[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [O] Emacs 22 compatibility

From: Achim Gratz
Subject: Re: [O] Emacs 22 compatibility
Date: Wed, 19 Dec 2012 10:21:15 +0100
User-agent: Gnus/5.13 (Gnus v5.13) Emacs/24.2.91 (gnu/linux)

Bastien writes:
>> I just tried building. it was successful although there are lot of
>> warnings.
> Good to know it builds fine, even with all those warnings.

That's a definition of "builds fine" that I wasn't previously aware of.

As to the question "does it run", yes — sort of.  You'll get a bunch of
warning messages each time you open an Org document and I wouldn't trust
it to handle any real data.  But that's only a very superficial look
I've had and I certainly didn't use any advance functionality in Org.

As I said, I've already looked into some of these errors and there are
several affecting core functionality.

I've been ignoring when some packages are entirely missing in Emacs 22
(like VM and wanderlust) and assumed that the associated functionality
was unlikely to be used from it.  If this is expected to be normal
however, then we should check for the existence of these packages and
conditionalize on the result.

There is another bunch of warnings that is triggered by the fact that
declare-function is doing nothing in Emacs 22.  This is used extensively
in the new exporter to avoid circular requires, but there seem to be a
few places where a require or an autoload should have been used instead.

Then comes a longish list of variables and functions that Emacs 22
doesn't yet have (window-system-default-frame-alist) or have a different
signature (like using delete-directory with more than one argument or
kill-buffer with none).  Also things like the use of internal or
undocumented variables and functions (e.g. timer--time) that should be
fixed anyway.

The seemingly most serious I've found is that using a defaliased special
form (org-condition-case-unless-debug) inside a macro doesn't seem to be
recognized by the byte compiler and it tries to compile a symbol list as
a function call.  Short of yanking out the defalias and replacing with
condition-case, nothing seems to work (maybe another macro expansion
too, haven't had time for this yet).

If  I fix things enough so that the test suite will run I'll get this:

Ran 401 tests, 388 results as expected, 13 unexpected
3 expected failures

13 unexpected results:
   FAILED  ob-exp/evaluate-all-executables-in-order
   FAILED  ob-exp/export-call-line-information
   FAILED  ob-exp/exports-both
   FAILED  ob-exp/exports-inline
   FAILED  ob-exp/mixed-blocks-with-exports-both
   FAILED  ob-exp/noweb-no-export-and-exports-both
   FAILED  ob-exp/noweb-on-export
   FAILED  ob-exp/noweb-on-export-with-exports-results
   FAILED  test-org-element/headline-comment-keyword
   FAILED  test-org-element/headline-interpreter
   FAILED  test-org/beginning-of-line
   FAILED  test-org/comment-dwim
   FAILED  test-org/end-of-line

While this doesn't look so bad on first sight, it will still require a
significant effort to fix and we know that the test suite coverage is
nowhere near complete.

+<[Q+ Matrix-12 WAVE#46+305 Neuron microQkb Andromeda XTk Blofeld]>+

Factory and User Sound Singles for Waldorf rackAttack:

reply via email to

[Prev in Thread] Current Thread [Next in Thread]