IanMayo
commented
11 months ago Legitimate patterns in the data mean we get lots of false-negatives for this warning.
We need to use a different method to find absolute content that is out of sequence in the sources.
For our
run-1processing, we are spotting elements are missing a top value.This is to help me locate the files where an absolute positioned table (or image) needs to be physically moved to where the area of whitespace has been provided for it.
This test is reporting around half of the files. But, in these hundreds of failures 95% of the time there isn't actually an issue - since the absolite-positioned div just contains a single element (a table):
The above content parses satisfactorily, so I don't actually have to intervene (except to remove that warning).
I can tell by the filenames that I'm going to find the same pattern in those files. This is obscuring instances that do require my attention.
This is the scenario that I could do with being flagged up: where an element contains more than one child, and the children don't all have the same positioning strategy: both absolute, or both relevant (no
top).If we could flag those up - I can work through them.
I have introduced an example where the parent just contains one div: https://github.com/DeepBlueCLtd/LegacyMan/commit/35d39dcd82282fdfb7fedb3403d002b87c837044 . But frustratingly, this doesn't currently throw a warning anyway - though it matches the target content.