From: Brian B. <ca...@ai...> - 2001-07-10 22:57:52
|
I just checked in all my odbc stuff which is mostly stubs at this point but it shouldn't cause any compile errors. Want to send me the last few lines of the compile when it dies on parser.h (which should be generated by yacc/bison) ? Brian On Tue, 10 Jul 2001, Michael Wood wrote: > Brian Bruns <ca...@ai...> wrote: > > > > When compiling it should detect readline or not and just not > > use it if it doesn't find it. I believe the rpm is built with > > readline enabled, but I can't imagine a modern linux distro > > not having readline, too many of the gnu tools use it. > > Yes, but not all distributions use the same version of readline, > and readline 4.1 has some small incompatibilities with previous > versions AFAIK. > > i.e. this person probably has readline-2.2 or something > installed. (RH 6.2 uses readline-2.2.1 RH 7.0 uses > readline-4.1) > > > As for glib, it is absolutely a requirement. All the hashes, > > dynamic arrays and such use it. RPMs (install the devel one > > too) should be availible or grab source from gtk.org. > [snip] > > His problems might just go away if he installs readline-devel, > glib and glib-devel and then compiles it instead of trying to > use the binaries. > > Just a couple of things. With latest CVS on a RH 6.2 box, > configure works fine, except that it complains about a missing > Makefile.in in the src/odbc directory. There's actually not > even a src/odbc directory until after ./configure is run. After > running configure, there's just an empty Makefile in src/odbc. > (I suspect it the odbc directory has not been checked into CVS?) > > Immediately after running configure, "make" fails. It tries to > run aclocal, which fails with the following error: > aclocal: /usr/share/aclocal/gconf.m4: 4: duplicated macro `AM_PATH_GCONF' > > I got around that by copying /usr/share/libtool/libltdl/aclocal.m4 > to the mdbtools directory. "make" then gets much further, but > dies on the sql stuff when it can't find parser.h. > > |