Expand description
Discovery of kizz.md files.
Discovery starts in one directory and never looks above it. It finds every file
named exactly kizz.md in that directory and every subdirectory, hidden
directories included, with a concurrent walk of its own on the command’s worker
pool (WorkerPool), which consults no Git or ignore-file rule, so none hides a
kizz.md. The walk reads the directories of each depth in parallel, then those of
the next depth; it starts no thread of its own.
Symbolic links to directories are followed, except a link that would make the walk
recurse into one of its own parent directories, which is skipped with a warning:
a link to a directory with the device and inode of the directory that holds it or
of any directory on the walk’s path to it. A path that cannot be read is a warning,
with the system’s error, except one that vanished during the walk, such as a
dangling symbolic link, whose name is not kizz.md. A kizz.md reached through
several paths is kept once, under the first of them in the default order of
default_order. The files and warnings are sorted, so the order in which the
walk finds them changes nothing.
Structs§
Constants§
- KIZZ_
FILE_ NAME - The name of every file kizz reads to find work.
Functions§
- default_
order - The default order of
kizz.mdpaths relative to the discovery root. - discover
- Find every
kizz.mdunderroot, which should be an absolute directory path, with a concurrent walk onpool.