Problem with haddock and binary repos in checkout from github
David Peixotto
dmp at rice.edu
Thu Jun 30 01:42:51 CEST 2011
On Jun 29, 2011, at 5:57 PM, Ian Lynagh wrote:
> On Tue, Jun 28, 2011 at 02:55:09PM -0500, David Peixotto wrote:
>>
>> Oddly, this commit does not seem to be in the github repo, even though the commit date is 2007.
>
> Perhaps github has only recently started rejecting patches with this
> error?
I was confused though because I thought that a commit in git would depend on all its ancestors reachable from the commit, so I couldn't see how this commit could be in one repo but not the other if it was actually committed in 2007.
I think what happened was that these commits were actually added recently, but since they were imports of earlier commits they kept the original commit date. The last commit in the github repo is 31fce20bf6a0191e26e640d1c464deea00c8ac4d. If we look at what comes after that commit in the haskell.org repo, I can see some 755 commits by running
$ git log 31fce20bf6a0191e26e640d1c464deea00c8ac4d..
and it looks like several of these commits are from the 2007 time period.
>> $ git show cc2f67f98fbc246f010516c71722d5c3432e0a2b
>> fatal: bad object cc2f67f98fbc246f010516c71722d5c3432e0a2b
>>
>> I'm rather confused about what is happening here.
>
> Where are you running that? It works in my local repo, FWIW.
I'm running that from a fresh github checkout. I don't have a problem if I run it from a checkout of the darcs.haskell.org version.
$ git clone git://github.com/ghc/packages-Cabal.git github-Cabal
Cloning into github-Cabal...
$ cd github-Cabal/
$ git show cc2f67f98fbc246f010516c71722d5c3432e0a2b
fatal: bad object cc2f67f98fbc246f010516c71722d5c3432e0a2b
That commit is in the haskell.org cabal repo.
More information about the Cvs-ghc
mailing list