#tux4kids chat log for 2002-05-11

Summary

The purpose of this IRC meeting was to discuss some coding issues and progress on the new management transition of Tux Typing. By way of an introduction, there has been a great deal of activity these last few weeks from the new project manager Jesse Andrews and from one of the developers (and current Release Manager) Calvin Arndt.

Calvin has been working on making the code run more efficiently on slower hardware as well as supplying some new Win32 test binaries. He has also been attacking several memleaks

Jesse has been rewriting much of the code that could be modularized as well as cleaning up ugly sections of code, and dealing with those issues involved in becoming a new project manager.

Now, without further ado, the summary of the chat log:

• There has been some confusion regarding the SDL function SDL_CreateRGBSurface. Calvin claims that removing it solves a lot of memleak troubles. [It's probably not SDL_CreateRGBSurface(..) itself that is creating this memleaks. If anything, it's problems in our own code, as SDL has undergone some pretty intensive development and testing over the last many years. - Sam, Ed.]
• UNICODE and i18n support is coming along nicely. Initial SDL_TTF (TrueType font) support has been added, and Jesse is committed to make UNICODE work for the 2.0 release.
• It was clarified that we will still be using SourceForge's CVS hosting for Tux Typing rather than switch over to the new Development Server. This is done for consistancy, and because SF.net's CVS server still works fine for our needs (i.e., there is no reason to make the switch).
• Jesse said that he would merge his work and Calvin's together and commit them to the development branch.
• Related to that, a developer can tell when they are working in the developmental branch by reading the README file in the tuxtype/ root directory. If they are working in the development branch, the top of the file will say: "Tux Typing (DEVELOPMENT BRANCH):"
• There was some discussion of the proper use of diff, patch, and cvs together: How they could be used together, and what made them seperate and distinct.
• Jacob Greig will be writing up a submission policy and information page for the Tux4Kids Bug-Tracking Group (which Jake also runs). He intends to model it after those documents in use at AbiWord.
• There was a question why fprintf to stderr was used for verbose output instead of the more conventional printf to stdout was used. The original reason was to keep compatibility with some strange legacy compilers which would ignore all stdout in the event that the application was not run from the console. However, with an increasing ammount of POSIX compliance in many other (non-Linux/BSD) OSes, this concern may not be an issue any more.
• Sam Hart has agreed to do further animations for Tux Typing. He has some un-used animations from other programs (such as TuxMath) which he will be placing in the Images CVS on the new dev server.
• There was further emphasis that for an SDL program to be truely cross-platform, it must follow this guideline for including SDL related headers.
• There was some discussion of dirty-blits, what they are, why we use them, and why not use them in the main menu. I wrote a small explaination on the mailing list about them, and why they couldn't be used in the main menu. You can read the message here.
• There was quite a bit of discussion that Calvin's changes make the main menu go too slow (Jesse claimed to be getting less than 10fps). This will be looked into further, but there was no resolution.