> On Jun 23, 2019, at 12:53 PM, Lars Ingebrigtsen wrote: > > I'm not that familiar with the comint/shell code... but this is done > in the central comint code, so it would do this for all the modes that > use comint? Couldn't that lead to problems in these other modes that > aren't doing this timestamp thing? Yes, it does apply for any comint buffer that uses comint-read-input-ring. i had originally thought this was only used by shell-mode, but you are correct that it is used by others. I’ve added defcustom handling to the patch to properly account for this. In addition, I have made some revisions to better account for the extended history prefix. Any additional help on this would be appreciate!