OK... If I understood you correctly, you're telling me that I can implement pessimistic locking by using transactions. The document you mentioned does tell me...
Hi Corwin, sorry, but I think I'm lost... I tried both defines, but no luck. I tried to debug it, but at some point, the DataObj seems not properly ...
... mysql, but ... I got that too. fedora 3, postgresql-odbc-7.3, postgres 7.4.8, unixODBC 2.2.9 I assumed that the postgres odbc drivers 'packaged' with...
By default, the select...update iterator uses an update of the form update tablename set value = new_value where primary_key = x; Whether this lock is...
I am trying to use the nullable<> template class to define a member in a DTL_TABLEn()-defined row class, and am getting compiler errors. I am using MS Visual...
Hi Moshe, I haven't tried using nullable in those DTL_TABLEx situations, I assume they work. I use nullables a lot, and I thought < would work. However, there...
Hi, Paul. Thank you for sending me your version of nullable.h. It is definitely different from the one I was using. When I tried yours, it eliminated the ...
Hi, Yeah I did have to change a few other things to get nullable<> working as it should. Looks like you are well behind. hmm. Personally, I would do what I...
Actually I think the code in CVS is reasonably stable at this point. What I really need to do is run and examine the test suite thoroughly and make a new...
Hello Corwin: I have added the bulk_copy code to the sql_iterator class (per your message below). For reads it executes as expected, but on inserts it will not...
Jeff, The reason the bulk_copy code from DB_select_iterator does not work for inserts is that it was designed to assume we were only fetching data, therefore...
Hi Corwin: Of course, I just assumed that the bulk_copy code was shared between the select_iterator and insert. My mistake and illustrates the problems ...
Corwin, The select_update iterator is indeed in USE_ALL_FIELDS mode and SQL Server doesn't like the resulting command when some of the original fields in the...
Vasco, Probably we should agree on terminology here - I don't really like the terms "optimistic lock" and "pessimistic lock" that much since databases tend to...
i'm jumping in without probably knowing the full story, but i've had to fix DTL where clause generation when it comes to NULL types (for the nullable<> class)....
Hi, I've already fixed this, I'll have a look at the source code in a week or so (its really busy right now), but if you look at previous emails by me, you'll...
Today I tried to compile dtl taken from cvs. I had problems compiling tests with borland free compiler. It appears that recent changes force inlining in a...
I have made a new minor release to correct the issues reported here over the last few months and put into CVS. Version 3.6.10; Released Aug 14, 2005 ...
Hello! Recently I've run in some inconsistency problem when working with dates through jtime_c class. To store current time in the table of MS SQL Server I've...
Yes, using field = (NULL) instead of field IS NULL was the problem, IIRC. So, from the latest messages, it looks like this problem has already been addressed...
I discovered an issue while compiling new release of dtl on Fedora core 2. linux with stlport 5.0. The problem is in RootException.h at class ImplErr. STLPort...
The DTL_STRUCTN classes try to define a stream operator for the struct when they create it. The error you are getting: 'operator <<' is ambiguous" is (I guess)...
Hi all, I just uploaded a little fix for nullable.h, the DTL now compiles for boost 1.30, 1.32, 1.33, none (ie no boost), and for gcc 3.0, 3.3, 4.0.1, and...
For those that are interested, there is a renewed discussion on the Boost developers' list about a database library and now some code in the sandbox CVS. It's...
Richard, Thanks for pointing this out! Corwin ... From: "Richard Jennings" <richard.jennings@...> To: <DatabaseTemplateLibrary@yahoogroups.com> Sent:...
Andrej, What happens inside DTL is that we just have a flag inside the DBConnection object that gets set to true the first time a connection is made and after ...
Note: forwarded message attached. __________________________________________________ Do You Yahoo!? Tired of spam? Yahoo! Mail has the best spam protection...
Hylke, Honestly, this is a stumper. My first instict is to call SQLGetConnectAttr just like you have done to test if the connection is active. The only other...
Why not just forget the whole thing and have your app do a simple SQL statement every hour to keep the connection alive? Otherwise I would say this is a bug in...