Re: Refactoring hardcoded SHA-1 constants
- From
Junio C Hamano <gitster@pobox.com>
- Date
- Apr 19, 2014, 01:06 UTC
- Message-ID
- <xmqqy4z29lg0.fsf@gitster.dls.corp.google.com>
- In-Reply-To
- <CACsJy8BMuj8jcsODGLPYxakh2pMV83AqsiMb7XFNkNwdGc7NLg@mail.gmail.com>
Duy Nguyen <pclouds@gmail.com> writes:
Show 12 quoted lines
> On Sat, Apr 19, 2014 at 7:06 AM, Michael Haggerty <mhagger@alum.mit.edu> wrote: >> Let the brainstorming (and bikeshedding) begin! >> >> 1. GIT_OID_RAWSZ / GIT_OID_HEXSZ >> >> 2. OID_RAWSZ / OID_HEXSZ >> >> 3. OID_BINARY_LEN / OID_ASCII_LEN >> >> 4. BINARY_OID_LEN / ASCII_OID_LEN > > 5. sizeof(oid) / ASCII_OID_LEN
Can we safely assume sizeof(struct { uchar oid[20]; }) is 20, or on some 64-bit platforms do we have to worry about 8-byte alignment?
In any case, if the pair of names in 1. are what is already used, I do not see a need for us to waste time bikeshedding at all.
In an ideal world, an implementation of cmd_foo() that defines a variable to hold an object name and then calls into libgit.a services may link in the future with a different implementation of the services that are currently offered by libgit.a but are reimplemented on top of libgit2, right? Having the same name would help us, not hurt us, with that direction in some future, no?