You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Possibly, you might not be able to reproduce it at your side with this example as it has strange behaviour - if objIO1 is noted to be with type 'base_t' instead of 'dmc.base_t' then this test case ends without errors. In larger program however, with multiple types and their inherited subtypes, this is is not a protection.
Additional note: If we don't get the 'name' value from objIO1, the test case ends without errors. In general, if no string members are accessed or if we don't try to access string members from the subtypes, larger programs seem to work fine.
Considering your outstanding reaction time from #132 , I hope to hear from you soon.
Best Regards,
Yani
The text was updated successfully, but these errors were encountered:
The issue was happening when trying to access to derived type members while base type cached information was already computed (stupid unnecessary check in my previous commit).
This is now fixed.
ocilib::Object::ToString() behavior in #132 was not correct as it was not throwing an exception (number field conversion error due to an invalid offset computation).
I've fixed root cause of #132 and #133 but ocilib::Object::ToString() behavior was still not correct in case of failure.
Thus I've creae issue #134 and committed a fix for it. Thus if there is any other issues in trying to access/convert ocilib::Object members, you will get now the expected C++ exception
Hello,
This is connected to fix #132. With the test data from it I've encountered a new, problem.
After the fix it seems that when accessing string object members the current code 'accumulates' some kind of error until it crashes.
Its a bit hard to set a precise reproduction case due to it's strange behaviour.
With user 'dmc' I've created the following types:
The connection is with the same user:
Possibly, you might not be able to reproduce it at your side with this example as it has strange behaviour - if objIO1 is noted to be with type 'base_t' instead of 'dmc.base_t' then this test case ends without errors. In larger program however, with multiple types and their inherited subtypes, this is is not a protection.
Additional note: If we don't get the 'name' value from objIO1, the test case ends without errors. In general, if no string members are accessed or if we don't try to access string members from the subtypes, larger programs seem to work fine.
Considering your outstanding reaction time from #132 , I hope to hear from you soon.
Best Regards,
Yani
The text was updated successfully, but these errors were encountered: