When a document contains any unicode character which is not part of the allowed range of characters configured in the Parse Settings, the following errors are reported many times in the stdlog files:
pdm_k_reindex 5980 SIGNIFICANT BPebrReindex.c 929 EBR::REINDEX::There is one document to be processed in KD.
pdm_k_reindex 5980 ERROR String.c 632 55555, Invalid characters encounter in ...
pdm_k_reindex 5980 ERROR String.c 643 Invalid characters encounter in ...
There is a fix which makes the "ERROR" messages become "SIGNIFICANT" and makes pdm_k_reindex to complete the execution, but the messages keep appearing in the logs.
When a customer is implementing his knowledge base, it is important that such messages appear so he is aware of which needs to be configured in SDM.
But in a big base which has already been implemented, whenever a user creates a document which contains a character which is not part of the Parse Settings, the messages may fill the stdlog files, making them unusable for that period.
Trying to find out which document is causing the errors is an unrewarding activity, as well as time consuming.
Increasing the stdlog file size may not minimize the impact as well.
And there may come a time when the size of the Parse Settings field will not store all the "weird" or "unpredictable" characters which may be added to SDM via a copy/paste from a Text Editor which allows font size formatting (MS Word, for example).
The idea is to have the specific messages directed to a separate log file (as the one from pdm_k_reindex creates when it's executed).