Re: [PATCH 0/2] Synchronous versions for asynchronous functions

[Iago Toral <itoral igalia com>]

On Mon, 12 Jul 2010 20:21:03 +0200, "Juan A. Suarez Romero"
<jasuarez igalia com> wrote:
> On Mon, 2010-07-12 at 19:10 +0200, Iago Toral wrote:
>> As I said in some other ocasion... why do we want the sync versions of
>> the API? 
>> Are they worth the extra effort/code? What kind of applications would
>> use these 
>> sync APIs? do we really want to target those applications?
>> 
> 
> Because all asynchronous functions in GIO, and also in other components
> too, have their synchronous counterparts. I think the same reasons to
> have them can be applied here.

GIO and Grilo are different things and have different targets, GIO is general 
purpose, should be useful in any kind of application, any field while  Grilo 
is very specific to one type of applications. I don't think the argument above 
by itself is a valid one.

> Besides of it, sync versions are easier to use than async versions.
> Developers should have the freedom to choose if they prefer using sync
> versions (easier but less efficient) than async versions. Very likely,
> if I would develop a non-graphical applications, pretty sure that sync
> versions would be enough for me.

Sync versions are slightly easier to use but they are also probably the *wrong* 
way to do things in 95% of the situations. I guess only apps that are not interactive 
should be interested in using these APIs, and even in that case I wonder if 
having a sync version really makes a difference (hadn't rygel-grilo had to implement 
a sync dbus call it would have no use for a sync version of the APIs, for example)

Anyway, I am not against having this either. I still don't see a very good point 
for having sync versions of the APIs but I agree it should not be a lot of work 
and it might be useful in some situations, even if they are not many, so if 
you are willing to do this I am ok with it.

Iago