cbm.wikipedia wrote:

One issue with this is that if page A transcludes page B, and B trips the spam filter, then saving A will fail. To make this less bad, the spam filter should be put deeper into the parser code, so that as soon as B is parsed the spam filter is checked against it. That would allow a more detailed error message - that page A cannot be saved because page B trips the spam filter with a certain link.

mike.lifeguard+bugs wrote:

(In reply to comment #1)

One issue with this is that if page A transcludes page B, and B trips the spam
filter, then saving A will fail. To make this less bad, the spam filter should
be put deeper into the parser code, so that as soon as B is parsed the spam
filter is checked against it. That would allow a more detailed error message -
that page A cannot be saved because page B trips the spam filter with a certain
link.

That would require a more thorough rewrite, I think, and should therefore be requested separately. This might conceivably be done at the same time as bug 4459.

Isn't this how it works now? I'm sure i've seen issues with bots on en.wiki because they are failing when their is blacklisted links on the page.