I have added a pre-release of 3.6.11 which fixes the latest gcc 4.1.2
compile errors you mentioned. It can be found at:
http://tech.groups.yahoo.com/group/DatabaseTemplateLibrary/files/DTL36/

and is labelled DTL 3.6.11.

Tests seem to run OK under linux. Once I finish running under windows
/ visual C++ I will go ahead and release this (probably next week).

Corwin


--- In DatabaseTemplateLibrary@yahoogroups.com, "Corwin Joy"
<corwinjoy@...> wrote:
>
> Hi,
> I went ahead and had a go compiling DTL under gcc 4.1.2 on redhat
> linux and got the same error. This is just gcc being stupid and not
> recognizing member functions inside of a template. The lines you are
> getting this error for look like:
>
> setstate(this->badbit)
>
> change them to
> this->setstate(this->badbit)
>
> and the error will go away.
> I got a couple other minor syntax errors related to the fact that I am
> running a 64 bit version so that I needed to change SetStmtAttrs to
> take a SQLPointer for the 2nd argument (the value argument) rather
> than SDWORD which is what I had previously. I will try to post a
> minor release with these changes next week since there have been a
> couple questions on the forum about running DTL under 64bit linux.
>
> Corwin
>
> --- In DatabaseTemplateLibrary@yahoogroups.com, "buschnick"
> <Soeren.Meyer-Eppler@> wrote:
> >
> > This is a repost since my first message two weeks ago seems to have
> > been swallowed by the moderation queue or something.
> >
> > I'm trying to compile the DTL 3.6.10 on ubuntu linux with gcc 4.2.1 (I
> > have tried several older versions as well). The process chokes on two
> > errors though. The first one was easy enough to fix, it was just a
> > repetition of the classname in a function declaration in a header.
> > Something that AFAIK is simply redundant but legal C++.
> >
> > However, the second error completely baffles me as the error message
> > doesn't seem to have anything to do with reality. gcc complains it
> > cannot convert parameter one from int to char * for function "char *
> > setstate( char *)". This is _not_ the function declaration from the
> > dtl headers, but a global function from cstdlib or math. So it seems
> > gcc is messing up the name lookup or something and finds an incorrect
> > function declaration. If I rename the dtl's setstate method to
> > setstate_dtl the error changes to something else, indicating that gcc
> > indeed can't find the method declaration at all. The necessary header
> > "DB_base.h" is included though.
> >
> > Question: did anyone else run into this problem? How can I fix this?
> > It is quite important for us to get a linux build of our product
> > working and the dtl is the last thing stopping us... So thank you for
> > any tips/pointers.
> >
> > regards,
> >
> > Sören
> >
>