Last commit breaks desktop file loading
Hi! The last commit, 3474346fba687aaf265037ca8fd9ede705727262, that doesn’t enable evil-local-mode when load-in-progress is t breaks loading of files through a desktop file at Emacs boot time. The commit message notes that Emacs loads a package into a buffer called “ *load” with load-in-progress set to t. Shouldn’t the condition then be ((and (string= (buffer-name) " *load") load-in-progress))? The only reference to “ *load” that I could find in the Emacs source was, however, “ *load*” and is the buffer used by mule in load-with-code-conversion, so the check should probably be ((string= (buffer-name) " *load*")). This problem occurs because desktop (load)s the desktop file, which creates the buffers while it’s being loaded, and, since load-in-progress is global, all created buffers will see it as t when evil-local-mode is run.
On 2012-04-25 10:40 +0200, Nikolai Weibull wrote:
The commit message notes that Emacs loads a package into a buffer called “ *load” with load-in-progress set to t. Shouldn’t the condition then be ((and (string= (buffer-name) " *load") load-in-progress))?
Yes. I have an unpushed commit in my local repository that does exactly this. I will push it after some further testing. -- Vegard
On Wed, Apr 25, 2012 at 11:20, Vegard Øye <[email protected]> wrote:
On 2012-04-25 10:40 +0200, Nikolai Weibull wrote:
The commit message notes that Emacs loads a package into a buffer called “ *load” with load-in-progress set to t. Shouldn’t the condition then be ((and (string= (buffer-name) " *load") load-in-progress))?
Yes. I have an unpushed commit in my local repository that does exactly this. I will push it after some further testing.
Great, thank you!
participants (2)
-
Nikolai Weibull -
Vegard Øye