Some IDS 9.2/4GL 7.3 strangeness

Recently updated to engine 9.20 and 4gl 7.30 and had a few suprises: ---------------------------------------------------------------------------------------------------------------------- Had this one today for the bug catalogue if it is not already in there. The insert action shown inserts into a table via the below view which accidentally referenced the same column in the physical table twice from different fields in the view. After correcting that error and updating statistics the error did not occur. ---------------------------logfile---------------------------- 14:19:37 Informix Dynamic Server 2000 Version 9.20.UC4 SoftwareSerial Number XXX#X55555 14:19:37 Assert Failed: No Exception Handler 14:19:37 Who: Session(1020, lpe@edb1, 38234, 1074804916) Thread(1048, sqlexec, 400d9318, 1) File: mtex.c Line: 413 14:19:37 Results: Exception Caught. Type: MT_EX_OS, Context: mem 14:19:37 Action: Please notify Informix Technical Support. --------------------session log---------------------------- Sess SQL Current Iso Lock SQL ISAM F.E. Id Stmt type Database Lvl Mode ERR ERR Vers 1020 INSERT dsr1 CR Not Wait -640 0 9.03 Current SQL statement : insert into af_b1_20 values ( 967119208, 967119208, "lpe", 1021, "B1", 20, 56, 1122, "123456789012345", "F24", "", "", "", "", "", "", "", "", "" ) Last parsed SQL statement : insert into af_b1_20 values ( 967119208, 967119208, "lpe", 1021, "B1", 20, 56, 1122, "123456789012345", "F24", "", "", "", "", "", "", "", "", "" ) ------------------check up on error---------------------- finderr 640 -640 QPlan sanity failure line-number. Running UPDATE STATISTICS might fix this internal error. If the error recurs, please note all circumstances and contact Informix Technical Support. --------------------------------------------------------------------------------------------------------------- also a problem concerning 'matches' and foregin letters {CLIENT_LOCALE=da_dk.8859-1 DB_LOCALE=da_dk.8859-1 Table content NOT using GLS} {drop table a;} create temp table a(b char(3)); insert into a values ('a'a'); insert into a values (''aa'); insert into a values ('aa''); select '' in front - works',* from a where b matches ''*'; select '' behind - works',* from a where b matches '*''; select '' anywhere - fails',* from a where b matches '*'*'; select 'workaround - works',* from a where b matches '*[']*'; and some trouble with CURRENT in 4GL: { CURRENT <any> TO <any> does not work, the result always equals CURRENT without limits } --------------------------------------------------------------------------------------- Question is: Are these known bugs being worked on correcting, totally new ones, or am I just doing something plain wrong here ;) //finn PS! Also had the "second select from a view when statement cache is on fails" bug - but I think that one IS wellknown (Rereading from stmt cache It cannot see that <view>.* includes <view>.<field>, and when <view>.<field> is in an order by list, select fails) Finn E. Theodorsen///theodor@inet.uni2.dk///AtCbM///Legend#60219605 J-M-B Type: INTP. Homepage: http://www.theodor.suite.dk/index.htm All advertisments sent to the above address will be treated as requests for computer support, and charged accordingly. Sending these kind of messages equals an acceptance of these terms. The minimum fee is $500.

theodor@inet.uni2.dk (Finn E. Theodorsen) wrote: > { CURRENT <any> TO <any> does not work, > the result always equals CURRENT without limits } > >--------------------------------------------------------------------------------------- This was due to our setting the GL_DATETIME environment to control the display of standard daytime variables, this affects the output of CURRENT as well, ignoring the "<any> to <any>" part every time ;( even in statements like define a daytime hour to minute a = current hour to minute Any way around this? I really don't want to track down all those displays of ordinary daytime variables who need the GL_DATETIME and set it specifically only for them. It worked in 6.04 so there should be a way to make it work again without having to stop using CURRENT <a> to <b> alltogether. //finn Finn E. Theodorsen///theodor@inet.uni2.dk///AtCbM///Legend#60219605 J-M-B Type: INTP. Homepage: http://www.theodor.suite.dk/index.htm All advertisments sent to the above address will be treated as requests for computer support, and charged accordingly. Sending these kind of messages equals an acceptance of these terms. The minimum fee is $500.