pl1gcc-development — Development of the PL/I front-end for GCC

[pl1gcc-development] Fwd: Re: Question for pl1gcc

[Henrik S. <hen...@ba...>]

Hi Bjarke,

You have to clarify your contract situation before you endeavour into any
 open source projects. Are you also writing compilers for your company or are
 there any other obvious borderline conflict of interests ?
I guess you will have to get a _written_ and signed statement from your
employeer that they either waive any rights they might have to your work done
for pl1gcc, or at least they are accepting the work your do is released under
the GPL license. This is very important you have clear lines here !!

The initial version of the compiler is based on this version:
 http://www-306.ibm.com/software/awdtools/pli/plimvs/library/

The pl1 preprocessor seems like a good isolated project.
Someone mentioned that a macro language named m4, but I haven't actually
 tried it. I would probably use a combination of lex and yacc, but fell free
 to come up with something completely different.
One thing to consider is how to integrate the pl1gcc preprocessor (pl1pp)
 into the gcc command and how the output gets passed to the pl1gcc compiler.
 Traditionally on pl/i the preproc can be activated via options given to the
 compiler, and I guess the same is the case for gcc.
When you start to read on the pl/i preproc, forget everything you know about
the c preprocessor. pl1pp is very different. But try to make up your own
opinion and let us discuss further.

Henrik

On Tuesday 26 October 2004 19.49, you wrote:
> Hi Henrik
>
> My biggest concern at the moment is the copyrights issue. Not that
> I want to keep any rights, but since I am employee of a software
> company there is a contractual side. Basically they own the code I
> write, even though it has no direct link to their business. I would
> have to get written permission from them saying that they have dont
> want any rights to this. I don't think it is problem, but it might be.
> Companies are strange that way sometimes.
>
> That being said. It seems to me that the best place to start
> for an outsider are these items:
>
> - Writing a preprocessor: It is rather self-contained units, and
>   won't compromise any already produce code.
>
> - Restructuring pl1gcc manuals: A good way to find the right
>   questions and learn the details of the project. Although I not
>   entirely sure what you mean by manuals.(writting man-pages?)
>
> - Start writing real test cases: This is also a good way
>   to learn about the project, and might benefit from less
>   PL/I experienced people like myself. Since I would make
>   mistakes that an experienced PL/I developer might not.
>
> - Define good structure for errormessages: This also seems
>   a good place to start. It will help anyone understand
>   the current code-base.
>
> Clearly the two most interesting tasks are the first and last
> in the above, and those would be the ones that initially would
> interest me.
>
> I have downloaded and read the PL/I manuals from IBMs homepage,
> and I assume they are the ones you mention. I haven't been monitoring
> the gcc discussion on gcc 4.0, but I will have a look ;o)
>
> Best regards
> Bjarke
>
>
>
> -----Original Message-----
> From: Henrik Sorensen [mailto:hen...@ba...]
> Sent: 26. oktober 2004 04:41
> To: Bjarke Wegge Andersen
> Subject: Re: Question for pl1gcc
>
>
> Great you want to help !
>
> Are you already registered as sourceforge user ?
> It is completely free and harmless, but if you are registered you can
> update the CVS etc.
> Use link http://sourceforge.net/account/newuser_emailverify.php
>
>
> ok where do we start...
> Firstly do you mind it we continue this discussion on the
> pl1...@li... ?
>
> that way we can engage some more people in the discussion, and also use the
> list more active.
>
> The README is currently the best source for ideas. I know it isn't much,
> and if more developers join the effort this will have to be changed to
> maybe some
> kind of sourceforge tasks or any other kind of structure.
>
> You are absolutely spot on regarding the integration with gcc. If you
> follow the gcc discussion, you will know that the tree-ssa branch got
> accepted into the mainline and will form the basis for gcc 4.0. This branch
> has a much simpler intermediate tree structure, that I always wanted to
> use. Since this branch is settling down, it might be a good time to
> actually try to use it.
>
> Due to the nature of the PL/I  programming language, we have to do a
> two-pass
> parsing. My first plan was to use an pl1gcc specific tree for the first
> pass,
> and then on the second pass create the gcc tree structure.
> Pass 1:
>  Scope boundaries (partly done)
>  Declare structures
> Pass 2:
>  build matching gcc tree structure.
>
> btw, do you have links to the PL/I manuals ?
>
> ideas for tasks:
>  Restructure pl1gcc manuals.
>  Make pl1gcc work with the gcc-4.0 branch again.
>  Find out how to integrate external libraries into pl1gcc, like GMP and
> glibc.
>  Write a PL/I preprocessor for pl1gcc (pl1pp?)
>  Start writing real test cases.
>  Define good structure for errormessages.
>  Continue the development of pl1gcc itself ;-)
>
> One important last point, since it is my intension to have pl1gcc
> integrated into gcc, once we have a more workable version, you must be
> prepared to handover any copyrights you might have within this project to
> FSF.
>
> let me know how you see this.
>
> Henrik
>
> On Monday 25 October 2004 21.51, you wrote:
> > Hi again
> >
> > After doing some reading on GCC frontend it seems that you still
> > need to integrate bison with the tree and rtl structure of GCC.
> >
> > Although I never done any work with GCC frontend, I do have some
> > experience using flex, bison and have been doing a bit of C for a
> > few years. I would be glad to lend a hand, if you can use it.
> >
> > Is there anywhere that I can read up on what the plan for
> > the next few steps are(other than the README file)? That way I
> > could assess when it would be possible for me to assist (Since
> > my PL/I skills are not that great).
> >
> > Best regards
> > Bjarke Wegge Andersen
> >
> >
> > -----Original Message-----
> > From: Henrik Sorensen [mailto:hen...@ba...]
> > Sent: 24. oktober 2004 20:16
> > To: Bjarke Wegge Andersen
> > Subject: Re: Question for pl1gcc
> >
> > On Wednesday 20 October 2004 22.43, you wrote:
> > > Hi
> >
> > Hello !
> > Sorry for answering a bit late. My laptop had a nervous breakdown, after
> > the latest upgrade ...
> >
> > > I just downloaded and installed pl1gcc using cygwin. The instructions
> >
> > where
> >
> > > easy to follow, and I encountered no problems. :o)
> >
> > thanks for letting me know. Currently all my machines are running Linux
> > so I really appreciate any feedback regarding cygwin.
> >
> > > I am, however, a bit confused. Is it correct that the compiler "only"
> > > parses pl1-code and doesn't produce any executeable code? I am quiet
> > > new
> >
> > to
> >
> > > PL/1, so foregive me if the question is immensely stupid. :o)
> >
> > Well thanks for asking anyway, and it is not that stupid.
> >
> > Currently the compiler is under development, so it will take a while
>
> before
>
> > any executable code will be generated. The first diffucult part was to
>
> find
>
> > a
> > way to structure PL/I, and now that the language can be parsed with flex
> > and bison, only some hard work is required to create the compiler.
> >
> > I just need to find some time again.
> >
> > > Best regards
> > > Bjarke Wegge Andersen, DK
> >
> > Henrik, CH, ex DK ;-)

