Would it make sense to use the built-in before-init-time in order to determine the startup time of the current instance? It might make uptimes more accurate and less sensitive to when it is loaded during an Emacs session.

It appears that only the cl-lib bundled with the latest Emacs versions has cl-flet*: the standalone cl-lib package does not. This is likely an issue with that package, which Stefan Monnier would want to fix if one of us lets him know, but it appears that a workaround in the meantime would be for uptimes to use cl-letf* instead.

(You can see the byte compilation errors/warnings in the various jobs of this build: https://travis-ci.org/purcell/emacs.d/builds/216883163)

Hi,

In a thread in the emacs-devel maillist, the licensing situation for emacs packages provided through Emacs package archives has been under focus. I have volunteered to contact the authors of packages that have a license that is incompatible with Emacs, which is now under GPL-3+.

See https://lists.gnu.org/archive/html/emacs-devel/2017-07/msg01069.html

So I wonder if you could consider to change the license of your package to GPL-3+?

Yours,

Jean-Christophe Helary

I keep the saved uptimes outside of ~/.emacs.d, but sometimes I find that a copy of the saved file turns up there. I suspect the problem is that I use use-package to configure the location and I also have uptimes save as early as possible. I should probably drop that and just allow the auto-save to kick in when the timer fires (if auto-save is enabled).

Thinking here being: that save-on-load-of-package might sometimes kick in before the configure via use-package does its thing. I'm not 100% sure that's the case but I can't see a downside to not doing it like that.