Re: [PATCH 05/16] remote: remove dead code in read_branches_file()

Ramkumar Ramachandra [off-list ref] writes:

Junio C Hamano wrote:

quoted

quoted

quoted

Reminds me of the strategy to deprecate functionality in X (cf.
http://lwn.net/Articles/536520/) ;-)

Leaving dead code around to confuse readers? :\

We broke the use case to access jgarzik/netdev-2.6 only by having
jgarzik remote accidentally, and waited for quite a while (since
early 2008 until now) to see if anobody screamed.  Nobody did.  We
now know we can remove that feature.

Did we _not_ leave dead code around to confuse readers (no comment or
log message indicating intent), or am I missing something?

The largest thing you are missing is that it does not have anything
to do with my "Reminds me of ..." comment.  The silent breakage by
df93e33c is exactly "broke silently and started waiting for bug
report, which never came."  It is not about those reading the code
(only to them, dead code matters) at all, but all about users who
use the system without reading code at all.