************************** Pascal-P5 1.4 RELEASE *****************************

You can also find this news at:

https://sourceforge.net/p/pascalp5/discussion/news/

WHAT'S NEW IN THE 1.4 RELEASE

* ISO 7185 negative and positive tests greatly expanded.

* Covers essentially ALL ISO 7185 error cases.

* Moved source configuration from SED to CPP (preprocessor).

* Now uses separate MPB includes for basic configuration.

* Certified for both 32 bit and 64 bit operation in Windows and Linux.

* Documentation updates.

* Increased available compiler options.

* Many bugs fixed.

* Source code tracking/debug system now more accurate.

* Enhanced regression reporting.

* Greatly improved error handling (no more crashes after error).

* Compiler/interpreter updated to tighter ISO 7185 standards checking.

* Spew checking (fault insertion checking).

* Added misspelled symbol and misspelled identifier recovery.

* Supressed duplicate errors on same undefined identifier.

* Supressed duplicate errors on files not in program header.

The 1.4 version of Pascal-P5 is the best tested, most ISO 7185 compliant
implementation of the Pascal language available. And at just over 10k lines
(total), it's also one of the smallest. It has been extensively checked for
operation in three different environments:

* Windows 32 bit with DOS shell.

* Windows 32 bit with Bash shell.

* Linux 64 bit with Bash shell.

It's test suite tests for both constructs that comply with the ISO 7185
as well as testing that the compiler rejects constructs that do not
comply. It in fact surpasses the old BSI test protocol, which is no longer
obtainable in any case. It has also been extensively fault tested by a
program that inserts random faults into the source code and tests that the
compiler will not crash or halt from simple source errors.

The 1.4 update incorporates many elements of Pascal-P6 that apply to the ISO
7185 standard only. Thus it was brought up to date with Pascal-P6.

This is expected to be the last major source code improvement for Pascal-P5.

Pascal-P5 comes from a long line of Pascal compilers, starting with the first
operational compiler for Pascal in 1973 from ETH Zurich by Niklaus Wirth's
group. Nearly 50 years old, it progressed through the versions Pascal-P1 to
Pascal-P4, and the versions Pascal-P2 and Pascal-P4 were used as the basis for
many operational Pascal compilers. Pascal-P5 brings it to full ISO 7185 status,
and Pascal-P6 extends the language greatly.

Questions:

Q1: Does Pascal-P5 generate machine code, or will it ever?

A1: No, it is designed to be an ISO 7185 compiler only. There has never been an
operational compiler that didn't extend the base language of Pascal, including
the original CDC 6000 compiler. Thus I decided to create a new version of
Pascal-P, Pascal-P6, that both extended the language and finished aspects of
Pascal-P that I thought needed addressing, like debug mode and machine code
generation.

Q2: Does Pascal-P5 have a use?

A2: Besides it's use as a model compiler, Pascal-P5 can be used as a "lint"
style utility to verify programs to the ISO 7185 language. For this you would
partition your code into ISO 7185 compliant and non-compliant sections, and use
the cpp provided "include" facility to allow compilation and checking of just
the core ISO 7185 part.

Q3: Does Pascal-P5 run on Mac OS X?

A3: Pascal-P5 needs a host compiler. GPC used to run on the system, but no
longer does. FPC runs on Mac OS X, but does not compile the current version of
Pascal-P5. There are mentions on the web of versions of Pascal-P5 that do run
under FPC, but have been modified from their ISO 7185 form and are likely
downrev (from a previous revision) as well. I understand from the FPC group
that most or all of the issues have been corrected in tip (the source you get
directly from git repository), so if you are able to compile a working FPC from
sources, that may work for you.

Q4: What is the advantage of the original Pascal/ISO 7185 language?

A4: Original Pascal as defined N. Wirth (with T. Hoare and others) was a type
safe or protected language, meaning that the program could not crash itself with
accesses beyond the end of an array, or wild pointers, etc. Protected languages
experienced a renassance in the 1990s with Java and C#, but Pascal did the same
(and arguably better) decades previously. Most people don't realize this because
neither the authors of those languages nor the computer press gave proper credit
to N. Wirth., and in fact the way pointers are kept type safe in Pascal (a
design by T. Hoare) is still better than that of Java and C# (so called "managed
pointers"). Futher, many of interpreted languages of today are popular because
they are also type safe, even though they are much slower. This is one reason I
still think there is a place for a language to replace C/C++, that is as fast
and as versatile, but still type safe/protected.