Improvements to predicate_property/2 #2902
Conversation
This one was a toughie: it turns out that using `ptr::align_of()`` was a bad idea, since the buffer in `Heap` itself is not aligned to `Heap::heap_cell_alignment()`, so `ptr::align_of()` would sometimes return lower values than expected. That made for an heisenbug: if the alignment of the heap happened to be 4, then the bug wouldn't trigger.
This reverts commit 3b58798.
mark done item
rebis-dev makes the speed difference especially apparent due to the linear scan of strings on the heap in partial_string_tail/2 which is now avoided for repositionable streams, notably files. This addresses mthom#2888 reported and analyzed by @haijinSk. Many thanks!
ENHANCED: dedicated faster branches for repositionable streams
This addresses mthom#2898 reported by @rotu. Thank you a lot! It now also supports nested module qualifications.
I expect that check_predicate_property/5 may crash or yield wrong results for variable Module in general. This must be addressed before merging.
|
The commit history makes it very hard for me to tell what's going on in this PR. I think rebis-dev must have been rebased after this PR was created. If it's not too much trouble, could you please rebase this PR? |
|
Look at the changes from just the last 2 commits: https://github.com/mthom/scryer-prolog/pull/2902/files/9acaeb38efa741633ebd76a1b05a7de19a37f8c2..afc3aa253a631d52161640338553c42ce5a3c547 |
|
@bakaq Thanks, you're right. Those commits even cherry-pick cleanly onto master. |
|
BTW, if you drop the https://github.com/mthom/scryer-prolog/actions/runs/14579523803/job/40892946221?pr=2902#step:2:62> Checking out the ref
That is, it's cleaner to fork |
The motivation for this is #2898.
However, throwing instantiation errors for variable module leads to Scryer crashing on startup. This must be addressed before merging.
If anyone has any input on this, I would greatly appreciate it. Thank you a lot!