From: Michael Haggerty Date: Sat, 13 Jun 2015 08:45:29 GMT Subject: Re: [PATCH 0/8] object_id part 2 Message-ID: <557BEDA9.30809@alum.mit.edu> In-Reply-To: <20150612222713.GD29305@vauxhall.crustytoothpaste.net> -----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 On 06/13/2015 12:27 AM, brian m. carlson wrote: > On Fri, Jun 12, 2015 at 03:14:25PM -0700, Junio C Hamano wrote: >> "brian m. carlson" writes: >> >>> While I did run the tests between each commit, I hadn't noticed >>> they were failing because I don't have Apache installed on my >>> laptop, so they were silently skipped. I'll resubmit with that >>> fixed. >> >> It is somewhat strange that _only_ http part had failures like >> this, and is unnerving, too, given that a few people seem to have >> given at least a cursory read over the patches and didn't spot >> anything obviously wrong. >> >> Was that because there was a single manual botch, or was that >> merely that other parts of the system do not have sufficient test >> coverage? > > It appears that I broke the change in "parse_fetch: convert to use > struct object_id" which modifies remote-curl.c, so I think it's a > single manual botch. I'm going to rework that patch anyway since > Michael said that he didn't like the idea of parse_oid_hex as it > stands, so it will end up being mostly moot. In the same email where I made those design suggestions, I also I pointed out a bug in the implementation of parse_oid_hex(). Maybe that is the reason for the test failures. Michael - -- Michael Haggerty mhagger@alum.mit.edu -----BEGIN PGP SIGNATURE----- Version: GnuPG v2 iEYEARECAAYFAlV77aYACgkQwg9mrRwfmAnCyQCeIp9aSOTtQ1ABpiSybcFQFP87 fNwAoLNhQtyFQ2/fqQIvzl1gGEEOlWxa =gI7/ -----END PGP SIGNATURE-----