Re: [patches] RFA: Use cross nm in NPTL tests

[Jim Blandy <jimb@xxxxxxxxxxxxxxxx>]

"Carlos O'Donell" <carlos@xxxxxxxxxxxxxxxx> writes:
> On 1/31/07, Jim Blandy <jimb@xxxxxxxxxxxxxxxx> wrote:
>> > Jim,
>> >
>> > Why are we commenting out code instead of removing it?
>>
>> Which do you think would be better?  My concern is ease of merging: a
>> deleted line doesn't look like anything, and 'svn annotate' isn't
>> helpful in finding deletions.  So a conflict between an imported
>> change and one of these changes would take more effort to sort out
>> than it would if the conditional had simply been commented out.
>
> I consider "svn annotate" nothing more than a toy. It doesn't do much
> of anything.

(Now, if you could annotate *forward* in time from a given point as
well as *backward*...)

> You are saying that "# old line" gives you context
> about the previous FSF state of the code, and through
> this context the merge will be easier?

Exactly.  That was the thought.

> I buy that.

Okay!

If someone else has strong opinions, I'm definitely open on the
question.  It's not standard practice.  But in EGLIBC, because of our
intention to continue to track the FSF sources very closely, we
generally need to prioritize merging slightly above legibility
(although our first choise is obviously "both").