alembic / sphinx puzzler

[Chris Lamb]

Hi James,

> I'll file a bug on sphinx's GitHub repository about the original issue 
> within the next few hours.

Thanks. Please feel free to quote my previous email, as well as link
to my WIP patch. If you do any of that, I would highly recommend
adding my braindump (starting "Why it hasn't been…"): I can't shake
the feeling something else is going on; something that the Sphinx
developers will probably have a better smell for.

Let us know when you have an issue number/URL.

> Could it be that other Sphinx documentation-generation issues tend to 
> have occluded this one?

It's always possible. But I'd actually fixed a lot of issues highly-
related to this one a couple of years ago (eg. frozenset instances),
and all the  remaining Sphinx issues were in entirely different areas
of the program.

> I also have to admit that I still don't understand what it is that 
> varies (and frequently varies, apparently) across builds that exposed 
> the problem in the first place.

Hm, isn't this just probability at work? As in, because there is 50%
chance that the 2-item set is serialised in any given order, it's only
going to be detected as unreproducible 50% of the time:

 +---------+---------+----------------+
 | Build A | Build B | Result         |
 +---------+---------+----------------+
 |  a, b   |  a, b   | "Reproducible" |
 +---------+---------+----------------+
 |  b, a   |  a, b   | Unreproducible |
 +---------+---------+----------------+
 |  a, b   |  b, a   | Unreproducible |
 +---------+---------+----------------+
 |  b, a   |  b, a   | "Reproducible" |
 +---------+---------+----------------+

That might explain why there is no "logic" to the diffoscope results
on tests.reproducible-builds.org.


Regards,

-- 
      o
    ⬋   ⬊      Chris Lamb
   o     o     reproducible-builds.org 💠
    ⬊   ⬋
      o