Restoring a back up from OEMR 4.0 in Kubuntu 10.10 returns an error …Error: The phpGACL directory path ’ $phpgacl_location = dirname(__FILE__).’/…/gacl’;’ does not start with ‘/’.
I really do need a solution, unfortunately I upgraded before fully checking the restore functionality. Now I am left without any recourse to a restore if anything happens to my main computer. Also I cannot restore my daily backup to my laptop.
Unless you are adding another site to the tree at /var/www/openemr then you need to first delete or move that old directory out of the way (assuming, of course, you do not want to keep it).
Regarding the first error, if you answer Y when asked if you want to specify new names and locations, then you will be given a chance to provide a valid path for the phpgacl directory.
OK, I removed the old openemr directory at /var/www and did a restore after entering the full path for phpGACL directory: /var/www/openemr/gacl. It worked.
But in this method of restoring has two differences from my previous restores in 3.2:
1: Have to delete /var/www/openemr everytime I restore
2: Have to provide a pathway to phpGACL everytime I restore
Thats a lot of work for daily restores to my laptop from my office computer!
It would not be hard to customize the restore script to do both of those things automatically. This is the kind of problem that shell scripting is designed to solve.
Yeah, but the point is the restore script worked very well without modifications in version 3.2, so having to redo the whole thing is kind of counterproductive!
Restoring in 4.x is a bit different than in 3.x, due to the need to allow for multiple sites. That’s why you now have to explicitly delete the directory if you are not adding a site to it. The phpgacl thing might be a bug (not sure), but use of an external phpgacl is strongly discouraged anyway.
Alternatively, consider hiring one or more of the vendors here for your various support needs. They put much effort into improving OpenEMR and will appreciate your business.
Brady - when we patch a release, ie: 4.0.0 shouldn’t we be rolling the release patch #, ie: now it’s 4.0.3 I think.
It’s very hard for end users to tell if the source code has been patched
3 digit release # is intended for that purpose, ie: MAJOR.MINOR.PATCH
At the very least we should update the version.php
$v_major = ‘4’;
$v_minor = ‘0’;
$v_patch = ‘0’;
Maybe the branch names (in the future) should be rel-40 leaving the patch level as internal?
hey,
More thought would need to put into this. For example, isn’t a version also kept in the database and would this discrepancy break Rod’s admin.php stuff. Also, the ubuntu package for the next version would need some changes to support this. Would need time to test/debug which currently don’t have.
-brady
Rod’s stuff is about Database version not the release version, I think. Curious how this would break the Ubuntu package install? Seems very odd that the version number would have any effect on the installer since you are not generation paths that are version specific, to my knowledge.
That said I don’t ever use the installers, too limiting for my installs which are not ‘single’ instances and would never use the default paths and db names in any case. So … I don’t have an experience with them to draw on.