Automake Support for cyrus-imapd 2.5

Bron Gondwana brong at fastmail.fm
Tue May 8 18:04:11 EDT 2012


Sorry, I should have reviewed all this a while ago too!

On Tue, Apr 17, 2012 at 10:50:08PM +0200, Дилян Палаузов wrote:
> >commit "Makefile.am: merge sieve/Makefile.in"
> >
> >+       sieve/test/testExtension ....
> >
> >Not only is this line far too long, but the directory is sieve/tests/
> >not sieve/test/
> 
> I renamed all sieve/test/ to sieve/tests/ .

As I commented elsewhere, this is godawfully unmaintainable.
 

And while we're talking about maintainability.  The old Makefile.in
was just as guilty, but we really should fix this stuff:

imap_libimap_a_SOURCES = imap/annotate.c imap/annotate.h imap/append.c \
>-------imap/append.h imap/backend.c imap/backend.h imap/convert_code.c \
>-------imap/convert_code.h imap/dlist.c imap/dlist.h imap/duplicate.c \
>-------imap/duplicate.h imap/global.c imap/global.h imap/idle.c imap/idle.h \
>-------imap/idlemsg.c imap/idlemsg.h imap/imapparse.c imap/index.h \
>-------imap/mailbox.c imap/mailbox.h imap/mbdump.c imap/mbdump.h \
>-------imap/mboxkey.c imap/mboxkey.h imap/mboxlist.c imap/mboxlist.h \

That's a nightmare to maintain and generates unreadable diffs.
Unless there's a good reason otherwise, these should be one-per-line
in alphabetical order.  It makes diffs readable and branch merging
a lot less painful.

Bron.


More information about the Cyrus-devel mailing list