Re: Build failure for Emacs master

[Phillip Lord]

On Thu, April 14, 2016 12:11 am, Andy Moreton wrote:
> On Wed 13 Apr 2016, Eli Zaretskii wrote:
>
>
>>> From: Andy Moreton <address@hidden>
>>> Date: Tue, 12 Apr 2016 22:49:32 +0100
>>>
>>>
>>> As another data point, I have a similar failure from the emacs-25
>>> branch building with MSYS2 64bit on Win7.
>>>
>>> On that machine, building with "make -j4" fails and produces a
>>> corrupted loaddefs.el. Deleting loaddefs.el and rebuilding with "make
>>> -j4" also
>>> fails. Deleting loaddefs.el and rebuilding with "make -j1" succeeds.
>>
>> Please debug the corruption.  I don't see the problem on any of my
>> builds, so I cannot debug this, and having all these reports without
>> anyone working on fixing the problem is very disturbing.
>
> The corrupted loaddefs.el file seems to be the right size, but the last
> 768 bytes are zero instead of the expected content.
>
>
> To debug this I have tried extracting the command from the autoload
> target in lisp/Makefile and running that standalone. This results in
> repeatably building a corrutp loaddefs.el (if it is deleted before each
> run).
>
> Running that command using "bootstrap-emacs -Q" (i.e. interactively)
> shows that the *autoload-file* buffer contains the 768 zero characters at
> end, so it appears that the corruption occurs before the buffer is written
> to disk.
>
> Any hints on how to debug `batch-update-autoloads' to diagnose this
> further would be welcome.


I haven't got this far, but I also have both Andy and Angelo's build
failures working (er, failing) reproducibly now. It also appears that my
autoload patch on master is causing a build failure (only on windows).

If no one solves it first, I will bisect this as soon as possible, but in
practice this is likely to be next week.

Phil