Re: [PATCH v1] convert: add support for 'encoding' attribute

From: Junio C Hamano <gitster@pobox.com>
Date: Dec 14, 2017, 23:01 UTC
Message-ID: <xmqq7etosymp.fsf@gitster.mtv.corp.google.com>
In-Reply-To: <C062CB10-9885-46B4-9614-65A39BD5FBE8@gmail.com>

Lars Schneider <larsxschneider@gmail.com> writes:

> That way you could control the encoding for a text file specific
> for each path similar to the "mode bits". That also means you could
> change the encoding of a file while the blob content stays the same.

That is exactly why I said that I doubt it makes sense.

When you have the same blob object (that is in UTF-8) appear at two places in a tree (or two different subtrees inside a single tree) and the two tree entries that point at the blob tells contradicting story, you would have a checkout of the same contents in two different encodings. If you have the same blob object appear in two adjacent commits at the same path, with one commit's tree recording its encoding differently from the other's, you would switch encodings when you switch branches. I doubt these are "features", but sources of confusion.

Be it a feature, or a misfeature, it is shared with the existing solution which is .gitattributes embedded in the tree, so you are not making anything better or worse by moving the information to tree entry.

What I actually expected to hear when somebody talks about "ideal" (which may not even be achievable given existing user base) was to make blob object declare its desired external representation. That would remove the two possible confusion cited above, because once you have a blob, you have everything needed to externalize it.

In any case, I do not think we'd go there anyway, so ...