TML Pascal Returns to the Mac

By David E. Smith, Editor & Publisher, MacTutor

TML Pascal v3.0

Tom Leonard, TML

Version 3.0 of the TML Pascal compiler will be a completely new implementation of TML Pascal. The most significant change for this product is that it will be converted to operate with the Macintosh Programmer’s Workshop (MPW) as an MPW Tool.

The following is a list of the major features provided by TML Pascal v3.0:

• Operates within the Macintosh Programmer’s Workshop

• Complete support for Object Pascal

• Compiles MacApp

• Native 68020 and 68881 code generation

• Conditional compilation

• Assembly source code output as well as object code output

• Constant expression evaluation in declarations

• Performance analyzer

• Complete integration with other MPW languages including MPW C and Assembler.

In addition to the Pascal compiler, the product will also ship with a Pascal pretty printer (TMLPasMat) and Pascal cross reference tool (TMLPasRef).

TML Pascal will support the development of any type Macintosh program. These include:

• Desktop applications

• “Textbook” applications

• MPW Tool applications

• Desk accessories

• Definition procedures (WDEF, MDEF, CDEF, LDEF, etc.)

• FKEY resources

• INIT resources

• HyperCard XCMDs and XFCNs

and more...

Examples of how to create all of these various types of programs are provided with either the TML Pascal product or the TML Source Code Library.

In addition, TML Pascal will provide several MPW “scripts” which will implement custom menus and other user interface elements to make using the MPW shell easy for the beginning programmer. Compiling and executing a program will be as simple as selecting the “Compile & Run” command from the “TML Pascal” menu. In fact, many users may never find it necessary to use other MPW commands to compile and create programs.

External Tool Specification

Syntax TMLPascal [ option ] [ file ]

Description Pascal language compiler.

Input: One or more Pascal program or unit source code text files. By convention, Pascal source file names end with the suffix “.p”

Output: The compiler can create two types of output: MPW object code (the default) or MPW assembly source code. The output is written to a file whose name is the input source file name with the “.o” or “.a” suffix respectively.

There is no output to standard output.

Diagnostics: Compile time errors are written to diagnostic output. Progress and summary information is also written to diagnostic output if requested.

Status: The following status values are returned to the Shell:

0 Successful completion

1 Error in parameters

2 Compilation halted

Options:

-align Align all data items on longword boundaries.

-asm Compile the source file to assembly source output rather than object code.

-d name= TRUE | FALSE

Set the compile time variable name to TRUE or FALSE

-i pathname [,pathname]...

Search for include or USES files in the specified directories. Multiple -i options may be specified. At most 15 directories will be searched.

-mc68020 Generate code to take advantage of the 68020 processor.

-mc68881 Generate code to take advantage of the 68881 co-processor.

-o outName

Specify the pathname for the generated output file (object or assembly). If outName ends with a colon (:), it indicates a directory for the output file, whose name is then formed by the normal rules. If outName does not end with a colon, the output file is written to the file named outName.

-ov Turn on overflow checking.

-p Output compiler version and progress information to diagnostic output.

-r Suppress range checking.

-t Report compilation time to diagnostic output.

-u Initialize local and global data to the value $7267. Used for debugging.

-z Turn off the output of embedded procedure names in the object code. This option is equivalent to the {$D-} compiler directive. By default, the compiler creates the embedded procedure names for use by TMON, MacsBug, and other debuggers.

Language

TML Pascal v3.0 will be a compatible ANS Pascal and MPW Pascal compiler. In addition, a significant amount of time will also be spent further improving the code quality of generated code. Because the current version of TML Pascal is already very much compatible with MPW Pascal, no major changes are required. The following is a list of the more significant changes to be made for TML Pascal v3.0. See the TML Pascal v3.0 Language Reference Manual for a complete definition of the language.

• Conditional compilation

• Short circuit boolean operators ( &, | )

• Exponentiation operator ( ** )

• Identifiers may begin with an underscore

• Implement the “C” subprogram directive

• Fully implement Object types

• Type REAL changed to be equivalent to SINGLE

• Support 96 bit extended reals for 68881

• Support constant expressions where constant literals are normally expected

• Support sets in the range 0..2039

• Untyped files

• Type casting between ordinal types of different sizes

LS C and Resources

Bill Rausch

I think you missed the point in the May 88 letters column regarding the LightSpeed C development system (as did Peter Korn whose letter you were responding to). You do not need any “code hacks” in LSC to handle the resource fork of the application being developed. All you have to do is use the following naming convention:

the project file name: xxx

the resource file name: xxx.rsrc

and keep the resource file in the same folder as the project.

