Wallet API documentation different from what it actually does

[Michael Leupold]

On Thursday 13 November 2008 08:41:20 Thiago Macieira wrote:
> Michael Leupold wrote:
> >Still I'd like to get this fixed but I'm unsure about which path to take
> > and would like your opinion:
> >1) Easy: Adapt the API documentation and suggest to use hasEntry()
> >2) Medium: Insert a hasEntry() call in all of the read* client functions
> > and make the methods do what the doc says.
> >3) Proper: Add additional readPasswordNew DBus calls and make the
> > current API use those to perform what's advertised.
> >2) has the drawback of making the apps issue 3 DBus calls in case they
> > already use the hasEntry() workaround. 3) has the drawback of
> > cluttering the DBus interface with 6 extra methods.
> >What do you think?
> D-Bus roundtrips should be avoided: you should get all the information you
> need in a single call. So either we change kwalletd to do what it says
> it's going to do, or we try to circumvent another way.

I agree with the roundtrips. I was just unsure abotu it.

> You can use D-Bus errors instead of return values to signify a reading of
> a key that doesn't exist.

I tried that but unless I did something wrong I wasn't able to use a user-
defined error but only the ones already provided (might have been my error). 
Apart from that it would change the compatibility with clients circumventing 
the API. So far I put quite an amount of work into not breaking that and I 
plan to follow that in the future.

Thanks and regards,
Michael
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 189 bytes
Desc: This is a digitally signed message part.
URL: <http://mail.kde.org/pipermail/kde-core-devel/attachments/20081113/57cf9b47/attachment.sig>