Re: OutBox purpose (and a RadioMenuItem question)
- From: Pawel Salek <pawsa theochem kth se>
- To: hector huron scouts-es org
- cc: balsa-list gnome org
- Subject: Re: OutBox purpose (and a RadioMenuItem question)
- Date: Sat, 11 Mar 2000 15:53:21 +0100 (CET)
On 11 Mar, hector@huron.scouts-es.org wrote:
> Unless anybody has another idea for implementing sending funtions
> (which we can discuss), I'll try to have it working by the end of
> next week and them I'll start to clean the code.
In my opinion, most of the code is ok. The only crucial remark I have
(apart from the fcc issue) is that the outbox role is unclear. I was
not really able to guess it from the code and it may mean either I am
dump or the _present_ code is not cristally clear.
I would also have some comments about the code (they are minor from the
functionality point of view but pretty important for a potential
maintainer). I know that writing multithreaded code is difficult, and
writing code that can work both with and without threads is extremely
difficult but this is one more reason to write such a code with an
extreme care. The improvement I seen in this field is to define macros
for mutex locking that do proper operation if compiled in the threaded
mode but are simply no-op when compiled without threads. the present
code uses something like
#if THREAD
lock/unlock
#endif
which - in my view - is not as clear as it could be. There are also
repeated large chunks of code for initialization of HEADER mutt-side
structure from balsa-side Message. These chunks do not clean up
temporary files they have created. I have identified a number of
memory leaks: for instance, results returned by make_string_from_list
are not freed. Last but not least, the initialisation and freeing
MessageQueueItem could be isolated to respective functions, too.
As I wrote, many of these remarks are minor but I think fixing them
would improve the overall quality of the code.
I feel I have to explain myself why I got interested in the
libbalsa/send.c I am about to implement charset support for postponing
(when I start writing message in, say, Greek and postpone it, it should
have proper charset when I continue it) - and improve charset support
for sending which has presently a workaround character. I intent also to
add the possibility to postpone and continue messages that contain
attachments.
Well, it's all what I would like to say. Let's keep in touch. Hector,
please let me know when you commit the fixes. By the way is there any
way to get e-mail notification when a commit to a tree is performed?
Oh, one more thing. I have noticed that selection of shown headers
breaks sometimes shortcuts so in certain configrations one cannot exit
balsa by pressing Ctrl-Q. As far as I can tell this is a bug in
RadioMenuItem implementation (I have seen the same problem with
gnotepad+ program) but I don't follow respective discussions. Is anybody
here oriented in the development in this respect?
Pawel
[
Date Prev][
Date Next] [
Thread Prev][
Thread Next]
[
Thread Index]
[
Date Index]
[
Author Index]