Skip to content
This repository has been archived by the owner on May 31, 2024. It is now read-only.

Latest commit

 

History

History
89 lines (64 loc) · 3.35 KB

TODO.md

File metadata and controls

89 lines (64 loc) · 3.35 KB

Pending early todo

  • determine why Linux build has abysmal write performance when using Badger (probably should batch writes more when using safe mode)

  • figure to describe arch

  • fix fstest failures if I feel like it (rename I do not understand, NXIO would be trivial to return)

    sudo prove -f -o -r ~/git/fstest/tests/rename

    => no errors; only when all tests are run it does error out. This implies that there is some timing constraint (possibly flush? parallel write from previous tests? that interferes)

  • fix pjdfstest failures?

  • test harder (?)

    https://stackoverflow.com/questions/21565865/filesystem-test-suites has plenty of resources, maybe some are still applicable

Pending later TODO

  • define some way to BlockBackend to index them by Status => can easily get blocks with statuses that are awkward (or come up with an algorithm that does not require magic state and instead stores its state outside fs root in a different tree)

  • could make storage/tree actually do multiple parallel read/writes; given SSDs and kernel buffer cache, it might still be faster than what it is now

Pending someday todo

  • report to Apple that their ls implementation crashes with dates far enough in future? :p

  • ask hanwen about sometimes nil out in Link call - is it bug or feature?

  • look at Windows support - dokany / winfsp?

  • figure how to get supplementary groups on OS X, provide the info to osefuxe project, and eventually use the ops' Access method

Test Summary Report (fstest on OS X)

/Users/mstenber/git/fstest/tests/open/17.t (Wstat: 0 Tests: 3 Failed: 1) Failed test: 2 ^ NXIO not returned for pipes without readers and E_NONBLOCK; noncritical

/Users/mstenber/git/fstest/tests/rename/00.t (Wstat: 0 Tests: 79 Failed: 4) Failed tests: 49, 53, 57, 61 Files=184, Tests=1944, 156 wallclock secs ( 0.84 usr 0.34 sys + 14.49 cusr 18.54 csys = 34.21 CPU) ^ rename does not touch ctime for some reason; possibly some sort of caching issue?

Test Summary Report (pjdfstest on OS X)

/Users/mstenber/git/pjdfstest/tests/chown/07.t (Wstat: 0 Tests: 132 Failed: 19) Failed tests: 7, 12, 20, 27, 32, 40, 47, 52, 60, 67, 72 80, 87, 92, 100, 107, 112, 120, 127 /Users/mstenber/git/pjdfstest/tests/rename/09.t (Wstat: 0 Tests: 2353 Failed: 16) Failed tests: 2269-2272, 2279-2281, 2284, 2289-2292, 2299-2301 2304 /Users/mstenber/git/pjdfstest/tests/rename/10.t (Wstat: 0 Tests: 2099 Failed: 8) Failed tests: 2056-2058, 2061, 2063-2065, 2068 ^ apfs has these same :p

/Users/mstenber/git/pjdfstest/tests/open/17.t (Wstat: 0 Tests: 3 Failed: 1) Failed test: 2 ^ NXIO

/Users/mstenber/git/pjdfstest/tests/open/06.t (Wstat: 0 Tests: 144 Failed: 9) Failed tests: 67-68, 70-71, 73-74, 76, 80, 84 ^ some unexpected EPERMs; kernel-side checks too strident?

/Users/mstenber/git/pjdfstest/tests/rename/00.t (Wstat: 0 Tests: 150 Failed: 7) Failed tests: 96, 100, 104, 108, 112, 116, 120 ^ ctime not being updated (deja vu)

/Users/mstenber/git/pjdfstest/tests/rename/21.t (Wstat: 0 Tests: 16 Failed: 1) Failed test: 5 ^ peculiar test; it has conditions for both success and failure of rename

Files=232, Tests=8677, 246 wallclock secs ( 1.82 usr 0.49 sys + 39.95 cusr 48.49 csys = 90.75 CPU) Result: FAIL