Does anyone know how to fix this? If you mess with the signatures in the functions, the syntax checker can throw up false errors. Sometimes they will go away with recycling, but this one has been stuck even after moving the functions from environment to environment. It still saves and works fine, but there errors in the error list are super annoying.
I have noticed the same, but didnāt quite know how to describe it at the time, so I didnāt make a ticket.
The way I fixed it was to copy the function manually, ie, recreating it.
Saved, deleted the original, saved, copied the one I just made to the old name, saved, and deleted
the copy.
It was a pain but I was back in business.
It was also inconsistent, I wasnāt sure what I had done to make it happen.
Yeah, not doing that. Itās workingā¦
At some point, Iāll try to explore under the hood and see if I can find the bugā¦ but Iām not quite desperate enough for that.
And frankly, if you (@klincecum ) havenāt yet, it must be harder that I think it is to do thatā¦
No, I was busy, and needed it done now. Only happened to me twice.
Got too much actual work at the moment.
Ok. Bumping this because itās happening to me again and itās annoying the crap out of me. Anyone have any ideas how to fix this?
Comment out the function code. Delete the signature and save.
Close the editor. Reopen, recreate the signature, and save, then uncomment your function code and save.
I have also had success by commenting out the code, saving, then exporting the library and reimporting it with force, and then uncommenting the code and saving.
On prem for the win!
So whatever is looking at the signature for the syntax check isnāt looking at the same thing as the one on save. This is a screen shot after I fixed it, but if you move things around, the ParameterID and Order arenāt the same number, which is understandable. However, the syntax check was getting the field order from the parameterID instead of the order, so it was giving you an error, when it doesnāt really have one since the compiler uses the order (like it should). So I adjusted the parameter IDs to match the order number, and now the error goes away.
Now. how the heck do I present this to support and actually have it work through as a bug???
ok, I submitted a ticket. Weāll see where it goes from here.
I feel like this one they canāt deny is a bug. Itās very reproducible. Honestly I would have expected it to be fixed by now, but I just tested it in 2025.1.10 and itās still borked.
I created two functions in a library.
- func1
-
- input1 (string)
-
- input2 (string)
-
- input3 (int)
- func2
-
- input1 (string)
I took input3 and moved it to the second spot so the arguments were then (string, int, string).
func2 code is
this.ThisLib.function_1("test", 12, "test");
and it freaks out that it canāt convert int to string
LAME!
Adding a data point to this.
I had this same issue, but my ParameterID and Order values were the same.
I added a second parameter that is a string, but was getting an error saying it cannot convert String to Bool.
I went into the DB and switched the DataType from System.String to System.Boolean, refreshed my function, then switched it back to System.String, refreshed my function again and then it worked. 
UPDATE Ecf.EfxFunctionSignature
SET DataType = 'System.Boolean' /*'System.String'*/
WHERE SysRowID = '45DE4DB5-9ED7-42D1-87EE-B34500F0EB0A'
2028.1.5 
EDIT: Weāre on 11.2.400.11 which is 2023.2.11 - so might 12.2.100 be 2024.2?
12.0.100.11 = 2025.1.11 (current SaaS version)
I AM GUESSING HEREā¦
12.2.x is probably 2026.1 (May/June 2025 release), since 12.1.x is likely 2025.2 (October/November 2025 release)
I wish we had a tag for @epicor here! See how confusing your version numbers are? The best we can ever get is an educated guess. Everyone hates it!
The only One place that lays it out is on the Kinetic Download page on EpicWeb:
Under Releases:
See Carlosās better answer below.
Thatās only for already released versions. Does not help you guess on the future versions.
12.0 ?! What is that? There was no 11.0.
And 2025.2 is 12.1 ???
I have been game for defending the numbering, but come on now, theyāre just toying with us now.
I still donāt get why 2025 is 12 anything and not 11.4.
Ugh. Going to yell at that cloud now.
The reason is Linux on SaaS FAQ For Cloud Customers (Kinetic 2025.1)
