Skip to content

docs: compare ODFileSystem with other Amiga CD filesystems - #27

Merged
reinauer merged 3 commits into
mainfrom
readme-comparison
Oct 31, 2026
Merged

reinauer merged 3 commits into
mainfrom
readme-comparison

Conversation

@reinauer

Copy link
Copy Markdown
Owner

Gives readers a reason to choose ODFileSystem, based on measured results, now that the unsupported "most complete" claim is gone.

  • docs: compare ODFileSystem with other Amiga CD filesystems adds a "Why ODFileSystem" section with a feature matrix and CDBench results, including where other filesystems lead. docs/comparison.md has all eight filesystems from CDBench, the full results and the test setup. The claim now reads as format support among the filesystems tested.
  • docs: focus the README on installation and usage is the README restructure this builds on. It conflicted with the OS3 and OS4 size-limit update on main, so docs/building.md carries the new limits.
  • amiga: use register arguments for OS3 handler builds is included because the published numbers depend on it: most of the lock and open improvement comes from this change, and it wasn't on main.

The feature results come from running each handler on test discs in WinUAE, not from documentation. Multisession isn't compared, because flat images have no multisession table of contents.

Links to BENCHMARKS.md show the new numbers once reinauer/CDBench#1 is merged.

Tested with make check, make golden-check and an OS3 make amiga LTO= build (80 352 bytes).

Pass -mregparm=2 when compiling and linking all OS3 handler variants
to reduce code size and CPU overhead. Keep standalone test programs
on the C runtime calling convention and leave OS4 and AROS unchanged.

Make handler objects depend on the Makefile so this ABI change rebuilds
existing objects instead of mixing calling conventions.
Lead with handler installation and mount configuration, document AIFF
selection, and keep format behavior and limitations in the README.
Remove transient test results, implementation notes and marketing
claims.

Move detailed build and release procedures into dedicated guides and
real-image fixture instructions into the testing guide. Replace local
absolute links with relative links and update documented size limits.
The README no longer told readers why to choose this filesystem, since
the unsupported "most complete" claim was removed.

Add a "Why ODFileSystem" section with a feature matrix and the CDBench
results, including where other filesystems lead, and restate the claim
as format support among the filesystems tested. Add docs/comparison.md
with the full results and test setup, so the claims can be checked and
reproduced. Document Activate = 1 and its empty-drive limitation.
@reinauer
reinauer merged commit e73f056 into main Oct 31, 2026
2 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant