Messaging around translation-string flow is incorrect

[stevejalim]

At the top of a page where the translations are being managed by Smartling, it states that changes made inside Wagtail will be overwritten at the next Smartling sync (which in real terms might be every 15 mins). However, this doesn't appear to be the case.

In the screenshot below, two extra exclamation marks were added to the translated text, but when the Smartling sync was triggered again, the string was not 'reset' back to its original translated state.

This may be because the job was complete as far as Smartling was concerned, so the strings were not re-updated, or something else.

Here you can see that the translated content still has the manually-set !!!

[zerolab]

@stevejalim just want to make sure I understand this:

1. translation (and Smartling job was created)
2. the string was translated manually in Wagtail (with 3 x !)
3. The Job was completed in Smartling (with 2 x !)
4. The translations were not synced from Smartling?

sync_smartling is configured to run on cron, right?

[stevejalim]

Ah let me clarify:

Source string from Wagtail had one ❗

Smartling translation had one ❗

Translated page was manually edited to have three❗❗❗

On next sync I was expecting, based on the message, for the translated page to be 'reset" to what was/still is in Smartling: just one ❗

[stevejalim]

This was local so no cron but I was definitely re running the sync command

[zerolab]

hmm.. in that case, the job was done, so any local changes become just that (local changes), until the next Smartling job. We don't process completed Job, as that would become quite the slog and would not scale in any way.

We should update the message correspondingly

[stevejalim]

Am I right/wrong in thinking that in this situation, if the source page changed and new translations were generated at the Smartling side, that would override the localy-changed !!! in the translated page? That seems pretty logical that it would happen in that situation, because that'd be a new Job, yep?

[zerolab]

That should be the case. And that should be fairly easy to test