Zachary Dremann

Hmm, looks like -3 from lockPixels is `ANDROID_BITMAP_RESULT_ALLOCATION_FAILED` according to bitmap.h. Not entirely certain what that means. Can you give a little more information about your environment? What bitmap, what...

I agree. I've set up something here: https://github.com/Dr-Emann/android-stackblur/tree/less_garbage. I'll have to clean it up a little, and submit a pull request.

duplicate of #335

Anything doing parsing should probably be fuzzed as well

No, the first several bytes (the superblock) depend on the results of building the whole archive.

The read-only-ness of the squashfs format is kinda baked in to the format itself, squashfs optimizes for size, not updatability. Files in a squashfs archive are split into chunks, each...

Files in squashfs are represented basically as a pointer to a start block, and a list of sizes of the blocks that follow. Because of this assumption that all blocks...

I ran the benchmarks on an M1 Pro, averaged over 3 runs each. | add_many benchmark | Old | New | New (add_bulk | |------|-----|-----|-----------------| | intvlen=1 density=0.200000 order=SHUFFLE |...

Benchmarks for `contains_bulk` vs `contains_multi` (contains_bulk uses a loop over the more general `roaring_bitmap_contains_bulk`, contains_multi is calling `roaring_bitmap_contains_multi` from #371, which works only with arrays). I think the performance difference...

It looks like vs16-ci is building the tests with `NDEBUG` set, which disables asserts, and my test currently checks for side effects from the argument passed to assert. It feels...