Taylor Blau <me@xxxxxxxxxxxx> writes: >> diff --git a/Documentation/git-notes.txt b/Documentation/git-notes.txt >> index ced2e8280e..9fc1979531 100644 >> --- a/Documentation/git-notes.txt >> +++ b/Documentation/git-notes.txt >> @@ -223,7 +223,7 @@ are taken from notes refs. A notes ref is usually a branch which >> contains "files" whose paths are the object names for the objects >> they describe, with some directory separators included for performance >> reasons footnote:[Permitted pathnames have the form >> -'ab'`/`'cd'`/`'ef'`/`'...'`/`'abcdef...': a sequence of directory >> +'12'`/`'34'`/`'56'`/`'...'`/`'789abc...': a sequence of directory > > I had to read this twice to figure out why the first 'ab' changed to > '12'. It appears that this is to avoid having to use 'gh...' in the > last directory, since 'g', 'h' and so on aren't hexadecimal digits. You can wrap-around to '0' after counting up to 'f', no ;-)? Having '/.../' between '56' and '789' to indicate "there are more levels possible here" is somewhat misleading with the new example. We could argue that the original objectname does not have to be "123456789abc.." but then the whole exercise becomes somewhat pointless as the objectname could have been 'abcdef...abcdef...'. Another minor nit: it probably makes it read more natural to start counting from '0' when writing hexadecimal, if we really want to use sequence of ascending hexdigits. Using a seemingly random example 85/b4/.../808d9ee6debdf167ced3 might be less confusing, because ... >> names of two hexadecimal digits each followed by a filename with the >> rest of the object ID.]. ... I think these two lines is pleanty clear than any example, so I am OK with either the old or the updated example, but I think a seemingly random example as long as the leaf level does not share the leading hexdigits as the pathname would work the best.
