still trouble with pci.c on i386

[b. f.]

Right, that's partly why I suggested Anton use svn.  But, as John
explained, he still finds cvs convenient for some purposes.

b.

On 12/28/09, Dag-Erling Smørgrav <des at des.no> wrote:
> "b. f." <bf1783 at googlemail.com> writes:
>> I think John is referring to my first reply to Anton, early in the
>> thread, in which I said that I thought it would be easier for Anton to
>> get help for problems with -CURRENT by using svn revision numbers,
>> because most base system developers were using that VCS.  Apparently,
>> John at least is not using svn exclusively, and is willing to look up
>> cvs revision numbers.
>
> The problem is that CVS revision numbers are per-file, while Subversion
> revision numbers are per-commit.  A single Subversion commit that
> affects twenty files will translate to twenty (file, revno) tuples that
> must be rolled back individually, unless you can figure out a date (or
> date range) that corresponds exactly to that commit and that commit
> only.
>
> DES
> --
> Dag-Erling Smørgrav - des at des.no
>