Actually, it gets stranger and stranger!
Nanako Shiraishi nanako3-at-lavabit.com |Lists| wrote:
First, git-config.1 does not mention gitattributes except for specific
settings. core.whitespace is not one of these.
gitattributes.5 has:
note that this is ! before the *name* of the attribute, and for
"Checking whitespace errors / whitespace":
...
So now, git attributes says to unspecify a setting with '!' in front of
the attribute *name*, which for core.whitespace means use the value from
git-config, which if not specified means to error on e.g. trailing-space.
Now, I notice that git's own .gitattributes has this:
Which I assume actually works, but is invalid according to git-config.1
and gitattributes.5 because:
For the first line, neither of the values 'indent', 'trail' nor 'space'
are valid for core.whitespace, and the '!' notation is not defined in
neither gitattributes.5 nor in git-config.1 for *values*.
For the second line, gitattributes states generally about attributes that:
A value of true (unspecified would have required:"*.[ch] !whitespace")
is not valid for core.whitespace either.
So now (as I read it), neither git-config.1 nor gitattributes.5 specify
how to turn off core.whitespace "elements", but .gitattributes (from git
itself) uses one notation (!) with settings that are invalid for other
reaons, while "grep core.whitespace t/*" uses another notation (-) to
turn off elements.
Also in t/* the elements have various different names, such as: "trail",
"trailing", "trailing-space". t/t4019-diff-wserror.sh also uses this:
just to make things interesting!
It seems that there is some inconsistency about this. Perhaps
documentation (and changing t/* and .gitattributes accordingly) isn't
such a bad idea! :D
Peter
--
Peter Valdemar Mørch
http://www.morch.com
--
To unsubscribe from this list: send the line "unsubscribe git" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html