When the source is EDITED and then is re-uploaded any target that is NOT_CURRENT should be changed to PENDING but it remains the same.
LingotekContentTranslationService::updateDocument() the status of all targets were set to STATUS_REQUEST and then when the form was being built there is logic for if there is a translation for that langcode AND the status is set to STATUS_REQUEST then set status to STATUS_UNTRACKED because it means there is already a translation and this situation would and should mean that we are not handling the translation. In LingotekContentTranslationService::updateDocument() we should NOT be setting the status to STATUS_REQUEST but rather to STATUS_PENDING. And a target should only be moved to UNTRACKED if it is disassociated.
| Comment | File | Size | Author |
|---|---|---|---|
| #55 | interdiff.txt | 3.62 KB | t.murphy |
| #54 | target_set_to_untracked-2850534-54.patch | 13.42 KB | t.murphy |
| #53 | target_set_to_untracked-2850534-53.patch | 13.34 KB | t.murphy |
| #52 | target_set_to_untracked-2850534-52.patch | 10.69 KB | t.murphy |
| #50 | target_set_to_untracked-2850534-50.patch | 10.68 KB | t.murphy |
Comments
Comment #2
t.murphy commentedComment #4
t.murphy commentedComment #6
t.murphy commentedComment #8
t.murphy commentedComment #9
t.murphy commentedComment #11
t.murphy commentedComment #12
t.murphy commentedComment #15
t.murphy commentedComment #17
t.murphy commentedComment #18
penyaskitoDidn't look at this in detail yet, but at least we need more tests for config and config entities.
Comment #19
t.murphy commentedComment #21
t.murphy commentedComment #23
t.murphy commentedComment #25
t.murphy commentedComment #27
t.murphy commentedComment #29
t.murphy commentedComment #31
t.murphy commentedComment #33
t.murphy commentedComment #35
penyaskitoRemoved the ToDo and changed tests as now #2850545: Check All Translation Progress Action not updating status correctly has been committed.
Comment #36
penyaskitoThis change mades readability worse IMHO.
Same here, a single line if using short array syntax should be better.
Short array syntax here too.
What's wrong here then? Do we have an issue?
Why is this needed?
Comment #37
t.murphy commentedComment #38
t.murphy commentedThe last patch has been updated to fit the comments. The only one that doesn't have a change is #4 because the test really needs to continue farther, but the module does not follow the translation flow for the config management and so the TODO is meant to wait for the module to function as expected and then the tests are already written and ready to be added to the end of the test.
Comment #39
penyaskitoThis looks better, thanks! I will fix a couple of coding standards issues at commit (extra spaces on in_array and short syntax arrays)
Comment #40
penyaskitoInterdiff before commit.
Comment #42
penyaskitoCommitted 6ec333e and pushed to 8.x-1.x. Thanks!
Comment #44
penyaskitoCommitted fa5906e and pushed to 8.x-2.x. Thanks!
Comment #45
t.murphy commentedThis changed made it so that when there are other targets that are in the REQUEST status they also move to PENDING even though they shouldn't. This updated patch fixes that.
Comment #47
penyaskitoIf our tests didn't detect that we definitely need tests. If this patch failed, is there a problem with requirement or a big in our tests?
Comment #48
t.murphy commentedThat last test added new tests and those new tests are what failed. So I have another version that hopefully will pass. The problem was in that this scenario was not tested.
Comment #50
t.murphy commentedComment #52
t.murphy commentedComment #53
t.murphy commentedComment #54
t.murphy commentedComment #55
t.murphy commentedComment #58
penyaskitoCommitted 9039a4b and pushed to 8.x-1.x. Thanks!
Committed 3d3a26e and pushed to 8.x-2.x. Thanks!