Any restrictions inside "CommandEnded"?

[Anonymous]

Not surprisingly, in many reactors, for example objectModified or pickfirstModified, you have to be careful what ObjectArx commands you use. Strategies are to just "log" things you need to do or use sendString to do them later.

Are there any similar restrictions in "CommandEnded", especially regarding using the Group Dict?

We sometimes get error logs from our customers with an error we can not reproduce. The critical code is called from commandended.

[Virupaksha_aithal · Accepted solution]

Hi,

Though “CommandEnded” is relatively safe to update the database,it is recommended to use “sendstringtoexecute” from command end call back to perform database modifications. It is always "safe" to modify the database inside a “command”.

[tbrammer]

---

@Virupaksha_aithal wrote:

Though “CommandEnded” is relatively safe to update the database,it is recommended to use “sendstringtoexecute” from command end call back to perform database modifications. It is always "safe" to modify the database inside a “command”.

---

"relatively safe" means: There are certain unsafe modifications in certain conditions that should be handled in a seperate command.

Can you give examples and explain why they are unsafe in "CommandEnded"?

[Virupaksha_aithal]

Hi,

I do not have any clear case where it fails – at present, having said this, doing any database modifications in a command is always recommended. So, if there is any opportunity to make the database modifications in a command, I would always do in a command – instead of reactor call back.

[Anonymous]

Thank you for your openness and your time. I will rewrite my code to use sendStringToExecute().