The resource file is automatically opened by LSC when you “run” the application during development and is automatically linked when you build the stand-alone application. My own convention is to name my project files “x.proj” and the resource files “x.proj.rsrc”, where the stand-alone application will have the name “x”.

Also, re: a MouseHole post by David Kosiur about fractional font widths. Both Microsoft Word 3 and Word Perfect have options for using the fractional font widths with the LaserWriter. In Word 3, the option is controlled via a check box in the print dialog box. In Word Perfect, choose Screen from the Special menu to find it. In both programs (and I assume all other programs) setting this option disables the “font substitution” option.

Hard Disk: Hit Here!

Mike Cohen

In response to the messages about Hard Disk problems in the Mousehole Report, here’s my two cents worth. My Mac II’s 40 Meg internal drive (made by Quantum) often refuses to boot when I turn it on. One morning it went on for several hours, and in desperation I called Developer Support. Dave Ramsey told me, “Oh yes, that’s a common problem with Quantum Drives. The head sometimes freezes to the surface & prevents it from spinning. Try giving it a whack on the side”. The drive worked just fine after I whacked it. [Ouch! -Ed]

Modula-2?

Davis R Bell

What has happened to Modula 2? In the old anorexic Mac days you carried articles on the subject and lots of good code, especially by Tom Taylor (now with Apple). The old M-code compiler was slow but now there is the Wirth 68000 compiler worked on by TML and SemperSoft. Is it a back-handed compliment (not ‘complemint’ as some of your authors seem to assume) to Modula which assumes that source translation from C or Pascal is a snap? I’ve just moved (deep breath) from the ETH (fount of all wirthwhile(sic) wisdom) MacMeth system to the TML MPW (Modula-2) setup. Strongly recommemded with lots of nice things like : a very versatile variant of the event record; generic ADDRESS types for Handles, etc.; source macros for the editor; full (I repeat ‘full’) access to IM Vol 5 calls; and lots of low-level primitives, which virtually allow the writing of primitive assembler routines as Modula code. Beefs: documentation could be better and, of course the slow but reasonably intelligent MPW linker. One thing TML/SemperSoft might think of doing is implementing the source level debugger from ETH which came with MacMet. This was a real benison for those lazy days when one can’t face TMON (which, in version 2.81 works fine with TML Modula). How about it? [Modula-2 is a language that seems to always end up being still-born on the Mac. TML has just released their MPW Pascal with object support, so once again, one has to ask how serious they will be about their Modula-2 product? They never bothered to send MacTutor a copy for Editorial support, so as far as we are concerned, there doesn’t seem to be much interest in Modula on the Mac out there. -Ed]

DataBase Challenge

Dr. H. M. Aus

For what it is worth, I would like to add my comments to your database debate.

We currently run a medium size database for the diagnostic labs here at our Institute. The database is homegrown and written in VAX-DIBOL. Our labs receive approximately 300 specimens a week on which up to 100 different tests, including AIDS, can be requested. Another collaborating lab processes 1000 specimens per week. My long term goal is to improve the reliability and user friendliness of the database. That is why we are looking at Macintosh and appreciate the recent comments and opinions in MacTutor.

Of all the databases that we have looked at, Double Helix II seems to be the most reasonable choice because you don’t have to program it. However, as you also point out there are too many icons involved. Odesta needs to allow the user to present the icons in a compact form. In our application, we need 100 icon data fields for the laboratory results, 100 icon flags to indicate which tests have been requested, 100 icon flags to indicate that the results have been sent to the physician or clinic, 100 icon flags to show that the cost of the tests have been invoiced, etc. Plus, all the patient, clinic and physician information need to be “iconized”. This is simply too many icons.

Lastly, based on our DIBOL experience, we do not want to program another database. Also, we do not want to either pre-process or post-process our data.

We’d enjoy hearing from other users about similar applications and their solutions.

P.S. Your recent articles on Mac II color have been very useful to us in our image processing project. Please, forward your author’s kit and we’ll tell you about it in the fall. Absoft’s FORTRAN 020 V2.3 compiler has been a big disappointment! [Guy Kawasaki thinks his 4th Dimension is the best way to go and has sent us an article on how to do it. Watch for it next month. -Ed]

TOPS is TOPS

Rick Cumings

This letter is in response to comments made by Richard Clark on the MouseHole BBS, as printed in MacTutor’s April 1988 issue. He had complaints regarding duplicate serial numbers of TOPS products, and delays in the shipment of upgrades. We’re sorry that he had problems receiving a TOPS PC upgrade in a timely manner. Just before we released TOPS II, we found a snag in the AppleTalk Transaction Protocol (ATP) lookup in AppleTalk over zones. It manifested itself in very slow response times, leading to unacceptable transfer delays. We developed a workaround which took some time to test and release.

