Re: [Evolution-hackers] Patch proposal to e-msg-composer and -attachment-bar
- From: Not Zed <notzed ximian com>
- To: smurfd <smurfd smurfnet homelinux net>
- Cc: evolution-hackers lists ximian com, Jeffrey Stedfast <fejj novell com>
- Subject: Re: [Evolution-hackers] Patch proposal to e-msg-composer and -attachment-bar
- Date: Mon, 04 Jul 2005 13:30:16 +0800
On Sun, 2005-07-03 at 22:34 +0200, smurfd wrote:
> Im abit confused, so does this all mean that the patch is okey as it is.
> Or does it mean it's not okey?!
The original is better than the last one or this one.
The problem is when discussing code there's always a lot of ways to do
the same thing. Maybe it is an experience thing, and it isn't always
good - I often see so many ways to solve even simple problems that I'm
blinded from the more straightforward ones. So in email or irc what
comes out is often a rush of partially thought-out possibilities, rarely
the ideal path, and sometimes not even a reasonable path.
Normally that process is iteratively repeated quickly in your head and
you just cull the less ideal solutions and then the more reasonable
solutions will just bubble to the surface, usually based on a compromise
of various competing variables. e.g. abstraction vs efficiency, space
vs speed, expression vs conciseness. But if you're in the middle of
something else you dont really have the time to stop your head, put it
in a different gear for anothe problem someone has asked you about,
think about it properly, then switch over again afterwards.
So I apologise if the various discussions just get you confused, they're
often thought experiments which aren't meant to go further, rather than
actual solutions.
> something like this perhaps then, i had to re-use the size_to_string()
> function aswell, since it was a private one too.
Eek no, definitely not. Having duplicated structures around is
definitely a recipe for disaster. Its also full of c99 stuff like
in-line variable declarations which are right out, and it still leaks
the output of size_to_string (at least).
Ugh, who on earth wrote all that 1e3L mess, hmm, messy. Whats so
unreadable about "1024". *shakes head* But I digress ...
I have an easier solution:
Convert the size_to_string in to a public static method on
e-msg-composer-attachment(or -bar) (i.e. just make it an external
function, say e_msg_composer_attachment_size_to_string). You dont need
to change anything but its name (this is more or less what you had in
the first patch right?).
Add a single public method to e-msg-composer-attachment-bar 'get_size'
which calculates the size of all attachments.
Then, make the new code use those (fixing leaks).
(personally i dont see why the list of attachments/etc is 'private'
anyway, it seems to come from earlier days in evolution development
where everything had to be abstracted behing a million accessors for no
apparent reason other than thats what someone learnt in a 2nd year uni
course was the right thing to do, under the grand assumption that every
object was going to be re-usable everwhere else, which isn't how things
turned out in the end).
[
Date Prev][
Date Next] [
Thread Prev][
Thread Next]
[
Thread Index]
[
Date Index]
[
Author Index]