"ZheNing Hu via GitGitGadget" [off-list ref] writes:

Beyond, `--format=%(raw)` cannot be used with `--python`, `--shell`,
`--tcl`, `--perl` because if our binary raw data is passed to a variable
in the host language, the host languages may cause escape errors.

"may cause escape errors" just says you are not escaping correctly
in your code (implying that this patch is not good enough and with
more effort we should be able to fix it to allow binaries), but the
problem is the host languages may not support binaries
(specifically, anything with a NUL in it) at all, which is
fundamentally unfixable, in which case, rejecting is the only
sensible choice.

    ... because the host language may not support a NUL in the variables
    of its string type.

+The raw data in a object is `raw`, For commit and tag objects, `raw` contain

s/contain/contains/, but more importantly, as we are not introducing
%(header), I do not see why we want to talk about its details.  For
commits and tags, just like for trees and blobs, 'raw' is the raw
data in the object, so beyond "The raw data of a object is %(raw)",
I do not think there is anything to talk about.

+`header` and `contents` two parts, `header` is structured part of raw data, it
+composed of "tree XXX", "parent YYY", etc lines in commits , or composed of
+"object OOO", "type TTT", etc lines in tags; `contents` is unstructured "free
+text" part of raw object data. For blob and tree objects, their raw data don't
+have `header` and `contents` parts.

+	const char *s = buf->buf;
+	size_t cur_len = 0;
+
+	while ((cur_len != buf->len) && (isspace(*s) || *s == '\0')) {
 		s++;
+		cur_len++;

Is NUL treated the same as a whitespace letter for the purpose of
determining if a line is empty?  WHY?