Gpc August 2010

gpc@gnu.de

32 participants
32 discussions

Re: [GPC]: Ada
by Pascal 07 Aug '10

07 Aug '10

Hello Adriaan. Few years ago, Mac Pascal Interfaces were translated into Carbon-Ada. You can found them on www.macada.org. Basically Ada 2005 is supporting all modern object oriented paradigmes : classes as package + tagged record, multiple inheritance with interfaces (as Java), static and dynamic binding... Would be the translation of Cocoa Interfaces into Ada achievable? What would be necessary ? Anyway, I'm really interesting in "such thing like Objective-Ada" and contributing if I can help. Regards, Pascal. http://blady.pagesperso-orange.fr Le 5 août 10 à 09:46, gpc-owner(a)gnu.de a écrit : > De : Adriaan van Os <gpc(a)microbizz.nl> > Date : 5 août 2010 09:01:36 HAEC > À : gpc(a)gnu.de > Objet : Rép : Ada (was: D > > > But there is no such thing like Objective-Ada, is there ? That > would be a clear disadvantage of Ada as an intermediate language. > > Regards, > > Adriaan van Os

2 1

pascal code for numerical computations
by Leo Brewin 06 Aug '10

06 Aug '10

Hi, I'd like to put in my 2 cents worth for GPC. It is fantastic! I've been using GPC for the past few years and it has been an invaluable aid to my productivity. It meets all my needs (most importantly the support for conformant arrays and modules). My main area of research is in numerical general relativity and I have taught numerical methods to science students for a number of years. With the current lively discussion on the future directions of GPC I though it would be good time to make my own contribution by way of making some of my code available. It's a collection of numerical codes covering a range of standard numerical methods (root finding, interpolation, integration etc.) as well as some more advanced methods (multigrids, adaptive integrations, improper integrals, boundary value problems, genetic algorithms and so on). You can find the code at http://users.monash.edu.au/~leo/software/pascal/files/GPCExamples.tar.gz More information can be found in the README.txt in the parent directory. Cheers, Leo Brewin School of Mathematical Sciences Monash University Australia

2 1

For the wish list: Vector Arithmetic
by John L. Ries 06 Aug '10

06 Aug '10

Fortran 90 and later allow arithmetic operations on whole arrays, like so: !Both arrays must be the same shape integer foo(37), doublefoo(37) doublefoo=foo*2 This is actually a highly useful feature as it reduces the need for looping (a common place for bugs to hide) and allows array operations to be distributed over multiple processors, if available. In Pascal, the above might look something like this: var foo, doublefoo: array[1..37] of integer; {...} begin {...} doublefoo:=2*foo; {...} end. --------------------------| John L. Ries | Salford Systems | Phone: (619)543-8880 x107 | or (435)867-8885 | --------------------------|

4 7

Quo vadis, GPC?
by Frank Heckenbach 06 Aug '10

06 Aug '10

Hi everybody, since GPC development has mostly stalled, I've thought about if and how its development could be continued. Here are my current thoughts about it. Since the text is quite long, I put it on the web: http://fjf.gnu.de/gpc-future.html As I write there, I don't see much future in the current way of developing GPC, but also alternative development models will not be a task for a single person. In other words, without some new contributors, there is probably no chance for GPC development to continue. I don't really know how many of you currently use GPC, and to what extent and in which ways, e.g., do you use it just to maintain some legacy code, or are you actively writing new applications? So in order to tell whether continuing GPC development is worthwhile, I'd like to know who of you would actually care about major new features in GPC (as opposed to just preserving the status quo), and who would be interested not only in using GPC, but also supporting its continued development, either by actively contributing to it, or -- perhaps in the case of companies that use GPC -- by paying for continued development. If it turns out there is no sufficient amount of interest, I'm afraid to say it's probably better to put an end to it now rather than further dragging along. (Of course, the existing GPC versions will continue to be available, and anyone who wants to can use and modify them, which the GPL already guarantees, but without prospects for the future, I would then retire from GPC development and start to rewrite my own Pascal programs in other languages.) Frank -- Frank Heckenbach, f.heckenbach(a)fh-soft.de, http://fjf.gnu.de/, 7977168E GPC To-Do list, latest features, fixed bugs: http://www.gnu-pascal.de/todo.html GPC download signing key: ACB3 79B2 7EB2 B7A7 EFDE D101 CD02 4C9D 0FE0 E5E8

