Re: Conventions on struct copying?

From: brian m. carlson <sandals@crustytoothpaste.net>
Date: Jun 20, 2014, 01:26 UTC
Message-ID: <20140620012608.GA856079@vauxhall.crustytoothpaste.net>
In-Reply-To: <5A7F0CEE2DEC4F1B825568DB8358D397@black>

On Thu, Jun 19, 2014 at 01:22:44PM -0400, Jason Pyeron wrote:

Show 11 quoted lines

> > -----Original Message-----
> > From: Junio C Hamano
> > Sent: Thursday, June 19, 2014 13:11
> > 
> > "brian m. carlson" <sandals@crustytoothpaste.net> writes:
> > 
> > > I don't know of any place we explicitly copy structs like
> > > this,...
> > 
> > which should be a reason enough.  The first concrete guideline is
> > "just imitate the existing code".

Right. The reason I asked is that most of our structures end up containing pointers, so it wouldn't make sense to do the equivalent of a memcpy on them anyways. This is just data, so I thought it might be different. I'll implement an oidcpy function to do the work.

Show 5 quoted lines

> > > but I don't know of any prohibition against it, either.
> > 
> > So now you know ;-).
> 
> To expand, on that do not trust the compiler to do deep copies.

I'm not.  The definition looks like;

  struct object_id {
    unsigned char sha1[20];
  };

I only want it to memcpy those 20 bytes; there's no pointers or anything to complicate it.

-- 
brian m. carlson / brian with sandals: Houston, Texas, US
+1 832 623 2791 | http://www.crustytoothpaste.net/~bmc | My opinion only
OpenPGP: RSA v4 4096b: 88AC E9B2 9196 305B A994 7552 F1BA 225C 0223 B187