https://bugs.kde.org/show_bug.cgi?id=402154
--- Comment #39 from Joachim Wagner <jwag...@computing.dcu.ie> --- (In reply to tagwerk19 from comment #38) > Symlinks provide a bit of an edge case :-) > [...] > Does the "mount point" idea work here? I don't know the internals of the indexer implementation so I cannot say for sure. I would have thought the current indexer calls `stat()` on every file and therefore will have no problem noticing it is on a different filesystem after following a symlink. If following symlinks would pose a problem to the current indexer this means the indexer works differently than I thought. Switching to using the mount point, filesystem ID and subvolid, I'd again have assumed these three are queried for every file to be index (using a volatile cache with stat.st_dev as the key to speed things up). If this check is performed for every file to be indexed I don't see how there would be any problem when following symlinks, other than surprising users who thought that adding a folder to "Do not search in these locations" (GUI) will exclude its contents from the index. -- You are receiving this mail because: You are watching all bug changes.
