From: Thamm, R. <rus...@ds...> - 2009-01-05 06:03:54
|
Hi Eric, Firstly, the ideal situation would be to have the subversion working copy on the data-base. I once looked at doing this. However, the amount of work to modify svnkit to do this is very large. Furthermore, ideally you need an XML compare algorithm rather than the standard text compare used by subversion. A second approach is to have working copy in the file system and to sync the data-base with the working copy. This approach does have some limitations but is much easier to implement. The major limitation that I can see with this approach is that it is not possible to determine automatically which copy has changed - the eXist copy or the Working Copy. Generally the file with the later date will be the modified one, but there are no guarantees. I have written a Java application to sync a group of eXist documents with a subversion working copy. However, the application uses the SOAP interface and has some features that are specifically designed for my situation. So the application would need some modification to make it generally useful. The application uses an XML compare algorithm and so doesn't care about indenting etc. There is also a visual component that provides a tree showing the XML differences between two XML documents and allows the differences to be resolved individually. The application scans the subversion working copy and the eXist directory and provides a side by side list of files with an indication of changes. The user can sync all changes with a single button click or pop-up the XML difference tree in order to view and resolve differences between a pair of files. The application also supports subversion functions, so I can commit, update, delete and add files to the working copy. If there is some interest, I am prepared to modify my application to make it more generally useful. Cheers Russell Thamm -----Original Message----- From: Eric Pugh [mailto:ep...@op...] Sent: Tuesday, 30 December 2008 9:07 AM To: exi...@li... Subject: [Exist-open] Trying to version database with Subversion.. Hi all, We've been trying to come up with a better approach for versioning our code and data in Subversion while easily editing/running it in eXist. We want to do this because we have multiple developers and content creators, and for now at least, Subversion does a great job of helping get content and code in one place. However, every time I edit a file in eXist, I have to remember to add it to Subversion. I tried mounting eXist as a webdav directory and checking out my database files from subversion to the webdav directory, but all the duplicate files in the .svn folder caused issues. Basically over taxing WebDAV as a protocol I think. So my latest crazy idea for the developers is that I can maybe checkout of Subversion into my exist data directory? In EXIST_DATA/ fs/ seems to be the files written out. I just checked out all my content to this directory, and restarted eXist, and it seemed to ignore the .svn dirs, but see the files. I edited a file through webdav, and then went into EXIST_DATA/fs/ and did svn st and saw the file was changed. So, is this a promising avenue for folks who are running eXist locally, or a rathole? Some things I am wondering is what if I create a new file, will it be added to exist? Even if this only works for modifiying existing files, it seems to be a bit easier... Eric ----------------------------------------------------- Eric Pugh | Principal | OpenSource Connections, LLC | 434.466.1467 | http://www.opensourceconnections.com ------------------------------------------------------------------------ ------ _______________________________________________ Exist-open mailing list Exi...@li... https://lists.sourceforge.net/lists/listinfo/exist-open IMPORTANT: This email remains the property of the Australian Defence Organisation and is subject to the jurisdiction of section 70 of the CRIMES ACT 1914. If you have received this email in error, you are requested to contact the sender and delete the email. |