On 09/15/2012 08:35 AM, David Engster wrote:
> David Engster writes:
>> Eric Ludlam writes:
>>> On 09/14/2012 04:10 PM, David Engster wrote:
>>>> It's working for Emacs 24.x, but still failing with Emacs 23.2. Pretty
>>> The below patch could add some debugging in to see if it is even trying
>>> to load the c templates. ei - in Emacs 23.2, is c-mode a child mode of
>>> c++ mode? If not, then something bovine/c.el is perhaps not loading?
>> It's the other way round (c++-mode should be a child of c-mode), but
>> otherwise you're right: it seems bovine/c isn't loaded on Emacs 23.2
>> during this part of the integration test. I have no idea why,
>> though. I'm afraid it is one of those dreaded buffer-setup thingies,
>> which are a nightmare to debug.
Yep, I got that backward. Thanks also for the output of the two Emacsen.
I think the attached patch will fix it. The problem is that in the
first call to fill a buffer with text using srecode, the request to load
templates is made, and the next line is to turn on semantic-mode.
I think the right fix is my patch that turns on Semantic during test
setup, plus removing calls to (semantic-mode 1) elsewhere which are
basically superfluous with my patch. Before the file rename, semantic
didn't need to be enabled, so this makes good sense as to why this happened.
If the patch works, I can make the change unless you want to do it.
I ran with this test change, and it continues to work w/ my Emacs.