A customer reported sporadic parsing errors for the Global Type Sequence table during endevor initialization:
TYPE 'TYPE01' DESC 'TYPE01 description'
BSTPPARS: E004 INVALID COMMAND WORDING, FOUND:
BSTPPARS: E004 INVALID COMMAND WORDING, FOUND: '.
6:36:05 B1TS011E TYPE SEQUENCE TABLE HAS ERRORS ON DSN=dsn
6:36:05 B1TS012E TYPE SEQUENCE PROCESSING IS NOW DISABLED.
The messages suggested that maybe an spurious character had squeezed in the TYPE statement which caused the parser to fail. This idea was supported by the fact that, in all cases, the same exact TYPE statement was flagged.
However, the contents of the member were correct (this point was checked by forcing an SVC dump to see the member in storage). Also, the member didn't change at all, but the error kept arising now and then.
It was not possible to determine the cause, mainly as it was not feasible to force an SVC dump at the correct routine in a production environment for a symptom that could not be reproduced at will.
However, it was observed that the affected line extended up to column 71 in the member. Although it was correct (the syntax allows up to column 72) it was the only one being that long and it was the one being flagged each time.
The errors cleared after shortening the line so that the last nonblank character was before column 71.