On Tue, Feb 27, 2018 at 02:10:20PM -0800, Junio C Hamano wrote:
quoted
I thought it solved that by the hosting folks never seeing the strange
binary-looking data. They see only utf8, which diffs well.
Ah, OK, that is a "fix" in a wider context (in a narrower context,
"work around" is a more appropriate term ;-).
The reason why I have been nudging people toward considering in-repo
encoding attribute is because forcing projects that already have
their contents in a strange binary-looking encoding to switch is
costly. But perhaps having them pay one-time conversion pain is a
better investment longer term.
Yeah, thanks for mentioning that. It should have gone in my "relative
merits" list. The conversion flag-day is definitely going to be a pain
for users, and doesn't help with diffing older versions.
-Peff