bug#70036: a fix that

[João Távora]

On Fri, Apr 19, 2024 at 6:59 AM Theodor Thornhill <theo@thornhill.no> wrote:

> I agree - that's why I want to focus on improving the code, same as
> you. I just want you to understand that being mindful on how you come
> across can't really hurt.

You're right of course.  I do try to be mindful, but maybe I made a mistake.
IIUC it was the part where I said this was a "one-off and uncommon" report
that you felt was unkind and harsh.  Because you wrote that.  I absolutely
think we miscommunicated: I wasn't trying to say you have a bizarre project
setup, I was simply trying to make a statement about statistics and
the relative importance.

> There are enough conflicts around these days,

In Emacs or in the world?  If the former I'd rather not know about
them.

> let's not make code be one.

Right.  So reverting another person's code isn't that big of a deal.
You "reverted" bits of my code when doing your commits, and that's
fine.

> Yeah, that's right. I assume that when json serde disappeared as a
> bottleneck others emerge. This is the obvious one to me right now.

When my hotspot for a given reasonable profile is 4%, I'm feeling
fairly good about my code.  But yes, of course, it can be addressed.

> Sure. That will make the change you make similar in spirit to mine, yet
> with a different implementation. I'm fine with that.

With the difference that my change also still ultimately uses truename,
just very frugally.  So it still understands symlinks, doesn't
mislead the server and doesn't expose the bug I demonstrated.

João