26 98

Pastel compiler (GCC's predecessor)
by Rugxulo 05 Aug '10

05 Aug '10

Okay, somebody mentioned the Pastel compiler (Pascal-ish, compiled itself) that was the original basis for RMS' hacking. He wrote his own C frontend and that later became GCC proper. It's very hard to find info on this, so I'm pasting here my recent findings for ya. (Note that I know less than probably anybody about this. I'm no knowledgeable insider or anything!) In particular, GCC's history and Wikipedia seem to omit mention of most of this. Somebody should also probably e-mail RMS to ask if he still has a copy of the sources around somewhere (interesting to me, at least, and I'm just vaguely curious!). "Jeff Broughton, then Amber project leader, implemented two extended Pascal compilers. The extended language was named Pastel (an off-color Pascal). A complete compiler targeted to the S-1 architecture was completed in several months. However, the first version of the compiler, which was implemented in slightly more than a weekend, produced PL/1 as its output! This was done to allow the continued use of Multics, which had no Pascal compiler, as the development site for Amber. At that point the project had a couple of man-years of code written in PL/1. The conversion of all PL/1 code to Pastel was accomplished in about 3 man months. (1998 historical note: at one point in the project Richard Stallman visited, and had the Pastel compiler explained to him. He left with a copy of the source, and used it to produce the Gnu C compiler. Most of the techniques that gave the Gnu C compiler its reputation for good code generation came from the Amber Pastel compiler.)" http://www.mit.edu/~cbf/thesis.htm http://killian.com/earl/ http://www.lbl.gov/cs/CSnews/CSnews073109a.html http://en.wikipedia.org/wiki/PathScale

2 3

Rewriting Pascal
by Prof. Harley Flanders 04 Aug '10

04 Aug '10

To Frank H. and all Pascal re-writers: One of the main, if not the main, point of keeping Pascal updated is to extend its usefulness. I want to address one area: scientific programming. It is what I do; I have done it in Pascal for vary many years, and continue doing so in Delphi. I even wrote a book, "Scientific Pascal", in 2 editions. So here is my wish list: Add types Complex Rational coded in assembler. All the standard operations and functions should apply to these types: Complex: +, -, *, /, ^ or ** Sqr, Sqrt, Sin, Cos, ..., Exp Ln Rational: +, -, *, /, Sqr, LowestTerms,... Even 40-year old FORTRAN has complex types. Obviously Complex is needed for numerical analysis, differential equations, electrical engineering, etc. A rational type would be very useful for number theory. Add higher precision floats (20 bytes, 24 bytes, user-chosen precision) again with all the standard operations and functions implemented in assembler. Borland introduced type Extended (10 byte floats), possibly in Turbo Pascal 4 or 5. This was an enormous improvement over the Real type, (4 byte), and made full use of the 80x87 mathematical coprocessor. Extended has remained the highest precision f.p. type in Borland Pascal and probably all other Pascals for 15-20 years, while computer speed and memory have improved by orders of magnitude. It is surely time for more accurate real types, coded in assembler. The current real types (in Delphi) are Single (4 bytes), Real48 (6 bytes), Double (8 bytes), Extended (10 bytes), [and Comp and Currency]. Synonyms are in order,skipping the factor of 8: Real4, Real6, Real8, Real10, etc. for the hopefully new ones. It is also time for more accurate integers. Borland introduced Int64 10-20 years ago, and it has remained the standard highest accuracy integer type. Delphi wisely added the names (and unsigned types) Int8, Int16, Int32, Int64, UInt8, UInt16, UInt32, UInt64. (My preference again would be to drop the factor 8: Int1, Int2, Int4, Int8, UInt1, UInt2, UInt4, UInt8.) I would love to see Int16, Int32, Int64 and Int128, and their unsigned counterparts. Add exponentiation, with the symbol ^ or ** overloaded for the special cases of the exponent: positive or negative integer [ (-0.02)^(-3) = -1.25E5 ], rational with odd denominator [ (-8)^(5/3) = -32 ]. FORTRAN and even BASIC have had exponentiation forever. Allow genuine operator overloading, such as operator "+" (const X, Y: MyType): MyType; operator "*" (const X: Integer; const Y: IntegerVector): IntegerVector; Delphi has made a mess of this so far, with overloading only possible with class functions and records. GNU Pascal and Free Pascal do much better. It should be possible to overload all the standard operators: + - * / ^ = <> <= >= > < There should be an abstract classes TRing which can be specialized to TField, etc. TRing should have a Ring Element that can be specialized. So one could define TMatrix, with elements in TRing, and define matrix multiplication A*B and Det (determinant) without specifying the elements in advance, and specialize them to any ring, such as the rings of rationals, of polynomials with complex coefficients, etc. This is what OOP, polymorphism, abstraction, inheritance should be about. Add Print Methods for TMemo, TStringGrid, TImage, and maybe a few other components. Obviously if I compute data and store it in a table or memo, or create a graph, I want to print it. Currently this requires searching the WWW for someone's code. One of the advantages of unicode (see below): one can mix alphabets in tables. For instance a table of number theoretic functions will have Greek headings: Omega, Phi, omega, etc. Any modern Pascal should use unicode, with all character and string procedures modified accordingly. This is a big advance Delphi started a year or so ago Add some more fonts that include Greek letters and many math symbols, and which can be easily added to code (which should be WYSIWYG). Fonts for scientific use should be non-justified, so that "9", "1" and "." are the same width in displays of numbers. As far as I can tell, Courier New (OEM CharSet) is the only satisfactory font at present. Harley Flanders -- ============================= Prof. Harley Flanders 3533 Windemere Court Ann Arbor, MI 48105-2867 Home: 734 668 1546 harley(a)umich.edu =============================

