From: Alec M. <ap...@co...> - 2005-09-28 14:53:16
|
On Wednesday 28 September 2005 06:43 am, Sylvain Th=E9nault wrote: > On Wednesday 28 September =E0 14:48, Sylvain Th=E9nault wrote: > > Hey =21 > > > > I'm having a hard time trying to figure how to make CMFEditions's > > versioning and AT objects with some fields using FileSystemStorage > > nicely play together... > > Today, the file of the working copy overwrites older version's file. > > I thought I fix this by writing a modifier and patching FSS to consid= er > > the version id, however I've encountered some problems... > > The main problem is to determine if the instance is the working copy = or > > a version. The version_id field which is added to the working copy wh= en > > saving the first version has a surprizing behaviour: it seems to be > > either 0 or the version_id of the latest rollbacked / shown(=21) vers= ion. > > Is this really expected ? IMO what would be nice would be to always h= ave > > version_id=3DNone on the working copy and version_id=3Dversion_number= on > > repository clones. Moreover showing an old version shouldn't have any= > > effect on the working copy. > > Hum, one point is wrong here: the version_id attribute of the working > copy is not changed when a version is only shown. However, if some > attributes have been added to the clone, they are copied back to the > working copy. Yes, I noticed this while doing things yesterday, very strange behavior. = It = appears that it's supposed to be undone at the end of the revert, but it'= s = not doing so. Alec |