|
From: <gai...@li...> - 2003-06-22 22:02:57
|
On Sun, 2003-06-22 at 13:56, gai...@li... wrote: > The whole point of gpgme is to make it uneccesary to call gpg from system= ()=20 > calls... so I think the keyserver functions would be an appropriate reque= st. >=20 True. However, as we have found, gpgme currently lacks some aspects that we would like implemented. Making the keyserver code into its own function(s) would make it easier to replace once gpgme supports this functionality. > I see your point too. if gpgme can snag a key from the server then we rea= lly=20 > don't care how gpg behaves on default, or about it's config file. we can=20 > just keep the users gaim-e pref recorded in our own files and make gpg=20 > behave the way we want it to. >=20 I already added the necessary code to the prefs system so that we don't have to worry about modifying gpg.conf (you had an earlier concern about this). I am currently debugging the plugin code, and will be adding the ui functions as I get to that file. Currently there are two new prefs, autofetch key (boolean for enabling/disabling this ability) and keyserver (a string containing the user selected, default keyserver). > however if we have to call gpg from a system() call there's no reason not= to=20 > in my mind. we literally have to do it once per key (cause gpg will polit= ely=20 > put it in the users keyring). that's an acceptable overhead hit. >=20 > Stu hmmm...forgot about the overhead cost involved in calling system....but I think you are right, the functions should only be called once per new key. Dario |