|
| From: | Paul Eggert |
| Subject: | Re: Excessive use of `eassert` |
| Date: | Fri, 26 Jan 2024 00:06:56 -0800 |
| User-agent: | Mozilla Thunderbird |
On 2024-01-24 07:09, Alan Mackenzie wrote:
The behaviour is not formally "undefined" and never has been.
Sure it is. In C, if you dereference a bad pointer the behavior is undefined, which means anything can happen. On common platforms, sometimes you'll get a crash and sometimes (as my example showed) you won't, and this is true regardless of whether symbols-with-pos-enabled is true. There has never been a guarantee in a normal build that XSYMBOL (S) crashes when S is not a symbol, and none of the recent changes have altered this fact about XSYMBOL.
| [Prev in Thread] | Current Thread | [Next in Thread] |