Sylvain Beucler <[EMAIL PROTECTED]> tapota : >> The whole point is avoiding creating useless stuff -- which may >> implies many other things. Having 20000 directories with 18000 being >> empty inside one directory is not the same as having 2000 >> directories. Not for the system itself. Not for third party scripts >> that may run an rsync commadn on the directory. Not for the sysadmin >> looking for a specific directory with the ls command. >> >> Scalability will always be an issue. Creating only what is necessary >> is, IMHO, a good start. > > I see the point, thanks. > > Projects are created with features at 'On' by default which reduces > the efficiency of this way to do the job,
In fact, this is site-specific (depends on your database content). But sure, it wont make sense to create projects with feature off. > but as you said this could be improved, and the cron jobs still can > check the database to see what they should take into account. That would be good. > > > Incidentally I would like to see more operations on the projects (like > rename, delete, propagate a group type change...) in the backend, but > I guess that's something we'll discuss later - there're other > priorities right now :) Can you details what would each action mean. For rename, when it happens, I usually rename the group via the frontend, and a few hours later move the content of the previous cvs. In delete case, well the backend handle it (without touching cvs trees). -- Mathieu Roy +---------------------------------------------------------------------+ | General Homepage: http://yeupou.coleumes.org/ | | Computing Homepage: http://alberich.coleumes.org/ | | Not a native english speaker: | | http://stock.coleumes.org/doc.php?i=/misc-files/flawed-english | +---------------------------------------------------------------------+
