bug#29220: 26.0.90; eieio-persistent-read fail to restore saved object.

[Eric Abrahamsen <eric@ericabrahamsen.net>]

pierre.techoueyres@free.fr (Pierre Téchoueyres) writes:

> Eric Abrahamsen <eric@ericabrahamsen.net> writes:
>
>> On 02/25/18 00:21 AM, Pierre Téchoueyres wrote:
>>> Eric Abrahamsen <eric@ericabrahamsen.net> writes:
>>>
>>>> pierre.techoueyres@free.fr (Pierre Téchoueyres) writes:
>>>>
>>>>> Hello Eric,
>>>>> Any news on this fix ?
>>>>
>>>> Hey, sorry again this is taking so long. I wanted to run CEDET's test
>>>> suite, as that seems to be the main library that relies on
>>>> eieio-persistent, and I wanted to make sure it passed with these
>>>> changes. I can't get CEDET to build on my machine, however (apparently
>>>> Arch has a too-new version of ImageMagick, and CEDET won't build against
>>>> it), and I haven't had any luck just faking it and running the tests
>>>> themselves -- everything's too dependent on a successful build.
>>>>
>>>> I've only got two Arch machines to play with -- if anyone could
>>>> successfully build and test CEDET using the fix/eieio-persistent branch,
>>>> that would be very helpful...
>>>>
>>>
>>> I've been able to run the "itest-batch" and "utest-batch" from the CEDET
>>> git repository (http://git.code.sf.net/p/cedet/git), but I don't know
>>> how to interpret the results. Whatever the emacs version I use (25.3 or
>>> 26.0.91 + fix), I get errors. So I don't know what to check for the
>>> fix/eieio-persistent branch.
>>
>> Bah, this is annoyingly difficult -- thank you for trying it out. I
>> suppose one terrible heuristic would be if the errors were different
>> between emacs-25, emacs-26, and fix/eieio-persistent. Would you mind
>> sending me the error output, in case anything jumps out at me?
>>
> Hope this could help you.

Thanks. They all end in "Corrupt object on disk", which is an
eieio-persist problem. I'll keep working on this.