Re: case 787122: Problem after FREEING prepared stmt in Dynamic-4GL

Colin M McGrath wrote in message <71kujj$4o5$1@news.xmission.com>... > >David Williams wrote: > >> I belive under 4gl 6.x (and hence probably D4GL) you need to free the >> cursor as well.. > >Tried that, still same error -404. >Plus, given the version we have (see list below), should we be freeing both? >Isn't freeing memory twice asking for eventual core dumps? > The problem is linked with the database version and not with the pcode version. With version 5.xx databases only free prepared_cursor (or free prepared statement) can be done. If both are done, the database returns error -404. With version 7.xx databases do free prepared_cursor AND free prepared_statement. If both free's aren't done, the server eventually gives an out of memory error. In 5.xx the free releases both the cursor and the statement. I think that in 7.xx you can do this (never tried it as our code has to be 5.xx and 7.xx compatible): prepare statement from string while ... declare stat_cursor cursor for statement foreach row in stat_cursor ... end foreach free stat_cursor end while free statement with this code the database only evaluates the plan for the statement once. In our NewEra program's that work agains't both 5.xx and 7.xx databases we have free stat_cursor if db_version() not like "5%" then free statement end if showing that the problem must be within the database and not in the 4gl(d4gl) pcode interpreter. Carlos Costa e Silva.