Re: [freesci-develop] cleanup patch

[Jordi Vilalta]

Hi,

2009/1/26 Max Horn <address@hidden>:
> I saw no attachment to that mail. Seems there are problems with this list &
> attachments. Maybe you guys should just upload patches somewhere, and then
> send around mails with URLs to the patches? Right now I read in the #freesci
> logs that you already noticed the missing patch, and that essentially things
> are now delayed because Jordi tries to resend the patch... Let's try not to
> get bad communication lines in the way of things :-). Like, Jordi could
> email the patch directly to walter if walter wants to review it, instead of
> trying to resend to this list and it then failing again... ;-)

Thanks for the advice :P Here's the patch I tried to send on my last mail:
http://www-plan.cs.colorado.edu/freesci/tmp/headers_cleanup.patch

> Also, I'd like to renew my offer with help on cleanup and/or integration,
> both in terms of advice, and actual coding help, when it comes to cleanup;
> or assistance with migrating stuff from C "pseudo classes" to real C++
> classes) (such as seg_manager_t, sbtree_t, and countless others).

Thanks for the offer! As we discussed yesterday on IRC, there's
another task previous to converting to C++, which is physically
organizing the source/headers in said C "pseudo classes", which is
what we're trying to do right now. For example, right now some headers
contain interfaces for more than one source file.

After that we could start converting classes to C++. Another question
raised was whether we must do the conversion to C++ on the current
FreeSCI or after branching, and whether to use the ScummVM APIs right
from the start or start converting it to plain C++. There seemed to be
agreement that converting it to the ScummVM while doing the C++
conversion would be faster and probably easier, since we can benefit
from ScummVM's common classes. In that case it would be done outside
the main FreeSCI branch.

The last remaining issue is preparing the FreeSCI tree to fit into
ScummVM's SVN (removing useless files, apply code formatting
conventions, etc.). This will be done on a separate branch, since it
will break the standalone FreeSCI.

So briefly we have:
1) split/organize headers
2) make a new branch to prepare the source to fit ScummVM's SVN
3) do the conversion to C++ and use the ScummVM API
4) import the source code into ScummVM's SVN

We mainly need advice on 2) and 4) to know what are the requirements
we need to achieve. Maybe we could even postpone 3) and work on that
after it's been imported into ScummVM's SVN...

Anyway it looked like a good idea to give Max an account on the Darcs
server so he can help us with 1), 2) and eventually also 3 if we
decide to do it before moving to ScummVM's SVN.

I've just tried to sumarize here what we discussed yesterday on IRC.
Feel free to express your opinions!