Skip to content

Latest commit

 

History

History
57 lines (42 loc) · 2.43 KB

TODO.md

File metadata and controls

57 lines (42 loc) · 2.43 KB
Raw

To-Do List

This file is part of extents, tools for querying and accessing file extents.

Written in 2019 by Eliah Kagan <[email protected]>.

To the extent possible under law, the author(s) have dedicated all copyright and related and neighboring rights to this software to the public domain worldwide. This software is distributed without any warranty.

You should have received a copy of the CC0 Public Domain Dedication along with this software. If not, see http://creativecommons.org/publicdomain/zero/1.0/.

This file lists TODO and FIXME items that aren't commented in the code, are likely to be missed on casual reading, or are particularly important or interesting.

To do:

  1. I believe Bash uses intmax_t for integers in arithmetic evaluation, in which case the stitch script should work with numbers from real-world disks just fine. But I need to verify that it does actually do that.

    Even assuming that is so, I should add checks to stitch that cover overflow cases, because:

    • it's supposed to be a robust parser that identifies obviously unreasonable input, and

    • in theory, the range of intmax_t could be exceeded because it's signed and __u64—defined in the Linux kernel headers and used by the FIEMAP ioctl—is unsigned. (Usually intmax_t and __u64 will be signed and unsigned types with the same size and alignment. So __u64 can usually represent some values too big for intmax_t.)

  2. Put short comments atop each function briefly describing what it does. A few functions have this; most don't. Functions prototyped in header files should have these comments there. Other functions, which are hopefully all marked static, should have those comments on their definitions, though, for them, the comments could be omitted in the few cases where they truly contribute no information.

  3. Makefile should support debug vs. release mode. (NDEBUG should not be defined in either case. If the compiler can't verify and optimize away an assertion, I want that assertion.)

  4. Makefile should support specifying whether or not the build includes support for long options (using getopt_long() instead of getopt()). Right now this can be controlled by defining or undefining a NO_LONGOPTS macro.

  5. Find out why GNU getopt() exhibits POSIXLY_CORRECT-like behavior with feature test macros weaker than _GNU_SOURCE. See the "TODO" comment in feature-test.h.