Hi TML,
Tor Lillqvist wrote:
må 2004-12-20 klockan 17:34 +0800 skrev Simon.Zheng:
1. Possible fix for e-account.h, e-account.c e-account-list.h
e-account-list.c.
As the copies in evolution/e-util have been developed further, merge
these changes into the copy in libedataserver and drop the copies in
e-util.
I'm all for this, good work!
Does it mean these patches can be check in?
2. Possible fix for e-util.h and e-util.c
These two files are very different between evolution/e-util and
evolution-data-server/libedataserver. Thus, I think merging isn't a good
idea.
Hmm, I think it should be possible to merge these, too. (If working on
this, feel free to ignore the Win32 stuff, I'll take care of that
afterwards.)
Or at least, for those functions that are duplicated: just one copy, in
libedataserver, should be left. That would be, hmm:
e-d-s: e_util_mkdir_hier() vs evo: e_mkdir_hier(). Functionally
equivalent, although the code has diverged somewhat, more or less by
accident. The one in e-d-s uses g_mkdir_with_parents() when building
against GLib 2.8, the one in evo has had Win32 portability changed
added. The version in evo can be removed, and callers adjusted
accordingly. camel_mkdir() is a third copy of essentially the same code,
so while at it, uses of camel_mkdir() presumably should be change to
call e_util_mkdir_hier() instead.
e_util_strstrcase() vs. e_strstrcase(): Identical code. The copy in evo
could be removed, and callers adjusted.
e_strftime() (same name in both places): The copy in evo has a
significant (one character) bugfix! That fix should be applied to the
copy in e-d-s, too, and then the copy in evo removed.
e_utf8_strftime(), e_filename_make_safe() (100% identical in both
places): The copies in evo should be removed.
Thanks for your suggestion, I'll try and make some new patches for
e-util.* tomorrow.
--tml
Regards,
-Simon
|
