bug#36929: 27.0.50; Build on master broken

[Andy Moreton]

On Mon 05 Aug 2019, Paul Eggert wrote:

>> Should we maybe define CHECK_STRUCTS unconditionally to prevent issues
>> like this?
>
> Let's not, because CHECK_STRUCTS is more trouble than it's worth when doing
> serious development. See commit 9994bf17cf532f2e1d4310341da7180342202191 and
> related discussion.
>
> If you run into this problem again, feel free to do what Eli did and commit a
> change to the hash to get your --enable-checking=structs build to work again.
> I don't bother because --enable-checking=structs is a waste of time for me,
> and I recommend against it.

It is reasonable to decide that CHECK_STRUCT is not useful for you
personally. It is not reasonable to introduce patches with known
breakage or to encourage others to do so.

Either this feature is useful and all developers must keep it up to date
when making changes, or it is not useful and should be removed.

    AndyM