bug#43152: 28.0.50; when building emacs 28.0.50 ./temacs is stopped with core

[Paul Eggert]

On 9/3/20 4:17 PM, Angelo Graziosi wrote:
> Maybe c47be1b8440883b07b6cf918235a13b65e3d7be6 does not revert exactly the code 
> or it is the wrong revert...

It should revert exactly the right code. If reversion didn't fix the problem, 
then surely the problem is not related to the reverted code, and the reversion 
should therefore be reverted.

You reported that 7605060d51bbce88307c09bd2e9be60f2750ee3d worked and that its 
successor 2ff930d861b772466b9f6b95d1776696298f3e0b did not. Now that I look more 
carefully, though, I see that those two commits are not adjacent. (The commits 
are adjacent in the 'git log' output, but that output is a linearized version of 
a DAG.

The parent of 2ff930d861b772466b9f6b95d1776696298f3e0b is actually 
886ba068c82dcf5e0e2e1244bf99841d4ff5690c, its parent is 
6593d73928da6c9fb1ccc57930566ddd2a37c737, its parent is 
be2ef629eea4bd4a7b16f6db91aab155db3489c7, and so forth. Please bisect based on 
this ancestor relationship, not based on 'git log'. In particular, if 
886ba068c82dcf5e0e2e1244bf99841d4ff5690c fails to work then the reversion was 
surely unnecessary, and something else needs to be reverted.

You said that Emacs worked for you on August 9. If 
886ba068c82dcf5e0e2e1244bf99841d4ff5690c fails to work, I suggest bisecting 
based on master as of August 9 (commit b799cc271d69fc494da1fe04ca8ec6c529a19a19, 
say?), versus the failed commit 886ba068c82dcf5e0e2e1244bf99841d4ff5690c.