First error ili1reader.h:53: error: invalid pure specifier (only `= 0' is allowed) before ';' token make[3]: *** [../o/ili1reader.o] Error 1 make[3]: *** Waiting for unfinished jobs.... then: make[3]: Leaving directory `/var/tmp/portage/sci-libs/gdal-1.3.1/work/gdal-1.3.1/ogr/ogrsf_frmts/ili' make[2]: *** [ili-target] Error 2 make[2]: Leaving directory `/var/tmp/portage/sci-libs/gdal-1.3.1/work/gdal-1.3.1/ogr/ogrsf_frmts' make[1]: *** [sublibs] Error 2 make[1]: *** Waiting for unfinished jobs.... ... and ends with /frmts/pcraster' make[1]: Leaving directory `/var/tmp/portage/sci-libs/gdal-1.3.1/work/gdal-1.3.1 /frmts' !!! ERROR: sci-libs/gdal-1.3.1 failed. Call stack: ebuild.sh, line 1621: Called dyn_compile ebuild.sh, line 973: Called qa_call 'src_compile' ebuild.sh, line 44: Called src_compile gdal-1.3.1.ebuild, line 106: Called die Reproducible: Always
Forgot to mention: These are the packages that would be merged, in order: Calculating dependencies... done! [ebuild N ] sci-libs/gdal-1.3.1 USE="geos gif gml jpeg jpeg2k odbc png python sqlite -debug -doc -fits -hdf -hdf5 -netcdf -ogdi -postgres" 0 kB
I don't recall having any such build errors with that version, but it's pretty old... Any reason why you can't use a newer version?
I just built this version with your USE flags and with no issues, on a machine with a minimal "server" install (after it emerged half a dozen dependencies). I would suggest that something on your system is broken; have you tried running revdep-rebuild lately, and verified your toolchain is properly installed (and without any eselect compiler cruft confusing the issue)?
Steve, On returning from my holidays ... I ran /usr/portage/scripts/bootstrap.sh then emerge -e system followed by emerge -eN world, and still get the same error. So even though I seem to have naively rebuilt my entire system I'm no further forward. Perhaps you could let me know what sort of "eselect compiler cruft" may be confusing the issue. And, if such cruft exists, where I can find it on my system, or what I need to do to clear it. Thanks David
A while back, I think "eselect gcc" was going to replace gcc-config, but was pulled due to problems with how it switched from one version to the next (more or less). I'm pretty sure I never invoked that eselect module, but it bit me a while back on another machine. Basically, it messed up the binaries and symlinks under /usr/x86_64-pc-linux-gnu/ (or whatever your arch is). Check the subdirs in the above directory tree and make sure the links make sense. Only one version of gcc and binutils should be active, and you shouldn't see any weird binaries mixed in with a dir full of symlinks. You may also want to invoke "gcc-config" a couple of times to switch versions and make sure you're set for the version you think you are.
I'm assuming you have this resolved; please open an appropriate bug if you feel you need to (ie, toolchain, gdal, etc).
*** Bug 209862 has been marked as a duplicate of this bug. ***
I'm getting the same error with 1.3.1 which is currently the latest stable version. 1.3.2 compiles fine. Maybe its time stabilize one of the later versions?