On Thu, Jan 1, 2009 at 7:33 AM, TG <tg-php@xxxxxxxxxxxxxxxxxxxxxx> wrote: > Thanks for the recommendations, everyone. I decided to give Subversion > another try, since I had a server already set up from when I tried this a > while ago. Reinstalled TortoiseSVN for Windows and got it all configured. > But I couldn't figure out how to update existing files in the repository > without checking them out. you cant. you can basically just browse, but you can also do things like branch and delete (well things always stay in svn but still). if you want to change the contents of files or do merges, youll need to have a working copy. > I tried using the RepoBrowser and I could add > new items by dragging them to the browser, but if I tried to drag something > that already existed, I got an error. never touched RepoBrowser.., i tend to stick to the cli :) > Is there any way to reimport as a new version? In particular, I'd love to > have it just import changed files. everything in svn is based around 'changesets'. its pretty easy to pull them up and apply them to working copies. for example say you wanted the diff between version 1 and 4, svn diff -r 1:4 svn://mysvnurl I'd love to also be able to select a date and get a copy of the directory in > question as it appeared on that date. easy, use svn log and grep for the date in question. once youve found that you can get a snapshot of the entire repository as it appeared on any commit from that day. I'm sure there's ways to do this with SVN, but it's eluding me. I'll keep > looking, but I'm thinking that SVN is overkill for what I need, even though > it's in the right ballpark. svn is pretty simple, but you will have to spend some time on it for it to make sense. it sounds like you want something very simple. maybe something like quilt, http://savannah.nongnu.org/projects/quilt/ is better? im sure its limited, but there is a newer system called StGit as well that uses git as the backend. The other option mentioned, FileHamster, has potential, but I'd rather > control my own server and storage requirements. Plus I didn't see a way to > either whitelist of blacklist file types. if you mean, youd like to prevent certain files from making it into the repository then w/ svn, a commit hook would be sufficient. > I really just want to back up > script files and not the PDFs and other large files. So either > whitelisting by file type (or blacklisting if necessary) or even doing it > by file size would be fine. rather than a commit hook or in tandem w/ it, most people just choose not to commit certain files to the repository out of practice. these can also be hidden from view, essentially by placing filenames / patterns via the svn:ignore property My web host supports rsync, but I don't think I've seen any version control > options for rsync. does your webhost support svn clients? you can do svn co then svn up, svn switch etc. -nathan
