Here's a final note on my FPC branches. While I could gather much
information about the FPC implementation, during these experiments, I
understand now that FPC - as a production compiler - will not undergo
much refactoring. So I'll discontinue work on these branches, and
continue further activities outside the FPC project and repository.
The recent (rudimentary) implementation of an Oberon/Component Pascal
(CP) parser suggests to me provisions for an extensible/experimental
compiler, based on the FPC infrastructure and back-ends, code name
FpcEx. My contribution to that project will be a transparent layer
between the FPC infrastructure and the parser, that makes the
implementation of modified or new parsers easier, without much knowledge
about the FPC implementation details. Modifications of other aspects of
the compiler will be supported as well, at least on demand (AnsiStrings,
hashtables, file/dir cache, garbage collection, threads...). For
demonstration purposes of the intended features and changes I've added
the CP to the AltParser branch, further updates to that branch are not
planned. If somebody is interested in such a project, further details
can be discussed in fpc-other, for now.
The FpcEx project will not take into account small machines, with
limited RAM and processing power, because such machines can be supported
by cross compilation. This restriction will allow to implement several
improvements, that may speed up compilation on current mainstream PCs.
It also will allow to use more classes and virtual methods, what's not
acceptable in an production compiler, but is essential for an
experimental/extensible compiler.
DoDi
_______________________________________________
fpc-devel maillist - fpc-devel@lists.freepascal.org
http://lists.freepascal.org/mailman/listinfo/fpc-devel
