Skip to content

Commit abc4434

Browse files
committed
Dropped the idea of linking to the code in the README.md
It's just far too likely any links to the source code will quickly fall out of sync. Maybe if GitHub allowed linking to symbols in the source, but since they already prevent non-users from doing even simple searches I don't expect that will happen any time soon...
1 parent 399bb38 commit abc4434

File tree

1 file changed

+9
-9
lines changed

1 file changed

+9
-9
lines changed

README.md

Lines changed: 9 additions & 9 deletions
Original file line numberDiff line numberDiff line change
@@ -378,7 +378,7 @@ There are a couple implementation tricks worth noting in ramcrc32bd:
378378
The end result is still $O(n^e)$, but limited only by your CPU's
379379
shift, xor, and branching hardware. No memory accesses required.
380380

381-
See [ramcrc32bd_read][ramcrc32bd_read] for an implementation of this.
381+
See `ramcrc32bd_read` for an implementation of this.
382382

383383
## Caveats
384384

@@ -397,9 +397,9 @@ And some caveats:
397397

398398
Still, it's good to be aware of this tradeoff.
399399

400-
ramcrc32bd's [`error_correction`][error-correction] config option lets
401-
you control exactly how many bit-errors to attempt to repair in case
402-
better detection is more useful.
400+
ramcrc32bd's `error_correction` config option lets you control exactly
401+
how many bit-errors to attempt to repair in case better detection is
402+
more useful.
403403

404404
2. Brute force doesn't really scale.
405405

@@ -410,11 +410,11 @@ And some caveats:
410410
in terms of message size, so this performance may be excusable if
411411
messages are small and bit-errors are rare.
412412

413-
ramcrc32bd's [`error_correction`][error-correction] config option can
414-
also help here by limiting how many bit-errors we attempt to repair.
415-
If you set `error_correction=1`, for example, the runtime reduces to
416-
$O(n)$ worst case, which is roughly the same runtime it takes to read
417-
the data from the underlying storage.
413+
ramcrc32bd's `error_correction` config option can also help here by
414+
limiting how many bit-errors we attempt to repair. If you set
415+
`error_correction=1`, for example, the runtime reduces to $O(n)$ worst
416+
case, which is roughly the same runtime it takes to read the data from
417+
the underlying storage.
418418

419419
But if you need a performant error-correcting block device, consider
420420
ramcrc32bd's big brother, [ramrsbd][ramrsbd], which brings the

0 commit comments

Comments
 (0)