Release Notes

Windows GUI: The image properties were not correctly updated when changing between different image files in the diff viewer. Now it's fixed.

Windows GUI: When creating a new workspace, Plastic got locked when typing a non-existing root path (for example j:) in the path textbox, and then clicked the 'browse' button. Now it's fixed.

Windows GUI: The pending changes, items and branch explorer views were not refreshed after checking changes with an update operation without conflicts. Fixed.

OS X GUI: Since we released Gluon for OS X (release 6.0.16.1779), we accidentally hid the semantic differences feature.

Now it's fixed, and the diff mode selection (text diff vs semantic diff) is visible again in the diff viewers.

In after-replicationwrite triggers, the PLASTIC_BRANCH env variable indicates the replicated branch.

There was a bug and the env var was not filled for "push" operations.

It is fixed now.

Performing a push operation from server A to server B:

* The server B wrongly executed the after-replicationread triggers.

* The server A wrongly executed the after-replicationwrite triggers.

These are the executed triggers performing a push operation from server A to server B:

* On server A: before-replicationread & after-replicationread

* On server B: before-replicationwrite & after-replicationwrite

Performing a pull operation from server A to server B, the server B wrongly executed the after-replicationread trigger. Also, the server A was wrongly executing the after-replicationwrite trigger. Now, none of them are executed.

These are the executed triggers performing a push operation from server A to server B:

* On server A: before-replicationread & after-replicationread

* On server B: before-replicationwrite & after-replicationwrite

The import package and the push operation are not properly setting the PLASTIC_BRANCH for the before-replicationwrite and after-replicationwrite triggers. As a result, when a child branch is pushed (or imported using a package) with any of those triggers registered the operation was failing.

The repository trigger filters were ignored for the before-merge trigger. It's fixed now.

Example:

If you'd like the before-merge trigger to affect just the repository 'default', you'd run the following command to register the trigger:

cm mktrigger before-merge mergetest "c:/tmp/trigger.bat" --filter=rep:default,* 

There was a log issue with requests: send time (sendt) was also wrongly used in receive time (rect). Fixed.

java cm client (a.k.a. java CLI ): Sometimes the checkin operation could throw a NullPointerException when multiple threads were used for data transferring. This could be achieved by tuning the client.conf configuration file. (It's important to remark that the integrity of the data never was affected). Fixed.