-------------------------------------------------------

[pl1gcc-development] RE: Question for pl1gcc

[Bjarke W. A. <bw...@ad...>]

Hi Henrik

My biggest concern at the moment is the copyrights issue. Not that
I want to keep any rights, but since I am employee of a software
company there is a contractual side. Basically they own the code I
write, even though it has no direct link to their business. I would
have to get written permission from them saying that they have dont
want any rights to this. I don't think it is problem, but it might be.
Companies are strange that way sometimes.

That being said. It seems to me that the best place to start
for an outsider are these items:

- Writing a preprocessor: It is rather self-contained units, and
  won't compromise any already produce code.

- Restructuring pl1gcc manuals: A good way to find the right
  questions and learn the details of the project. Although I not
  entirely sure what you mean by manuals.(writting man-pages?)

- Start writing real test cases: This is also a good way
  to learn about the project, and might benefit from less
  PL/I experienced people like myself. Since I would make
  mistakes that an experienced PL/I developer might not.

- Define good structure for errormessages: This also seems
  a good place to start. It will help anyone understand
  the current code-base.

Clearly the two most interesting tasks are the first and last
in the above, and those would be the ones that initially would
interest me.

I have downloaded and read the PL/I manuals from IBMs homepage,
and I assume they are the ones you mention. I haven't been monitoring
the gcc discussion on gcc 4.0, but I will have a look ;o)

Best regards
Bjarke



-----Original Message-----
From: Henrik Sorensen [mailto:hen...@ba...]
Sent: 26. oktober 2004 04:41
To: Bjarke Wegge Andersen
Subject: Re: Question for pl1gcc


Great you want to help !

Are you already registered as sourceforge user ?
It is completely free and harmless, but if you are registered you can update
the CVS etc.
Use link http://sourceforge.net/account/newuser_emailverify.php


ok where do we start...
Firstly do you mind it we continue this discussion on the
pl1...@li... ?

that way we can engage some more people in the discussion, and also use the
list more active.

The README is currently the best source for ideas. I know it isn't much, and
if more developers join the effort this will have to be changed to maybe
some
kind of sourceforge tasks or any other kind of structure.

You are absolutely spot on regarding the integration with gcc. If you follow
the gcc discussion, you will know that the tree-ssa branch got accepted into
the mainline and will form the basis for gcc 4.0. This branch has a much
simpler intermediate tree structure, that I always wanted to use. Since this
branch is settling down, it might be a good time to actually try to use it.

