Re: Memory usage report

[Eli Zaretskii]

> From: Lars Ingebrigtsen <larsi@gnus.org>
> Cc: emacs-devel@gnu.org
> Date: Fri, 18 Sep 2020 14:59:24 +0200
> 
> Eli Zaretskii <eliz@gnu.org> writes:
> 
> >> That can have significant memory?
> >
> > Some of them could, yes.
> 
> Do you know which ones?

bidi_cache, for one.

> > Why would an image cache take that much? we flush it from time to
> > time, and I have difficulty believing that all those users reporting
> > large memory footprints load thousands of images every hour of every
> > day.
> 
> It is unlikely that "all those users" use that many images, but some
> may.  (As you no doubt remember, calling `image-size' in a loop will
> kill Emacs with an out of memory fault.)

That loop is highly unlikely to happen in practice, certainly not
happen constantly in a long-lived session.

> >> > Which variables did you have in mind in this context?  Can you show an
> >> > example?
> >> 
> >> It'd just traverse all the variables and compute the "largest" ones.
> >
> > Again, they are included in the GC report.  So what do you expect to
> > see?
> 
> I expect to see what variables take "all that size"?  I'm not sure what
> you're asking here.

I guess I don't understand the utility of knowing, e.g., which Lisp
string in the current session is the longest one.  What would you do
with such information?

> > So, while a more detailed report might be nice to have, I don't see
> > how it would help to diagnose "leaks" of the kind we are discussing
> > now on the bug list.
> 
> I am not talking about any bug report in particular.  If I were, that is
> where I would have posted this.

Then I guess I don't understand the purpose of the features you'd like
to have.  Can you explain?