[Date Prev][Date Next] [Chronological] [Thread] [Top]

Re: Really Weird Problem



At 01:10 PM 1/4/99 -0000, Jon Parry-McCulloch wrote:
>I'm trying to build the latest development release on Solaris 2.6.

Development snapshots are not releases and are primarily for
reference purposes.  Active developer's are encourged to use
AnonCVS or CVSup to track latest developments.

>I need to
>do this becuase some of the changes from Release 1.1 are needed to make it
>all work properly.

Like what?

>First, if I try a clean config-make then it dies in util-int.c.

Development of -lldap reentrancy/thread-safe features are in initial
stages of development...

>It seems
>that configure thinks everything's POSIX but compiler thinks otherwise. The
>result is that for three of the _r functions, it uses the wrong prototypes.

Updates to the configure system and util-int.c were made over the
weekend to ease this issue.  Previously -devel configure only
checked for the existance of ctime_r and not it's prototype.  

>To fix this I just changed the HAVE_CTIME_R definitiions in portable.h so
>the code will build cleanly.
>
>Horror upon horror... the code compiles well enough, but the linker
>core-dumps.

The linker itself core dumped?  or a specific application.

>Anyone else seen anything like this?

>At present I'm using a slightly hackd version of Release 1.1 (includes
>ldbm_datum_init(), the cursor fix, and the ldiff-ified str_ functions.

The DB2 changes (ldbm_dataum_init and cursor changes) are in 1.1.2.
Why do you need the ldif changes?  (ldif routines are not externalized).