Re: GSOC | Extending Common Lisp support

[Charles Turner <chturne at gmail dot com>]

On 1 July 2012 09:34, Per Bothner <per@bothner.com> wrote:
> Well, I think we need to break it up before it gets checked in.

OK, so my first patch is for the type refactoring. This isn't 100%
complete, but it's not far off. Scheme#formatType, string2Type seem
like good candidates for refactoring, but they haven't been yet.

The size of the patch is a bit misleading. Most of it is accounted for
from moving some large amount of code straight out of one class and
into another.

The two main changes were moving the unitNamespace out of Scheme and
into LispLanguage, and moving the types common to Lisp out of Scheme
and into LispLanguage. The other changes were fixes to other languages
that had previously been calling methods from Scheme to instead call
them from LispLanguage. Language also got a new abstract method,
getNamedType, which needed to be implemented in some of the other
languages. There are a few fixups to the imports that Netbeans alerted
me to.

-    Values forms = new Values();

This was removed because it isn't actually used, but might have been
confusing whilst reading the patch.

The getTypeFor method seems like it could quite easily be optimised.
Instead of all those string comparisons, we could instead have map
going from class names to LangObjTypes. This method is used quite a
lot IIUC.

I found in the past that whilst a patch is being reviewed, other
commits happen and envitable cause conflicts with the change logs in
my working copy, which are a pain to resolve. Is it OK to write the
ChangeLog entries once the patch is reviewed, and then resubmit?

I've seen on the news some pretty bad storms on the east coast of
America, which looked bad even for America's standards. I hope nobody
here is caught up in that!

Charlie.

Attachment: types.patch
Description: Binary data