Re: pushing for a new hash, was Re: [PATCH 2/3] rebase: Add tests for console output

Jonathan Nieder <jrnieder@gmail.com> writes:

If it helps, I _can_ make any set of declarations to make it sound more official, e.g. (the remainder of) June is the "make sure our tests are ready" month where we try to eradiate uchar[20] by advancing the object-id effort, replace EMPTY_TREE_SHA1_HEX and friends with EMPTY_TREE_OID_HEX, add a build-time configuration that allows us to build a Git binary that uses a phony hash e.g. ~sha1 truncated to 16 bytes, and build Git with such a configuration and then run tests, and we concentrate on this effort without adding new things until we make the tests pass.

And make similar declarations for the remainder of the year.

But the actual input for formulating what each step does and the timeline we aim for needs to come from the list wisdom; it won't have much impact if the project's official declaration is not followed-thru, and a unilateral declaration that is pulled out of thin-air likely will not be.

One good thing is that we can prepare the framework to adopt a new hash before knowing what the exact hash function is; crypto-minded folks can work on the hash selection in parallel without waiting for the framework to become ready. And I do agree with Dscho that having crypto experts would be very helpful for the latter.