[Mono-devel-list] Pascal compiler?

Willibald Krenn Willibald.Krenn at gmx.at
Tue Mar 15 06:40:38 EST 2005


Thanks very much for your detailed answer.

> Why would it be in the Mono cvs? Chrome is a closed-source commercial
> product, it's not open source. I made that clear in my original post.

Of course you made that clear - perhaps I did not explain what /I/ meant clearly 
enough: I somewhat missed a pointer in svn (some module/binary) that links to 
Chrome, if Chrome really is considered to be the standard mono pascal compiler.

> Letting alone any Open Source vs. Commercial debate, i think that in
> general it's superfluous if 35 different groups reinvent the wheel
> simultaneously, just for the sake of "doing it themselves". I honestly
> believe that Chrome will bring a lot to the table for Mono Pascal
> developers, and there's a lot to be gained by working together rather
> then against or parallel each other, duplicating the effort.

Finally, I downloaded a copy of the command line compiler and took a look at the 
examples. You got a lot of things in many developers always wanted, but I have 
to confess that your Pascal looks much more C# than Delphi.NET I guess.

You've modified with to work as 'with a := new XY do begin', and a 'for' now 
reads 'for i:integer = x to y do begin': Why did you leave the 'do begin' in if 
you were creating a new Pascal dialect? And why do you have to declare i as 
integer when a in the with statement is some undeclared variable where the 
compiler infers type information (I assume)?  What does 'var' indicate on 
class/instance fields? I mean they are all variables..
Are you allowed to declare value types in Chrome? If no, then I'm of the opinion 
that you should re-think that and add the possibility of creating value types 
too, because you give up optimization potential by the JIT compiler.

> As i said before we're open for input, feedback and assistance with
> Chrome on /many/ levels, not necessarily restricted to pure testing and
> feedback, and i'd be very interested in getting people involved in
> actual coding tasks related to the project (although, one month before
> release of our 1.0 version it's doubtful that it would make sense to
> have someone new come in and - say - rewrite the core parser ;-). 
> Stuff that needs work (and would directly benefit Mono in particular)
> are things like integration with MonoDevelop (which we have planned but
> currently don't have resources to spare on, as we do have a very small
> team working on Chrome), which should be a very fun project for one or
> two people that want to get involved with both Chrome and Mono (we've
> got quite a bit architecture in place already from our VS integration
> that could be leveraged, so this wouldn't be to extensive a project
> either). 
> Also interesting to do would be testing and prepping the compiler under
> the "more exotic" platforms supported by mono (which, to me means
> non-x86-Windows/Linux <g>)

I think I understand your offer now. Nevertheless I'll ask the following 
question: Could you imagine opening up the basic compiler you give away for free 
like Ximian/Novell did with mono? You would not lose the chance of selling the 
compiler to other companies on some non GPL basis and you still would have 
control over who has write access to your repo.

> See Todd's reply: we certainly have. Looking back thru my inbox, i
> actually do see /you/ participating in discussion about Chrome, a
> several months back ;-)

This must have been somewhen in summer(?) 2004 - and I deleted all the old 
messages a few weeks ago to free up space on my imap account. Anyways, since 
then I stumbled over Chrome on lots of different places, so please forgive me, 
if I forgot about the debate back then on this list.


More information about the Mono-devel-list mailing list