Compilation of the port failed giving the following error message. ./obj/gomni.o: In function `OpenDevice': ./obj/gomni.o(.text+0x151): undefined reference to `bmp_map_16m_rgb_color' ./obj/gomni.o(.text+0x15b): undefined reference to `bmp_map_16m_color_rgb' gmake: *** [bin/espgs] Error 1 *** Error code 2 Fix: The Makefile supplied with the port only builds the CUPS driver, however it is dependent upon the BMP16M driver to supply the functions listed as undefined during compilation. Changing the Makefile so that it built the BMP16M driver as well resulted in the port building, installing, and operating correctly. The supplied patch enables the port to build successfully - on my system at least. How-To-Repeat: Go to the /usr/ports/print/cups-pstoraster directory and type 'make clean all'. It will take some time for the error to occur as there a number of files to be compiled prior to the file giving the error.
On Wed, Oct 09, 2002 at 07:07:47AM +0800, glastech@iinet.net.au wrote: > >>Number: 43845 >>Category: ports >>Synopsis: [PATCH] - The cups-pstoraster port does not build due to undefined functions Thanks. However, the port builds successfully on my 4.7-STABLE system w/o modification, so until either (1) someone tells me why it is failing on some systems and not others or (2) I figure the reason out, I will not modify the port. Sorry folks, but I don't "fix" a bug until I know what is really causing it. E.g., it must build on bento or else I'd have had Kris K screaming at me by now. The evidence suggests somehow an interaction between certain persons's system configuration and the ghostscript configuration. I'm not changing anything until I know what that is. That may be frustrating to some of you, however I've got 20+ years of experience successfully finding and fixing difficult softare defects, so please be patient. The objective is not just to remedy your build problem, but to also discover why you have the problem in the first place. Even though it works, the fact that it is unncessary on my 4.7-20021007 systems suggests that it is not the *correct* solution. A copy of the config.log from one of the systems where it is failing would be very helpful as a starting point, as well as the output of 'pkg_info -Ia' on the same system. Please send directly to me; there is no need to either attach it to the PR (I'll summarize cause/solution) or spam any of the mailing lists with it. Thanks much. PS. Feel free to use this patch locally on your own systems, but it will not be committed. -- Alan Eldridge Unix/C(++) IT Pro, 20 yrs, seeking new employment. (http://wwweasel.geeksrus.net/~alane/resume.txt) KDE, KDE-FreeBSD Teams (http://www.kde.org, http://freebsd.kde.org/)
State Changed From-To: open->feedback I need info to determine the true cause of the problem.
Class Changed From-To: sw-bug->change-request It's not certain that this is a bug in the port.
Responsible Changed From-To: freebsd-ports->alane I'm the maintainer.
On Wed, Oct 09, 2002 at 07:07:47AM +0800, glastech@iinet.net.au wrote: > >>Number: 43845 >>Category: ports >>Synopsis: [PATCH] - The cups-pstoraster port does not build due to undefined functions Interesting.... You have a non-ports version of glib installed, and it's finding it, and I have a ports version of glib installed, and it's not finding it. Other than that, our config.logs are the same. Yours says that support for Omni can be compiled in, apparently as a result of finding that glib, and mine says it can't, for the opposite reason. This is starting to get weird. I'm gonna patch it so it sees glib and see if I get the same error then. -- Alan Eldridge Unix/C(++) IT Pro, 20 yrs, seeking new employment. (http://wwweasel.geeksrus.net/~alane/resume.txt) KDE, KDE-FreeBSD Teams (http://www.kde.org, http://freebsd.kde.org/)
State Changed From-To: feedback->closed Found the problem. If it found a glib, which would only be a non-ports glib, it would try to build Omni(?) drivers. We turn that off at config time and no more problem. Patch committed to port Makefile.