viking-1.2 fails to configure due to missing symbol in libgps or faulty configure test. Reproducible: Always Steps to Reproduce: 1. emerge viking-1.2 with gpsd support Actual Results: configure error Expected Results: successful configure+build checking whether to enable Realtime GPS Tracking stuff... (cached) yes checking for gps_poll in -lgps... no configure: error: libgps is needed for Realtime GPS Tracking feature, but not found. The feature can be disable with --disable-realtime-gps-tracking objdump -T /usr/lib64/libgps.so.20.0.0|grep gps_poll yields no symbol with that name either. seems that libgps really lacks this function. Question remains: Is the function gone (after being deprecated) or is it a bug in the configure script?
Depends. What's the error in config.log ?
1) Please attach the config.log. 2) Please attach the entire build log. 3) Please post your `emerge --info' output.
Created attachment 277305 [details] emerge --info
Created attachment 277309 [details] build.log
Created attachment 277315 [details] config.log
Guess what, neither build.log nor config.log won't tell you anything more than the original post. Who could have guessed ....
(In reply to comment #1) > Depends. > > What's the error in config.log ? configure:8527: checking for gps_poll in -lgps configure:8552: x86_64-pc-linux-gnu-gcc -o conftest -O3 -march=native -pipe -Wl,-O1 -Wl,--as-needed conftest.c -lgps -lz -lm >&5 /var/tmp/portage/sci-geosciences/viking-1.2/temp/ccUdlbar.o: In function `main': conftest.c:(.text+0x7): undefined reference to `gps_poll' collect2: ld returned 1 exit status Which is obvious, since libgps does not have any symbol (and thus function) with that name, as I already pointed out.
Making a long story short: The function got dropped from gpsd and replaced. Some other stuff changed around versuion 2.9 of gpsd roughly a year ago. A minor new version of viking 1.2.1 but does not yet include the necessary changes.
And to complete this: http://sourceforge.net/tracker/?func=detail&aid=3271943&group_id=83870&atid=570954 Is a bug ID concerning the bug about the gpsd API changes. Does masking gpsd or fixing the dependency on gpsd within the viking ebuild sound reasonable to the responsible dev.
(In reply to comment #6) > Guess what, neither build.log nor config.log won't tell you anything more than > the original post. Who could have guessed .... To get your attitude problem fixed, please file a new bug report.
Oh, almost funny. Better get the chichiness fixed in the first place *D'OH*.
Fixed in main tree. Most of your investigation on the issue were useless so next time just listen what you are asked for by wranglers. The issue is around since gpsd-2.96 (gpsd-2.95 would work}, so the breakage is in main tree for ~30 days with possible workaround by downgrading from 2.96 version.
(In reply to comment #12) > Fixed in main tree. > > Most of your investigation on the issue were useless so next time just listen > what you are asked for by wranglers. That much for attitude. Anyway, ever considered that some of the demanded data imposes privacy issues? (I am talking about 'emerge --info' in the first place here) Every time an ebuild fails there's the output please attach x,y,z. - If this is considered standard procedure anyway, where is the tool-set to report a bug in a straightforward manner? > > The issue is around since gpsd-2.96 (gpsd-2.95 would work}, so the breakage is > in main tree for ~30 days with possible workaround by downgrading from 2.96 > version. So gps_poll() got deprecated and removed in 2.96? I am sorry, I didn't know that, my fault then. mea culpa.