I can not reproduce this on latest version 6.15.11+250909 locally.

I just updated to 6.15.11+250909 and it's still wrong in the same way as before.

I have attached 3 files, which i took in order:

• batchsize-config: my configured batch size (via settings -> general -> email)
• invite_participants-nothing-sent: an example with 5 participants that are not yet valid and 1 that is valid. I executed invite_participants on this survey without supplying specific participants to invite. Nothing was sent at all, as you can see in this screenshot.
• invite_participants-1-sent: i updated 1 participant somewhere in the middle (so now there are 2 valid participants). I then repeated the invite_participants call. Only the one in the middle was invited, as you can see in this screenshot.

That last participant is just not invited, as long as there are at least [batchsize] participants before it that have not received an invitation yet.

invite_participants-1-sent.png (89,172 bytes)   

invite_participants-1-sent.png (89,172 bytes)   

invite_participants-nothing-sent.png (54,004 bytes)   

invite_participants-nothing-sent.png (54,004 bytes)   

batchsize-config.png (34,457 bytes)   

batchsize-config.png (34,457 bytes)   

Any update on this? Do you now understand what I mean, and can you reproduce it? If not, do you want me to try something else?

Since this issue is unfortunately not fixed yet, I have now refactored my code to work around this issue. My code now provides specific participant-ids to be invited (instead of only the survey-id).

The change now allows me to invite the relevant participants i expect, but I was a little bit surprised by the statusmessage in the response. I guess you could have a debate over whether it's correct, but in the example from the steps to reproduce, the response is "5 left to send", even if those 5 participants cannot (currently) be invited. I would have expected "0 left to send", if no participants can actually be invited.
I feel like the original issue and this confusing (to me at least) statusmessage are both caused by the same issue: the query does not consider the validity.

@ginosupport, maybe you can provide a pull request with your updated code based on the latest version? Then developers could review and if all checks are passed integrate the fix.
@tibor.pacalat, correct me if I am wrong. I am not sure if this is the default procedure you prefer?

I'm sorry, but I do not have a pull request. I did not refactor any limesurvey code.
Instead, I refactored my own code that uses the limesurvey API and changed how that calls the different remotecontrol API endpoints. Instead of calling the endpoint to invite_participants without providing any aTokenIds (and then expecting it to invite everyone that is ready for an invite), I have decided to do that calculation of who needs an invitation on our end and supply the aTokenIds myself. So effectively I've worked around this bug now

@ginosupport: Thanks a lot for your feedback, makes sense!

@Mazi and @ginosupport, we would gladly accept the PR with bugfix for this.

Fix committed to master branch: http://bugs.limesurvey.org/plugin.php?page=Source/view&id=41221