Reason for this is that the gorder_id "placeholder" gets deleted in case some gengo jobs already arrived. However if we have "some" jobs from gengo arriving it does not mean that they are all for the given gorder_id. If we then delete these jobs, we loose track of them, and therefore never be able to poll them, or to operate workflow over them.

Comments

blueminds’s picture

Status: Active » Needs review
StatusFileSize
new3.08 KB

See the patch... should work

berdir’s picture

Status: Needs review » Fixed

Ok, this works as a stop-gap fix. Can confirm that the comment buttons and comments are now there for all data items.

It is also quite an overhead atm. Every created job results in a callback to us with status 'available'. Every time, we request the order and process the job id's. That slows down the whole process a lot and is completely unecessary, we have all the information that we need as we're getting invoked with the job data already.

I expect that we'll change these it like this as part of the mapping refactoring: On the callback, just directly check if we have a mapping for the given job_id. If not, create it, just for that job item, as all information that we need is in the custom_data string: $tjid][$tjiid][$data_item_key. That is enough to create the mapping.

We can keep the order-based mapping check for the poll putton, where we could even go a step further and always check the order and add any possibly missing mappings.

Status: Fixed » Closed (fixed)

Automatically closed -- issue fixed for 2 weeks with no activity.