Mr. Clark also alluded to Sun MicroSystems bearing the ultimate responsibility for the problems. To be fair, the problems were many and varied, but TOPS bears the responsibility for any delays. We experienced the delays and bad product shipments due to our inexperience at providing a new upgrade to customers. We learned much during this time about how to increase our efficiency and responsibility to our customer base. As a result, our upgrade department has turned-around the situation. If you call for an upgrade today, you won’t experience shipment delays or other probloms.

If developers or customers have any technical problems with TOPS products, they can reach TOPS Technical Support at 415-769-8711; the TOPS Talk BBS at 415-769-8774; AppleLink address D0098; CompuServe (subtopic TOPS within the Apple Vendor Forum); or BIX (TOPS Vendor Support Conference).

The Trouble With FONDs

Cliff Joyce

By the way, I noticed the discussion in a recent MacTutor issue regarding the problems with bitmapped fonts behaving strangely in applications that use fractional character widths.

The problem goes back to older versions of the Font/DA Mover Which improperly set bit 14 of the ffFlags word when building FOND resources from scratch. This flipped bit tells the application that the FOND contains a fractional width table. The application then uses ffWTabOff to find the address of the width table. But since ffWTabOff is zero in this case, the application points to the beginning of the FOND--usually causing characters to overwrite themselves when being drawn.

We have removed all FONDs from our World Class Fonts collections, and included Font/DA Mover v3.6, which properly sets bit 14 of the ffFlags word.

However, a safer solution might be for programmers to make two tests to determine if a FOND contains a fractional width table. First test the bitflag, then test ffWTabOff. If ffWTabOff=0, assume there is no fractional character width table.

By the way, Font/DA Mover v3.6 now improperly sets bit 15 of the ffFlags word when creating generic FOND resources (unless my Inside Macintosh volume IV is not accurate in describing the setting of bit 15). This may be a moot point as I am unaware of any applications that use this bit for anything (yet).

Oh, yes--enclosed is my check for another year’s subscription. That’s it for now. Keep up the good work...

Display Postscript

Michael Nelson

Basically, some Mousehole Report posting listed in the April issue of MacTutor related to a developer’s poor perception of Display PostScript’s performance. I recently had the opportunity to see a demonstration of Display PostScript, and receive an overview of its architecture from Charles M. Geschke, an Executive V.P. for Adobe. While I didn’t take copius notes on this presentation I do have a good feel for Display PostScript, and would like to set the record straight.

The writer’s claim that Display PostScript is slow is just not accurate.

The architecture of Display PostScript is “OPEN.” That is the product comes with a full software implementation of its graphics core, with hooks for vendors to add value through. Vendor value added can be implemented via a windowing system interface which they have full control over, and/or through LINKS TO HARDWARE ASSIST FOR ACCELERATING GRAPHICS PERFORMANCE.

While a quick and dirty (i.e. dumb) port of Display PostScript to a given platform is likely to be less than blazingly fast (although the video tape I saw of Display PostScript running on a MAC II, IBM PC RT, IBM PS/2 80, SUN s/60, and VAXStation 2000 all looked pretty darned impressive!), an intelligent port utilizing available hardware assists will realize the available performance in the box (note: the VAXStation and IBM PC RT in the aformentioned video tape both had hardware vector drawing implemented).

From what I saw I was impressed. Clearly Adobe did their homework and put together a very tidy package (and DEC agrees, no SMALL OEM to land first). If there is a negative that I can think of, it is in Adobe’s attitude. They have a nice product, know it, but are not being particularly aggressive in pushing it, especially on the Mac II platform. For instance, they had no comment on when, if ever, they would release a Mac product. [This might be because they don’t want to offend Apple? -Ed]

I’ve Got a Pen If You Have an Idea

Alastair Dallas

Congratulations on your recent MultiFinder application article. With all that’s been going on, it’s not surprising that my letter of February 11 got lost in the shuffle, but I am still interested in hearing from you. Like every other Macintosh software developer, I find MacTutor invaluable reading. Unlike every other reader, I am an experieced writer. I’d like to sell you an article.

Object-oriented drawing programs have received scant attention in your pages, except for Bob Gordon’s promise to cover that topic “next month” last May. I am in the process of creating an application that will expand on the concept of object drawing programs like MacDraw, and I would be interested in explaining what I know and more importantly, what I learn. For example, I protptyped my program in HyperCard. Since your readers are serious programmers, I expect that they would be interested in an article that explains how to do that without using up as many pages as Danny Goodman’s Book. [Look forward to whatever work you would like to send our way. I personally am still very interested in how to clone the MacDraw technology. We still need more Draw type tools and have not published much on the subject on how to create and manipulate small PICT type objects. -Ed]