Programmer 7500/MAX_PROGRAMMERS.iso/PASCAL/T_REF.ZIP/T-REF.UPD

Text File | 1987-01-15 | 7.3 KB | 142 lines

At some time during the development process, one must say "Enough"! and there-
fore set the marker for the first version.  The program on this disk is just
that.  Of course, as Beta testing continues, problems are found and new ideas
are discussed to keep the development phase "alive".  We are dedicated to the
continued development of T-Ref.  The following are SOME of the features you can
expect to see soon.  Target dates have not been set beyond version 1.1, but we
are shooting for approximately three month intervals.  This means Version 2.0
(described below) will be ready in the Fall 1987 time frame.  Known bugs
(fixes) have priority but you can be sure we will "sweeten the pot" with
worthwhile enhancements with each new version.

The following changes are being considered for T-Ref, Version 1.1, to be com-
pleted in Mid-March 87:

   a) All known bugs with Version 1.0 will be corrected.  So far, these are:

      1. There is a bug in reading include files without a file extention
         specified.  If you have any include file directives without file
         extentions, you will have to put them in.
      2. Under certain conditions, if a Pascal source file begins with a blank
         line, T-Ref will loop (catch) on the first line and produce an
         infinite number of lines on the output source.  For Version 1.0, make
         sure there are no blank lines at the head of the main or include files
         that are to be read.  Eliminating them will prevent this from happen-
         ing.
      3. Version 1.0 does not work properly with EGA monitors and with certain
         systems not completely IBM compatible.
      4. During some screen refreshes, "snow" may appear on some CGA monitors
         due to memory contentions between the CPU and the monitor board.
      5. Multi-leveled (3 or more) comments (i.e., { (* { } *) } ) will confuse
         the parser.
      6. The program will try to parse a program even when both the source
         lister and the cross reference lister are "OFF".  This is not really
         a problem except there will be no output.

   b) Certain modifications will be made to make the program more user
      friendly.  In particular:

      1. The editor used to created headers for the printed code will be
         enhanced.
      2. The user will be able to select (and save) the foreground and back-
         ground colors he/she wishes to use with the program.
      3. Screen writes, especially at program start-up, will be much faster.

   c) Up to 10 different "sets" of parameters will be able to be saved.
      Presently, only one set is allowed.  The user will select the appro-
      priate "set" with a "key" of from 1 to 10 characters in length.  The
      parameter set may also be designated on the command line on program
      start-up.

   d) To facilitate input source file selection, a new feature will be
      incorporated.  A function key will call up a window that asks for drive,
      directory, and file mask of the input file.  File mask may include the
      wildcard characters "*" and "?" as in DOS.  All matching files are
      brought up and the user selects the desired file with the arrow keys.
      The selected file is then automatically loaded into the "Active Input
      File" parameter.

   e) We are looking at methods to speed up the parsing process by a minimum
      of 10 to 20 percent.

   f) In addition to the four lettering styles, additional switches will be
      added to double strike and/or underline reserve words.

   g) Top and bottom margins of pages will be have variable settings: from 1
      to 7 lines.

   h) For the source listing, a top of page and bottom of page column marker
      line will be optional.  The marker line, which will look like:

      ....:....1....:....2....:....3....:....4.... etc

      will help show column positions of the source code.

   i) An option to ignore column setting will be provided.  Presently, only
      the source code up to the column (set with the F5 function key) is read
      on input.  This option will override whatever the column setting is
      and read the entire input line.

   j) Several techniques will be used to compress storage space required for
      maintaining data on cross referenced identifiers.  Therefore, parsing
      larger programs should be possible.


The following modifications are scheduled for Version 1.2, which will be
completed in the June 87 time frame:

   a) The biggest enhancement will be "automatic multiple file parsing".
      By allowing the user to include wild card characters ("*" and "?") on
      the "Active Input File" parameter line, he/she can then set the program
      to automatically parse all files in a directory that match the specified
      mask.

   b) Along with the Proc/Func Mapping, each line will reference the file
      the procedure/function was declared within.  References will be by
      number, and a file listing will give name, date and time of last update,
      and size of file.  The user can then check .COM and .CHN files to
      guarantee all date/time tags succeed date/time tags of source files, as
      well as give approximate LAST DATE OF MODIFICATION for procedures.

   c) The Proc/Func Mapping will show page numbers of the declarations, as
      well as line numbers.

   d) A new optional method of printing cross referenced identifiers will be
      available.  For extremely large programs, even the best compression tech-
      niques are helpless.  The new method will cross reference identifiers
      when the procedure the identifier was declared within is exited.  This
      will free storage space more readily and keep references near the source
      listing relevant to the identifiers.

   e) A new optional method of printing source listings will be added.  This
      method simply performs a page break at the end of each procedure/func-
      tion, rather than "compressing" the code together.

   f) A new diagram will be added to show the procedure/function calling
      techniques of the parsed source code, thereby giving all routines called
      from within each routine as well as all routines that call it.


Although this lists most of the scheduled changes to T-Ref over the next six
months, it far from completes the software package.  We are constantly thinking
of new enhancements to T-Ref and will incorporate them as time permits.

Version 2.0 ---

Finally, we are already looking at methods of integrating Modula-2, C, and
dBASE III into a single comprehensive cross referencing tool.  Of course, we
will change the name from "T-Ref" to "The-Ref".  For modular languages such as
Modula-2 and C, we have already discovered an innovative way to handle the
complexities of predefined libraries (or "packages").  We expect to have a
completed, multi-language cross referencing program available by the end of
1987.

Once your name is on our mailing list, we will keep you up to date on all
changes as they develop.  If you have any ideas at all as to how to make the
program better, please feel free to drop us a post card or letter.  We will
respond as time permits.  Thank you for your continued support.

-- Synergy Software of Nebraska
   904 Lariat Circle
   Papillion, NE  68046