Re: I "might" have found a bug on 8.2.1 win32
От | Richard Huxton |
---|---|
Тема | Re: I "might" have found a bug on 8.2.1 win32 |
Дата | |
Msg-id | 45C36170.3010607@archonet.com обсуждение исходный текст |
Ответ на | Re: I "might" have found a bug on 8.2.1 win32 (Tony Caduto <tony_caduto@amsoftwaredesign.com>) |
Ответы |
Re: I "might" have found a bug on 8.2.1 win32
Re: I "might" have found a bug on 8.2.1 win32 |
Список | pgsql-general |
Tony Caduto wrote: > If pgAdmin III stored it's own passwords in the registry it would be up > to the user (as it should be) to use pgpass. > If they chose to use pgpass, libpq would override the passwords stored > in the registry anyway, which is what pgAdmin III is doing > automatically to my application without my or my users consent. It strikes me that the root of this is who/what is authenticating. Tony would like it to be his application (or rather libpq on behalf of his application). In fact, it is the OS-user (or rather libpq on behalf of the user). Since it is the OS-user that libpq deals with, it strikes me as right that the pgpass file should be shared by all apps. I'd expect my ssh config to be available to any ssh-based apps - scp / sftp etc, since it's clear that's user-based config. On the other hand, I'd be surprised if two email clients shared account/password details. Although, it might be nice if they did. I wonder if the Mac's keyring manages that sort of thing? And I'm endlessly irritated that my various web-browsers don't automatically share bookmarks. So - since there is at least one user who's suprised by this behaviour, we probably need two things: 1. A message that tells users the password is being remembered for *them* rather than for pgAdmin. 2. A simple way to allow an application to act independently of the user-wide configuration. Or perhaps just better explanation in the docs. -- Richard Huxton Archonet Ltd
В списке pgsql-general по дате отправления: