On 23.08.2011 22:08, Michael Terry wrote: > ede, making resume optional is a good idea (maybe a --no-resume option). But > that's a separate bug/branch.
it is, but assuming your agreement i'll open a topic on the mailing list about it. i actually vote for a --resume option, because currently it actively seems to disrupt backups of users trusting that it'll work. > > As for get_info() vs get_size(), I'm nervous about getting *all* infor we > can. That might be a lot, and some info is more or less expensive on > different backends. > > See, for example, all the info the GIO backend could request of a file: > file:///usr/share/doc/libglib2.0-doc/gio/GFileInfo.html thanks for the web link, actually i do not at all use or develop for linux desktop, therefor i am very grateful for any hints in that direction. > > I'm sure we could come up with a reasonable subset of metadata to query, but > adding more later isn't so hard. well hardcoding the info needed in the method name is not futureproof. actually i also don't wanna see a method for every info available. > > What about a get_info() method that returned a dictionary? And for an > initial pass, we would only have backends fill out a "size" entry. That way, > if we later discover some piece of metadata we want, we keep the same API and > can just edit the backends. this is a mapped list, right? i'd also suggest a path [mask?] parameter and a property list parameter. this way we can ask for get_info("path/to/file", ("size","last_mod") ) > > Now, as for whether get_info() should return information on a list or a > single file... ftp and sftp can operate on multiple files easily, but there > would be no savings to do so. We'd ideally want to verify after uploading > each file, so we're still making the same number of remote requests. And > we're not saving computationally, because getting info on the whole list is > asking for more information than we will use. i just considered the limitation of the backend protocol here. i am very aware that a backend capable of asking for the meta data of one file, should internally of course map it that way. i was just argumenting API-wise, and for the api it'd make sense to be able to return lists instead of atomic data sets. who knows what we'll need it for in the future. ede/duply.net -- https://code.launchpad.net/~mterry/duplicity/early-catch-498933/+merge/72607 Your team duplicity-team is requested to review the proposed merge of lp:~mterry/duplicity/early-catch-498933 into lp:duplicity. _______________________________________________ Mailing list: https://launchpad.net/~duplicity-team Post to : duplicity-team@lists.launchpad.net Unsubscribe : https://launchpad.net/~duplicity-team More help : https://help.launchpad.net/ListHelp