Due to the nature of the PL/I  programming language, we have to do a
two-pass
parsing. My first plan was to use an pl1gcc specific tree for the first
pass,
and then on the second pass create the gcc tree structure.
Pass 1:
 Scope boundaries (partly done)
 Declare structures
Pass 2:
 build matching gcc tree structure.

btw, do you have links to the PL/I manuals ?

ideas for tasks:
 Restructure pl1gcc manuals.
 Make pl1gcc work with the gcc-4.0 branch again.
 Find out how to integrate external libraries into pl1gcc, like GMP and
glibc.
 Write a PL/I preprocessor for pl1gcc (pl1pp?)
 Start writing real test cases.
 Define good structure for errormessages.
 Continue the development of pl1gcc itself ;-)

One important last point, since it is my intension to have pl1gcc integrated
into gcc, once we have a more workable version, you must be prepared to
handover any copyrights you might have within this project to FSF.

let me know how you see this.

Henrik




On Monday 25 October 2004 21.51, you wrote:
> Hi again
>
> After doing some reading on GCC frontend it seems that you still
> need to integrate bison with the tree and rtl structure of GCC.
>
> Although I never done any work with GCC frontend, I do have some
> experience using flex, bison and have been doing a bit of C for a
> few years. I would be glad to lend a hand, if you can use it.
>
> Is there anywhere that I can read up on what the plan for
> the next few steps are(other than the README file)? That way I
> could assess when it would be possible for me to assist (Since
> my PL/I skills are not that great).
>
> Best regards
> Bjarke Wegge Andersen
>
>
> -----Original Message-----
> From: Henrik Sorensen [mailto:hen...@ba...]
> Sent: 24. oktober 2004 20:16
> To: Bjarke Wegge Andersen
> Subject: Re: Question for pl1gcc
>
> On Wednesday 20 October 2004 22.43, you wrote:
> > Hi
>
> Hello !
> Sorry for answering a bit late. My laptop had a nervous breakdown, after
> the latest upgrade ...
>
> > I just downloaded and installed pl1gcc using cygwin. The instructions
>
> where
>
> > easy to follow, and I encountered no problems. :o)
>
> thanks for letting me know. Currently all my machines are running Linux so
> I really appreciate any feedback regarding cygwin.
>
> > I am, however, a bit confused. Is it correct that the compiler "only"
> > parses pl1-code and doesn't produce any executeable code? I am quiet new
>
> to
>
> > PL/1, so foregive me if the question is immensely stupid. :o)
>
> Well thanks for asking anyway, and it is not that stupid.
>
> Currently the compiler is under development, so it will take a while
before
> any executable code will be generated. The first diffucult part was to
find
> a
> way to structure PL/I, and now that the language can be parsed with flex
> and bison, only some hard work is required to create the compiler.
>
> I just need to find some time again.
>
> > Best regards
> > Bjarke Wegge Andersen, DK
>
> Henrik, CH, ex DK ;-)

[pl1gcc-development] PL/I Front-end for GCC v0.0.8

[Henrik S. <hen...@ba...>]

Cross posted to PL...@LI..., gc...@gc... and
comp.lang.pl1


August 2004

This is the eigth code drop of the GCC front-end for the PL/I programming 
language.

PL/I for GCC is released under the terms of the GNU Public License; version 2.

The version 0.0.8 marks an important milestone for the PL/I front-end project, 
because the entire PL/I language can now be parsed using only flex and bison.

If you are testing the PL/I compiler, please inform us about any syntax errors
that might occur. It suffice to send the offending line of code.

Changes in v0.0.8:
 - Implemented token based lookahead.
 - All PL/I Keywords can be used as identifiers.
 - Reduced number of conflicts in the grammar.
 - Addapted installation instructions to use the gcc-core-3.5 snapshot
 - Numerous new testcases has been added.
 - Simplification of INITIAL attribute.

Fixed syntax in v0.0.8:
 - Added FETCH and RELEASE statements.
 - Floating point constants can start with just a '.'
 - Repetition factors in edit format lists.
 - Allow nested DO in PUT/GET datalists.
 - PUT/GET options.
 - Preconditions can now be placed where a label can be placed.
 - ENVIRONMENT options has been reviewed.
 - REWRITE statement was fixed up.
 - LOCATE statement was corrected.
 - Added REFER to the array dimension attribute.
 - Structure declaration in signature for entry parameters.
 - Cleaned up variable declarations.

For a complete changelog look in the README file.

What is pl1gcc ?
The pl1gcc project is an attempt to create a native PL/I compiler using
the GNU Compiler Collection. The project is looking for more people to
join the development and testing. If you want to help speed up the development
of a free PL/I compiler please do contact us and join the fun.

Looking ahead.
The next release will focus on integration of syntax from VA PL/I and Multics.


If you do try the compiler on some of your own code, please let us know
how it goes.

For more information, comments, feedback and download, please visit

http://pl1gcc.sourceforge.net

Have fun

The GCC PL/I Team.
gcc...@us...