a2d/CONTRIBUTING.md

70 lines
3.3 KiB
Markdown
Raw Normal View History

2018-03-26 04:16:29 +00:00
# Contributing To The Effort
Contributions welcome! Preliminaries:
2018-04-12 04:39:38 +00:00
* Please review the [README](README.md).
* Read and adhere to the [Code of Conduct](CODE_OF_CONDUCT.md).
2018-03-26 04:16:29 +00:00
* Read the [Coding Style](CodingStyle.md) guide.
## Sub-Projects
1. Disassembly of the MouseGraphics ToolKit
2018-03-27 01:41:38 +00:00
* Continue the effort to understand this powerful GUI library.
* Make it relocatable, work on C bindings for cc65.
2018-12-10 04:43:56 +00:00
* Pure disassembly changes take place in the `disasm` branch, which builds identically to the original.
2018-03-26 04:16:29 +00:00
1. Disassembly of DeskTop itself
2018-03-27 01:41:38 +00:00
* The core bits of DeskTop.
* The various overlays.
* The DiskCopy overlay is basically a stand-alone app. Could be fun.
2018-04-11 03:36:40 +00:00
1. Disassembly of [Selector](https://github.com/inexorabletash/a2d/issues/63)
2018-03-26 04:16:29 +00:00
1. Bug fixes
2018-03-27 01:41:38 +00:00
* Try and tackle some of the bugs in the [issue tracker](https://github.com/inexorabletash/a2d/issues?q=is%3Aissue+is%3Aopen+label%3Abug-in-original).
2018-03-26 04:16:29 +00:00
1. Add new Desk Accessories
2018-06-02 18:32:30 +00:00
* List of ideas in the [issue tracker](https://github.com/inexorabletash/a2d/issues?utf8=%E2%9C%93&q=is%3Aissue+is%3Aopen+label%3A%22Desk+Accessories%22+label%3A%22feature+request%22)
2018-05-07 20:39:56 +00:00
## DeskTop Disassembly Burn-Down
To feel confident about making additions and fixes to DeskTop, we need to
make sure we're not breaking things. That can be done in some cases by
relying on API boundaries, such as between MGTK and the DeskTop application.
But DeskTop itself is a big, monolithic application with multiple overlays,
so we need to understand nearly all of it before we can start moving code
around.
The `res/stats.pl` tool provides is a quick and dirty analysis of the
progress in turning raw da65 output into something we can confidently
modify. Here's a snapshot of the output for some files:
```
Stats:
desktop_main.s unscoped: 246 scoped: 1109 raw: 60 unrefed: 29
desktop_res.s unscoped: 64 scoped: 0 raw: 4 unrefed: 64
desktop_aux.s unscoped: 83 scoped: 301 raw: 2 unrefed: 32
loader.s unscoped: 1 scoped: 0 raw: 20 unrefed: 0
mgtk.s unscoped: 0 scoped: 10 raw: 13 unrefed: 0
invoker.s unscoped: 0 scoped: 0 raw: 2 unrefed: 0
```
* **unscoped** counts the number of auto-generated labels like `L1234`
produced by da65 which are _not_ in _two_ nested scopes. A scope is
used for the overall structure of a module, and a nested scope
is used for procedures. This counts labels which are not in either,
2018-06-02 18:32:30 +00:00
and thus may have some affinity for a particular address
and therefore can't be safely moved.
2018-05-07 20:39:56 +00:00
* **scoped** counts the number of auto-generated labels like `L1234`
produced by da65 which _are_ inside two nested scopes. Once a label
is local to a procedure it is generally safe to move, although
actually understanding the purpose and giving it a more meaningful
label is even better.
* **raw** counts the number of 16-bit addresses (`$1234`) in the code;
these usually refer to routines, resources, or occasionally
buffer locations that need to be understood and replaced with
labels.
* **unrefed** counts the number of auto-generated labels like `L1234`
that lack references within the source file. Early on, these often
hint at bogus disassembly but can also signal that the code to
use a routine or resource has not yet been identified.