FWD » Core Development » User Interface

Created 11252a branch for the fix, #9709-620 for the root cause of this issue

Stanislav, SecurityManager.getInstance().getUserId() never returns null value. If an account is not detected in SecurityCache sc = ctx.getCache();, then it returns "". If an account is detected, this function can return account.getSubjectId() for the process account. Can this method of EnhancedBrowseConfigManager get a null parameter from the subjectId?

   private EnhancedBrowseConfig getCachedEnhancedConfig(String userId,
                                                        BrowseKey configKey)
   {
      Map<EhBrowseConfigKey, EnhancedBrowseConfig> cfg = userId == null ? configs : userConfigs.get();
      ReadWriteLock lock = userId == null ? configsLock : userConfigsLock.get();

      lock.readLock().lock();
      try
      {
         return cfg.get(new EhBrowseConfigKey(userId, configKey));
      }
      finally
      {
         lock.readLock().unlock();
      }
   }

This code :

         // config will be saved on the client side, set the user name
         case USER_DEFAULT:
            config.userName = SecurityManager.getInstance().getUserId();
            config.ehKeyBrowseName = null;
            config.ehUniqueKey = null;
            config.ehKeyProcName = null;
            config.columnConfigs = null;
            config.ehStartColumnKeys = null;
            break;
         case BROWSE_SPECIFIC_USER:
            config.userName = SecurityManager.getInstance().getUserId();
            break;

Is there a chance for 11252a will be reviewed this week?

Sorry for delay, I'm working on an immediate-priority task. I'll review today!

Stanislav, thank you!

16443 (11252a).

Looks good! Few javadocs are missing:

private final int accountType;

private TYPE(int type)
{
   accountType = type;
}

Constantin, could you also take a quick look?

Stanislav, I rebased and fixed java doc, cleaned the code rev 16450.(11252a). Please review.

16450.(11252a).

It is good. Minor thing: I may guess that according to our coding standards enum TYPE should be called Type.

Stanislav, thank you for findings. Those issues fixed in rev 16451(11252a).

rev 16451(11252a).

It's good.

Stanislav: Can this be put into testing?

I would also like Constantin to take a quick look at the update.

Stanislav Lomany wrote:

I would also like Constantin to take a quick look at the update.

I'm OK with the changes.

Stanislav, can we merge this? Did you do any testing?

I just executed ChUI regression testing for the case. Yes, I think we can merge.

Can 11252a be merged into the trunk?

Yes.

Sorry, I messed up the task

Please disregard my messages 20 and 22.

Stanislav, can we merge this? Did you do any testing?

I checked that saving enhanced browse configurations still works. Not sure what else to check. I think we can merge.

Alex: This is ready for merge and it is one that affects the customer for whom trunk is frozen right now. Should it merge?

Actually it is Sergey to do the merge.

Actually it is Sergey to do the merge.

Sorry about that.

OK. Rebasing 11252a.

This functionality is broken again in the trunk by rev 16477 and the fix is committed into 9709d branch.