subscribe 418642 [EMAIL PROTECTED] Hello, Michael.
I've made some tests with smart and rpm here and I could stress smart a little more. :-) What follows is a description of my actions. I will not include the errors, because I think they can be easily reproducible. If this is not true, I may post the errors. I've removed both rpm, smartpm-core and others. I guess I could get a cleaned system (yes, I've used purge). After installing smartpm (no rpm), there is no config file in /var/lib/smart. After running smart for the first time, the config file is created, detecting what your system has (deb-sys and aptsync). So, running as a non-superuser returns an error that smart could not read its config file (it is in read-only state). Removing the config file takes us one step back, so I think there is no need to purge, and install smartpm-core again. Running smart --shell works fine under non-superuser (if the config file was created using superuser). Installing rpm affects this whole process in nothing. However, if there is a rpm database, we have a problem. To create it, you can simply run rpm --initdb. As /var/lib/rpm does not exist, you gotta create it too. This is possible as superuser or giving access to /var/lib or /var/lib/rpm to some other user or group. If smart config file is already created, this again affects in nothing. But if you remove it and run smart again as superuser, it identifies the rpm database and configures a rpm-sys channel. Everything seems to work fine. Then, you remove rpm and disaster follows. smart will try to read rpm config files. If you purge rpm, then it's worse, because the database is also gone. I don't think smartpm should depend on rpm. Recommending or suggesting it would be fine, but depending on it is not the real solution for this problem. Neither removing or purging rpm should interfere with smart. I think smart should test for the existence of the database and, then, ask the user for removal or ignore the channel. There is a new upstream release, but I don't think it solves this problem. By the way, removing the channel (which does not trigger the errors) should solve the problem for the user. However, smart removal of channels does not seem to work for detected channels. Should I report another bug or should we fix the removal as a solution for this one. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
