On Fri, Mar 21, 2025 at 9:21 AM Ship Mints wrote: > On Thu, Mar 20, 2025 at 11:09 PM Stefan Kangas > wrote: > >> Ship Mints writes: >> >> > On Thu, Mar 20, 2025 at 12:57 PM Ship Mints >> wrote: >> > >> >> Text properties are added by common completion frameworks, and serve no >> >> purpose when stored in 'recentf-save-file', waste space, and slow >> loading >> >> the file. >> >> >> >> I did not add this to NEWS as I consider this non-user visible. Happy >> to >> >> add this if the maintainers disagree. >> >> >> >> I've been running recentf with 'substring-no-properties' for ages and >> >> others should benefit, by default. >> >> >> >> -Stephane >> >> >> > >> > savehist already does this implicitly. Perhaps it shouldn't even be >> > optional as I made it (it was an easy change). >> >> I think we should do this unconditionally, as the point here is to >> locate the file and nothing else. >> >> For the record, could you also provide details on which completion >> frameworks that add text properties? Do text properties cause any other >> issues than wasted space? It's enough to mention these detail here in >> the bug report, I think, no need to add it to the commit message. >> >> No need for NEWS here, I agree. >> > > Pretty much all completions, core and packages, that rely on metadata or > apply faces, add text properties to completion candidates. The variable > 'minibuffer-allow-text-properties' can be bound to nil but if you read > https://lists.gnu.org/archive/html/bug-gnu-emacs/2025-02/msg00952.html you'll > see there cases where it can't accomplish its claim. > > Revised patch attached. > Now with the bug number in the commit log...