail]>
**    Patch09
**    - Added proper support for unwritable strings
**    - strfuncs.c/strsplit():
**      Leading, and in particular, trailing space trimming may spin off
**      the end of the token string. This occurs when the string is empty,
**      or when it has trailing space.
**      Out of memory error doesn't call syserr() like the rest of the
**      routines do.
**    - xparse.c/parsecntl():
**      Documentation says that the fourth parameter for pc_ARGFLAGS and
**      pc_PARSEARGS should be either argMask_t (promoted to int) in the
**      case of writes, or argMask_t * in the case of reads. The code is
**      at odds with the documentation and requires (int *) on reads.
**      Also, should check for NULL cmd-name before passing it to strdup().
**    - vms_args.c/is_cmdline():
**      Mike Levin pointed out to me that if lib$get_foreign returns a
**      zero-length string then the proper actions are not taken. This
**      has been fixed.
**    - vms_args.c, parseargs.h, xparse.c
**      For vms_style I added a new parse-state called ps_NOTCMDLINE.
**      When this flag is SET, then vms_parse() knows for a fact that
**      the given argv[] array is NOT from the cmdline so it doesnt even
**      need to call is_cmdline(). SO Now I just have the proper routines
**      in xparse.c set this flag before calling vms_parse() (and unset
**      the flag when they are done).
**    - vms_args.c/vms_parse():
**      ARGVALGIVEN was set instead of ARGGIVEN (this was wrong).
**
**
**    08/15/91    Brad Appleton    <[e-mail]>
**    Patch08
**    - fixed some typos in the comments (SIDE-EFFECTS was misspelled
**      everywhere in the comments for the functions).
**    - fixed up external definition of ProgName to work for C++
**      (added extern "C").
**    - fixed up lparseargs() and vparseargs to free the argument vector
**      when it was finished with it.
**    - added some more comments to xparse.c and parseargs.c.
**
**    06/05/91    Brad Appleton    <[e-mail]>
**    Patch07
**    - fixed the "Intro" file to correctly show how to invoke parseargs(3)
**      (the arguments where in the wrong order).
**    - added stuff in #ifdef UNWRITABLE_STRING_LITERALS to work for machines
**      and compilers that dont allow the programmer to modify string literals.
**    - documented in the README file, all the various compile-time constants
**      that may change the behavior of parseargs.
**    - sparseargs neglected to free the argument-vector when it was done
**      using it.
**
**    05/11/91    Brad Appleton    <[e-mail]>
**    Patch06
**    - parsecntl(3) now correctly handles mispelled keywords for the
**      pc_ARGFLAGS request (it was looping forever).
**    - fixed an error in the documentation (a "NOT" was omitted).
**    - added #ifdef SVR4 stuff to unix_man.c to print System V Release 4
**      style manual pages.
**    - allowed "sname=aname" syntax in the ad_prompt field of an ARGDESC
**      in addition to the existing syntax. This new syntax allows argument
**      names to be be completely different from keyword names.
**    - changed get_name() & get_keyword to get_argname() & get_kwdname()
**      (respectively) in strfuncs.[ch]. Also sped up the algorithm to get
**      the keyword (since it no longer needs malloc).
**
**    04/25/91    Brad Appleton    <[e-mail]>
**    Patch05
**    - documented support for zsh in parseargs(1)
**    - added missing call to pclose() in pgopen.c:pgclose()
**    - added zsh to the list of shells supported by parseargs(1)
**      (I have yet to document this in the man-pages)
**    - changed the shell_info structure used by parseargs(1) to
**      represent different shells. It should now be much simpler
**      to add new shells to parseargs.c
**    - fixed a bug in strfuncs.c:basename() for VMS systems
**    - added more comments to explain whats going on in each <os>_args.c
**    - updated the README file to correspond with the patchlevel(s)
**    - updated dependencies in doc/Makefile
**
**    04/10/91    Brad Appleton    <[e-mail]>
**    Patch04
**    - fixed problem with prompting in xparse.c:prompt_user()
**    - under VMS, getenv was defined to be 'getsymbol', it should
**      be 'get_symbol' (note the underscore).
**    - now includes <errno.h> to declare errno
**    - parseargs no longer includes <fcntl.h>
**    - fix bug in vms_args.c printing '[]' for required args and vis-a-vis
**    - fixed bug in argInput and argOutput in argtype.c for VMS
**    - fixed bug in <os>_args.c where ARGVALGIVEN was improperly being
**      cleared if the argument was a list or a vector. This was causing
**      parseargs(1) to screw up on optional lists.
**    - fixed bug in strfuncs.c:strnicmp() (it was ignoring the length)
**    - changed strfuncs.[ch] so that my version of strdup is always used
**      (my version exits if malloc failed, others return NULL).
**    - Added some changes for SUN and VAX
**    - Added -1 option to parseargs(1)
**    - Unhid the -# option to parseargs(1)
**    - More documentation fixes
**
**    04/04/91    Brad Appleton    <[e-mail]>
**    Patch04 (continued)
**    - fixed bug in xparse.c with PARSECNTL='Prompt' being ignored.
**    - fixed newline printing error in perl-scripts for parseargs(1)
**    - added -C option to parseargs(1) and to the docs
**    - added "patchlevel.h" to distribution
**    - added "what" string with version and patch level for parseargs(1)
**    - added hidden -# option to parseargs(1)
**    - add "doc/effects.inc" to distribution
**    - fix docs
**    - fix test.rc for Plan 9 shell
**    - fix parseargs(1) quoting for perl
**
**    03/22/91    Brad Appleton    <[e-mail]>
**    Patch03
**    - ascii docs fix
**
**    03/21/91    Brad Appleton    <[e-mail]>
**    Patch02
**    - documentation fix
**
**    03/12/91    Brad Appleton    <[e-mail]>
**    Patch01
**    - documentation fix
**
**    03/02/91    Brad Appleton    <[e-mail]>    Created
***^^********************************************************************/

#define  VERSION     2
#define  REVISION    0
#define  PATCHLEVEL  11

#ifdef __STDC__
   static const char
#else
   static char
#endif
   _Ident[] = "@(#)parseargs  2.0  patchlevel 11";