Re: Deferring native compilation finalization until idle
Stéphane Marks <[email protected]> Sat, 25 Jul 2026 23:38:58 +0200
| Newsgroups | gmane.emacs.devel |
|---|---|
| Message-ID | <CAN+1HbpgCH0CX0X7P6V6YpRon6SZniDY38qrrzveaSgqkuBXQw@mail.gmail.com> |
On Sat, Jul 25, 2026 at 4:36 PM Perry Fraser <[email protected]> wrote: > Stéphane Marks <[email protected]> writes: > > In “production,” I run one of the jimeh binaries found here > https://github.com/jimeh/emacs-builds/releases which disable library > validation so I avoid the issue entirely. > > Gah, sorry, I had misunderstood what you were saying to mean that you > couldn’t reproduce the issue whatsoever. That clarifies things. > > I am not, though I think there are some additional details that need > mentioned in the PROBLEMS file for this to be an effective solution. > (Namely, macos-disable-library-validation.entitlements needs to be a > file with the appropriate entitlement enabled in it[1] > > These entitlements apply only to the bundle Emacs.app and its main > entrypoint binaries. > > Perhaps I wasn’t clear—the codesign command currently listed in > PROBLEMS will not function on its own as it needs an entitlement file > to read from, which is what the --entitlements flag takes. > That file already exists. See nextstep/templates/Info.plist.in which you can modify per etc/PROBLEMS. Also see https://github.com/jimeh/build-emacs-for-macos/blob/387c2c40aaca4f0140569383f9cabd8ea42ee692/pkg/sign/entitlements.go#L16 for how jimeh's build process does it. > I think you’re right that additionally it would need to be codesigning > the Emacs.app bundle to have any effect, due to what I mention below. > > I don’t know if that is the case. I suggested that as an area for an > enterprising someone to study in case it turns out that running > codesign on eln files as a post-generation step worked when those > files are contained in the bundle directory tree. > > Well, I wasn’t able to get the jimeh build to load without hiccups > until I let the eln cache dir be its default of being inside the .app. > I don't see that issue when starting a jimeh build from the bundle. > If you think it’s worth confirming, I can investigate further? > If you want to know what's going on, you'll have to investigate more, yes.