[Users] [Bug 2579] Installed headers redefine autotools-specific constants

noreply at thewildbeast.co.uk noreply at thewildbeast.co.uk
Fri Jun 29 22:21:33 CEST 2012


http://www.thewildbeast.co.uk/claws-mail/bugzilla/show_bug.cgi?id=2579





--- Comment #9 from Holger Berndt  2012-06-29 22:21:32 ---
No, in comment #6 it was offered to answer questions on -devel, but the related
mail was unfortunately ignored.

If this bug is invalid, I'd also be interested to know how to use autotools
with AC_CONFIG_HEADERS / AM_CONFIG_HEADER from a plugin. The mentioned rule of
"not redefining macros" is everything but simple (impossible?) to follow, as
the names are not chosen by the developer, but by autotools machinery, and are
thus doomed to collide.

In fact, as far as I can see, all external plugins that use this machinery
(notification, gdata, attachwarner, pdf_viewer, geolocation, perl, fancy,
python) have this problem. They just by coincidence redefine given #defines
with empty strings, about which gcc currently doesn't complain about (even
though it's likely invalid, or at the very least unclean).

Furthermore, the header might be incorrect, because Claws Mail might have been
compiled by the distributor, and a plugin might be compiled by a user, so Claws
Mail's installed config.h might not fit the actual machine.

The way I understand autotools is that it's supposed to be included in _all_
source files, before any other header, but it's not supposed to be included
from any public header, and it's also not supposed to be installed. A more
elaborate reasoning, mostly repeating what I wrote above, can be found in [1].

Now, I usually get autotools wrong, so if that's all wrong and the bug is
indeed invalid, please tell me (and Sal, and Mones, and Michal) how to fix our
code.

[1] http://inaugust.com/post/68

-- 
Configure bugmail: http://www.thewildbeast.co.uk/claws-mail/bugzilla/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug.



More information about the Users mailing list