fix: slightly better fuzzy scanning #115
Conversation
|
|
||
| if (patternOpcode != null && patternOpcode.ordinal != originalOpcode.ordinal) { | ||
| // reaching maximum threshold (0) means, | ||
| // the pattern does not match to the current instructions | ||
| if (threshold-- == 0) break | ||
|
|
||
| // look ahead | ||
| for (depth in 1 until threshold + 2) { |
There was a problem hiding this comment.
Kind of, I wanted to make it so that at a minimum at least the next instruction is considered. If threshold is 0 at that point, you'll only enter that loop once since you start at 1 and would go 'until' 2 (exclusive).
|
Before merging, can you rebase & test this PR against all patches for YouTube? |
oSumAtrIX
force-pushed
the
dev
branch
6 times, most recently
from
da2b438 to
9ec720e
Compare
|
Can this be merged soonish? PR has been open for a long while now, waiting for @oSumAtrIX to review. |
|
If possible, please test this against a majority of the existing patches. |
oSumAtrIX
force-pushed
the
dev
branch
3 times, most recently
from
646358d to
c5de9e2
Compare
|
Hey, apologies for the delayed reply. I won't be able to look at this for quite some time. If someone else can pick it up that'd be great or else I think it'd be safer to close the PR. |
|
Well it can stay open just fine until we can test it. |
|
@oSumAtrIX This should be reopened. |
|
What's keeping this PR from being merged after almost a year? Is no one available from the team to test this change, if necessary? |
|
It has low priority because it is currently unnecessary as well as needs to be tested if its breaking |
oSumAtrIX
force-pushed
the
dev
branch
2 times, most recently
from
64a5adf to
77dbee3
Compare
|
I believe, if we want to properly fix and implement fuzzy scanning, something like https://github.com/intuit/fuzzy-matcher should be used instead. |
oSumAtrIX
force-pushed
the
dev
branch
2 times, most recently
from
0fc5a47 to
ac1aff5
Compare
Closes #89