6 9

Re: Quo vadis, GPC?
by samiam＠moorecad.com 03 Aug '10

03 Aug '10

Frank heckenbach wrote: >samiam(a)moorecad.com wrote: >> P5 is a much better basis for a compiler front end, being 40 years >> old, very well documented, and completely ISO 7185 compliant (not a >> subset), and completely free of copyright or license encumberments. > Where is the copyright/license/public domain statement from its > authors? I didn't find them in p5.zip from your web site. Note that > "no copyright notice" doesn't mean "no copyright", but in fact the > opposite "all rights reserved" is the default. Like it or not, > that's the legal situation. Only since the Berne convention was adopted. Pascal-P predates the Berne convention (in the USA, I guess it depends on where and which treaty you are talking about, Berne dates back to 1886)[1], it was written in 1972. It was also openly distributed to anyone that asked for it. Steve Pemberton published it without problems (or agreement with the original authors) in his book, and made it available on his web site, again without restrictions (and I asked the author about it personally). You could also ask Wirth about it, but I don't suppose it would make a difference, since he didn't personally write it, his students did. Neither the old nor the new united states copyright laws specifically stated what the exact meaning of public domain was. The former law simply stated that copyright didn't apply by default, but the berne convention says it does. In any case, the courts have weighed in on the subject, and ruled time after time that works given out freely and freely published loose any copyright status. In fact, I have yet to see a single example of a work that was freely given out by the author, but later retracted and successfully sued over (note patents are a completely different subject). However, its not really my office to talk you into anything here, just point out possibilities. Scott Moore Berne convention: [1] This Convention shall apply to all works which, at the moment of its coming into force, have not yet fallen into the public domain in the country of origin through the expiry of the term of protection. US "Berne Convention Implementation Act of 1988" Sec. 12. Works in the public domain. Title 17, United States Code, as amended by this Act, does not provide copyright protection for any work that is in the public domain in the United States.

