Hi, all, On the last few days I've been working on a major mode derived from org-mode. While I was at it, I've encoutered some difficulties related to those reported by the OP (Stefan). Specifically: in a lot of places, org-mode will check if it's the current major-mode and, if it's not, will refuse to perform some important task.
But this poses a problem: a derived mode will always fail the test (eq major-mode 'org-mode) So, as long as org-mode relies on this test to perform some tasks, a derived mode will have non-functional pieces and bugs. Before this patch, all the checking was centralized in the defsubst `org-mode-p'. While I was developing my mode, I just changed it to (defsubst org-mode-p () (derived-mode-p 'org-mode)) Of course, this was just a temporary workaround, but I was actually planning on submitting a patch that implemented the (already discussed) `org-mode-or-derived-mode-p' alternative, replacing all calls to `org-mode-p' with calls to it. But then comes this patch, which removes `org-mode-p' altogether. So, now, there are several instances of the always-failing test on the code. Clearly, this is a big problem for anyone writing derived modes (there doesn't seem to be a lot of people doing that, but still). So, finally, here's my question: would it be possible to change all the (eq major-mode 'org-mode) tests and replace them with (derived-mode-p 'org-mode)? Is there any reason not to do this? That's effectively what happened when I changed the defsubst and I haven't noticed any problems. After all, org-mode will continue to always pass that test -- it's just that the derived modes will start to pass it too. If no ones' against it, I think I could create the patch. I haven't signed the FSF papers, but this change falls neatly into the "repetitive" category. Cheers, -- Cássio Koshikumo At Sat, 22 Oct 2011 11:31:21 +0200, Bastien wrote: > > Hi Tassilo, > > Tassilo Horn <tass...@member.fsf.org> writes: > > > Ups, I've slightly misread your suggestion. Currently, there is only > > one place in org-src.el that check for being a mode derived from > > org-mode or org-mode itself. The second is useless, because > > (derived-mode-p 'org-mode) is true for org-mode, too. > > > > But as I've said, I've misread your suggestion and made a patch that > > completely erases org-mode-p and all its calls. Feel free to apply or > > discard it. :-) > > I applied this patch -- thanks for it. > > IMHO another reason for using (eq major-mode 'org-mode) instead of > (org-mode-p) is that new contributors are likely to use this more > explicit (though longer) expression before they find out that a > defsubst is available for that. > > -- > Bastien > >
