Revert "libatalk: Restore invalid metadata cleanup in ad_open.c"

[rdmark]

Reverts #513

[rdmark]

@dgsga All of a sudden we have multiple reports that the EA metadata validation still fails in some environments.

#992
#993

I propose we revert this for now to buy us some time to get to the root of it, without blocking the 3.2.0 release. Thoughts?

[sonarqubecloud]

Quality Gate passed

Issues
2 New issues
0 Accepted issues

Measures
0 Security Hotspots
0.0% Coverage on New Code
0.0% Duplication on New Code

See analysis details on SonarCloud

[rdmark]

Some additional investigation needed first...

[ghost]

@rdmark, I don't have any issue with this PR but IMHO the CVE regression fixes that were pushed to main have never been as simple and clean as @andychen-syno's original PR that was never accepted:

9e864d1

I use it on my fork and it produces a clean log with Mac OS 9 clients (unlike current main), none of these invalid metadata errors, and also allows for icon substitution for the AFP share as it should with this client OS. @andychen-syno subsequently force-pushed a mod to this commit that allowed it to be merged after @anodos325's commits were merged with main.

[rdmark]

@dgsga I'm open to suggestions!

In main we have both Andrew's and Andy's patches, the latter applied on top of the former.

The former was merged with #174 and the latter with #178

I'm all for revisiting the solution, since we're still facing this apparent fragility in that area...

[ghost]

I'll prepare a draft PR...
UPDATE: PR #1007

[rdmark]

When working on the 3.2.0 release notes, I was reminded that we have a fix for macOS generated EA metadata with #575 – closing this PR for now while evaluating if this is enough to resolve users' issues.