Closed Arnedeklerk closed 12 months ago
We discussed this ondexId issue and I thought we agreed that we don't need to hide 'ondexId'. If that's still the case, it's better that we first remove the code doing it.
In the specific case, the problem is that this search returns an empty result and replaceOndexId()
receives an empty tableData
as parameter.
Which means code is to be improved, you can't ignore such obvious edge cases, at least in the most common situations, you need to validate the parameters at some point (eg, in the API response) and control the flow or correct the output accordingly (eg, show an error, stop processing the results, turn nulls into empty values, ensure the downstream code process empty values correctly).
It also means we need to do more testing, chasing these errors every other day isn't very good. Try to use the manual testing repo more often, like every 7-10 days of significant development, propose additions and changes to them, I didn't write them for myself only.
Also, use the project:*
labels in the GH issue descriptions.
I also thought the header changing code was removed, but I've downloaded a table and see indeed Ondexid is being changed to NodeID. @lawal-olaotan, please can you start by purging that code, then we test again to start.
@marco-brandizi more testing, indeed. Just don't have much time to work on it...
@lawal-olaotan Sorry :/ but this doesn't appear to be complete. There are issues remaining here:
[x] The exported evidence.tsv (tabular format export from evidence view) is still displaying nodeId in the first column header. That code should be removed and changed back to Ondexid.
[x] Searching with a gene like "TraesCS3D01G513000" no longer loads forever, but there is no error message popup telling the user what's gone wrong. We could at least catch the error and tell the user "Sorry, no matches were found!". Searching with an incompatible/wrong gene list and a keyword also seems to be a case to test, once this is working, please. Because that loads for a different period of time but also returns null and should be caught the same way.
If we agreed on it, remove this replaceOndexId() and check edge cases better in the remaining code.
I reverted the nodeId to ondexId.
I would like to know if this error message is acceptable.
Thanks @lawal-olaotan, I'll update the error, no worries. Thanks for reverting.
One thing, @lawal-olaotan, is that this text is black where it's usually red. I was going to just fix it in-line but it would be better done the same way as the rest, if you get time...
Done and working - thanks.
As in title, there is a bug where searching with a gene in the gene list box, such as: TraesCS3D01G513000. Loads forever. Try typing randomwordsgenelist, too, and it won't load.
It might be related to: Uncaught TypeError: Cannot read properties of null (reading 'map') at replaceOndexId (data-utils.js:1010:38) at genomicViewContent (data-utils.js:220:20) at Object. (data-utils.js:179:7)
at l (jquery-bstrap.min.js:32:23304)
at Object.fireWith [as resolveWith] (jquery-bstrap.min.js:32:24066)
at C (jquery-bstrap.min.js:32:82191)
at XMLHttpRequest.n (jquery-bstrap.min.js:32:86481)