Re: Preview: portable dumper

[Eli Zaretskii]

> Cc: address@hidden, address@hidden
> From: Paul Eggert <address@hidden>
> Date: Sat, 3 Dec 2016 09:41:43 -0800
> 
> Eli Zaretskii wrote:
> > the last two major bugs that were detected by GC were bug#24478 and 
> > bug#24640.
> 
> I don't see how this is counting. I don't see the relationship between 
> bug#24478 
> and GC. Bug#24640 was a crash but the bug report was not marked as serious.

Crashes in GC are always serious, because you can never know what
caused them.  In that case, it turned out that reading objects with
the #n=object form was broken, quite a serious problem.

> If we're merely counting GC-related things that broke Emacs, a 
> more-recently-fixed bug related to GC was not reported to bug-gnu-emacs at 
> all; 
> it's something I fixed two days ago in the release branch, which fixed a 
> broken 
> build with the Oracle Developer Studio C compiler.

Indeed, problems that are revealed in GC are not very rare.

> On the other hand, if we're talking about truly major bugs detected by GC, 
> then 
> Emacs currently doesn't work on FreeBSD 11 arm64 due to GC-related problems 
> that 
> Daniel's patches would fix. This is Bug#24892, marked as serious, and it 
> pretty 
> clearly outranks all the other bugs mentioned in this thread.

There's more than one problem, and no single solution can fix them
all.