openjdk11 fails during configure. configure:21890: checking that grep (/usr/bin/grep) -Fx handles empty lines in the pattern list correctly configure:21907: error: grep does not handle -Fx correctly. 11.2-STABLE FreeBSD 11.2-STABLE #22 r344491M
Here is what that looks like for me: checking that grep (/usr/bin/grep) -Fx handles empty lines in the pattern list correctly... yes This is on FreeBSD 11.2-RELEASE-p7
Hi, grep can be different: BSD or GPL. Could you please provide output of "grep -V"? Thanks!
(In reply to Michael Zhilin from comment #2) BINGO! I'd forgotten about that. $ grep -V grep (BSD grep, GNU compatible) 2.6.0-FreeBSD $ grep -i grep /etc/src.conf WITHOUT_GNU_GREP=true WITH_BSD_GREP=true
This seems like a bug in BSD grep then rather than something to do with the JDK. I'll update this accordingly.
Any quick workarounds for this? I tried installing gnugrep and setting BINARY_ALIAS= grep=gnugrep in the makefile. Still getting the error with openjdk 11 and 12 on CURRENT. checking that grep (/usr/bin/grep) -Fx handles empty lines in the pattern list correctly... configure: error: grep does not handle -Fx correctly. configure exiting with result code 1 ===> Script "configure" failed unexpectedly.
Well, temporarely swapping in a link to the gnu binary in /usr/bin, like a madman, did the trick.
I normally use BSD GREP on my 12-stable system and ran into this "grep does not handle -Fx correctly" error. I did the symlink trick referred to in comment 6 and that worked fine, but then I got hung up because openjdk11 doesn't want me to be using CCACHE. I went to /etc/make.conf and commented out "WITH_CCACHE_BUILD=yes" that was there. After running a "make clean" and "make" now still complains about the existence of CCACHE despite its being disabled in /etc/make.conf. Do I need to physically remove the ccache package from my machine in order for the configure script to be happy.
(In reply to M. Voorhis from comment #7) Oh, whoops; this thread is for FreeBSD11? I'm using '12, so my commentary should be elsewhere. Sorry about that..
Take; I think I fixed this one already (noticed with another port during PR229925 exp-run testing), but let me verify.
Original fix was committed in r352691, MFC'd to stable/12 in r353138, then just now MFC'd to stable/11 along with a whole host of other fixes in r354628. Unfortunately it will not have made it into any release yet.