Re: [Emacs-diffs] master d826037 3/3: Remove the need for temacs.in

[dancol]

On Apr 14, 2019 7:55 AM, Stefan Monnier <address@hidden> wrote:

>> > Paul, I think this discussion pointed out that there are no advantages
>> > to your change, while it does have disadvantages.
>> Hmm... it does have the advantage of speeding up compilation by
>> eliminating one `ld` call
> By how much?

I didn't bother to measure it, but I felt like it is a noticeable
difference in some of my use cases (because I sometimes noticed that
the make was stopped while linking temacs.in, and because it is also
often stopped while linking temacs).

> is it significant enough to countermand the disadvantages?

It all depends on your beliefs, not on technical issues, I think.

The fingerprint is not guaranteed foolproof when computed from the *.o
files and is not guaranteed foolproof when made from the temacs.in
file either. 

Isn't it? The two approaches are not equally robust. The temacs.in approach is as close as you're going to get to foolproof and future proof. Can you point out an instance, exempting hash collision, in which it fails?