Re: [PATCH] 21.1.4-mule: auto save doubles escape chars

What I'm leaning toward for 21.4 is Kyle's patch plus a charset sanity check, preferably at about the level of write-region-internal (or wherever it will signal on users trying to save with incompatible coding systems too). Rationale: >>>>> "Ben" == Ben Wing <ben(a)666.com&gt; writes: Ben> maybe. however, saving in many coding systems will Ben> positively lose data in many situations. Agreed. _But the user is just about to shoot himself in the foot anyway._ C-x C-s doesn't do the check you suggest, although it should. You could say "well, that's a reason to have more accuracy in the autosave" but oops! C-x C-s just deleted the autosave. Ben> also, in your scheme we'd have to save the coding system in Ben> the auto-save file. I don't see why an auto-save should have any differences from what would be produced by save-buffer at that point. The main issue AFAICT would be adding incompatible characters after the last save-buffer, when a Mule-aware user would explicitly fix the coding system. This would be taken care of by adding the pre-save check. Yes, users will often get this wrong until we provide better support for them. (For a start, suggest escape-quoted "if you don't know.") But sometimes they'll get it right, which is better than the current situation where they silently get a file full of "~" on save-buffer, and escape-quoted (which gets recovered incorrectly in VM) in auto-saves. Unless there are applications that do recover-file without user intervention. It's not obvious to me that we should attempt to support that; such an application really should be providing for its own needs. -- Institute of Policy and Planning Sciences http://turnbull.sk.tsukuba.ac.jp University of Tsukuba Tennodai 1-1-1 Tsukuba 305-8573 JAPAN Don't ask how you can "do" free software business; ask what your business can "do for" free software.