On dimanche 14 décembre 2008, Luigi Carlotto wrote: > Gambas2 - versione 2.9.0 > > The single anomaly verification in a circumstance, that I will try to > describe: > > - I have created a small static class that executes the logon to a > database, executes query and returns the data in a structure much > similar one to the logic applied to the Result/Field objects; at last > the logon comes sluice. This my logic is created in order to avoid to > leave the logon open and, since the Result object maintains the data > single to opened logon, my class copy these data in an appropriate > structure of Array of objects, that it comes returned to the procedures > that use this function. > - the calling procedures use this class, that the structure returns > given, that comes used for determined scopes. > - the logons can be made on database PostgreSQL, MySQL and SQLite; the > query they come correctly executed, returning the corrected values, > unless for MySQL. Like described in the first mail, one of the query > executes a loading of information from a table of system of MySQL, used > in order to determine the structure of the database. > - some fields of this table are of type “longtext” (BLOB). > - Strangely, these fields return a value NULL, although their content is > instead a valid data. > - Verifying the Result object, immediately after the execution of the > query, I have found that the value of this field is truly NULL, and this > excludes an anomaly of my procedures. > - To ulterior I have had confirmation, constructing it a small procedure > of test, in which it executes the same operations made from my program, > and that it uses the same static class; strangely, the query executed > from the test program, it has given back a valid value in the field > “longtest”. > - The thing has made me to rise a doubt: since my program is rather > complex, he is probable exists some problem in memory (stack) with the > management of fields BLOB? > - Knowing enough language C/C++, I have given to a glance to the code > source of the library, but the only thing that I have found is that > fields BLOB come managed in way detail, but has not found details > conditions that could make me to doubt on they a wrong management. The > only thing that has come to me in mind, is perhaps that a problem of > memory with programs exists many complexes, for which this type of data > (blob) comes lost… > - I have found this single anomaly in MySQL, but it is also true that > the others dbms do not use this type of data (blob) in the system > tables. > > Some suggestion? > > Thanks
Can you send me a little mysql database that implies some data in the ROUTINES table of information_schema? -- Benoit Minisini ------------------------------------------------------------------------------ _______________________________________________ Gambas-user mailing list Gambas-user@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/gambas-user