4 4

Re: Quo vadis, GPC? PL features for compiler
by Frank Heckenbach 03 Aug '10

03 Aug '10

spir wrote (by PM, but I supposed it was meant for the list): > > Yes, it's possible, but it's IMHO not conformtable. I know I'm > > repeating myself, but I'm really fed up with having to "reimplement > > the list", as I wrote. It's also a psychological thing IMHO -- if > > advanced data structures (objects, trees, hash tables) are readily > > available (as in C++), you tend you use them when useful; if they're > > not (as in Pascal, except for objects), you're much more likely to > > do quick ad-hoc work-arounds, which affect the efficiency, but also > > the design of your program, especially in the long run. GCC's > > TREE_NODEs are a particularly bad example of this, spreading all > > over the code, but also in my Pascal programs, I often find a lot of > > ad-hoc list (or other structures) handling code that obscures the > > actual purpose of my code and thus make it harder to understand > > later. > > This very clearly shows how much a (fully flexible and generic) > sequence type is a fondamental feature for any general-purpose > language. Thus, it must be part of new-GPC (meaning also > prerequisites such a generic & referenced types); then, we have > them available to write the new-GPC compiler in new-GPC ;-) I'd say not only simple lists (sequences), also hash tables, vectors (dynamic arrays) and trees are data structures I use regularly and would like to have readily available. > I take the opportunity to ask a side-question: is there a logical > or practical link between compiler target and implementation > languages? For instance, provided GPC's feature set is good enough > (see above) to write a compiler comfortably, At the moment it isn't yet WRT the above features. That's the chicken and egg problem I referred to. Either one uses C++ initially (which has the features) to implement a new compiler with the features and then ports the compiler to Pascal, or one uses the existing GPC to write the new compiler in a more awkward way and later cleans it up when it's self-hosting. Kevan Hashemi wrote: > I guarantee you that's the case for me. I read through Bison and I liked > it very much, and I am thoroughly convinced of its merit by your ^C > example. But my first look through the Bison manual page suggested that > the output was a C program that we would then compile and link to, which > means that at least part of our Pascal compiler would be written in C. The Bison output basically consists of generated tables (which are mostly numbers), and a fixed parsing algorithm that uses these tables which is more or less verbatim copied from a template file. It provides several such templates, the basic C algorithm, a GLR algorithm (that parses a larger class of syntax, which GPC requires) also written in C, and one in C++. So it's possible to add a Pascal template as well. Of course, it would have to be written (probably by us), which is some work -- though probably minor compared to the rest of the compiler, and of course, a one-time job, independent of grammar changes, and it would be useful to Pascal programmers who'd like to implement their own parsers. > Given the extent of my inexperience, but at the same time my eagerness > to help, perhaps you could explain to me how we could use Bison on the > way to a self-compiling Pascal, or how the C-code output of Bison (if > that's indeed what it puts out) could be part of the project in a way > that left us independent of the GCC back-end problems. It's nothing to do with backend problems. A newly written compiler implemented in C(++) with a non-GCC target (whether C++, Ada or LLVM) could still use Bison as is. Frank -- Frank Heckenbach, f.heckenbach(a)fh-soft.de, http://fjf.gnu.de/, 7977168E GPC To-Do list, latest features, fixed bugs: http://www.gnu-pascal.de/todo.html GPC download signing key: ACB3 79B2 7EB2 B7A7 EFDE D101 CD02 4C9D 0FE0 E5E8

2 1

Re: Quo vadis, GPC?
by samiam＠moorecad.com 03 Aug '10

03 Aug '10

1 0

Re: Quo vadis, GPC? (Managing by license)
by Objective Modula-2 03 Aug '10

03 Aug '10

John L. Ries wrote: > Such a policy would seem irrational to me, as I fail to see any real benefit > to the project, but it's possible I'm misunderstanding something. I didn't say it was policy. I said it has an influence. Many of the developers think that GCC can and should compete on technical merit alone, but a there are those who think that the design needs to protect the license. Sometimes decisions lean in one direction, sometimes they lean in the other direction. Both these groups have an influence. The decision of whether or not to introduce a plugin interface was certainly made at the highest level at the FSF, so you may want to consider than a matter of policy. But many decisions are made further down and they shouldn't be considered matters of policy. There was another incident many years ago where a developer had created a GCC to JVM translator or vice versa and was about to donate his code to the FSF. Richard Stallman was outraged and he trounced this poor guy in public, asking him to destroy his code so it would never leak anywhere. Clearly there is an influence of the school of thinking that says design has to follow the license. I wouldn't call it policy, but the influence is there.

2 1

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

2006

2005

2004

2003

2002

2001

2000

1999

1998

1997

1996

1995

Gpc August 2010