[exec] In file included from ./x10aux/double_utils.h:16:0, [exec] from ./x10aux/basic_functions.h:18, [exec] from ./x10aux/class_cast.h:20, [exec] from ./x10rt.h:11, [exec] from gen/x10/lang/OutOfMemoryError.h:4, [exec] from x10aux/alloc.cc:20: [exec] ./x10aux/math.h: In function 'bool x10aux::math::isinf(float)': [exec] ./x10aux/math.h:55:48: error: '::isinff' has not been declared [exec] inline type name(float v) { return ::name##f(v); } \ [exec] ^ [exec] ./x10aux/math.h:67:9: note: in expansion of macro 'ALIAS_MATH_FUNC1' [exec] ALIAS_MATH_FUNC1(bool, isinf); [exec] ^~~~~~~~~~~~~~~~ [exec] ./x10aux/math.h: In function 'bool x10aux::math::isfinite(float)': [exec] ./x10aux/math.h:55:48: error: '::isfinitef' has not been declared [exec] inline type name(float v) { return ::name##f(v); } \ [exec] ^ [exec] ./x10aux/math.h:83:9: note: in expansion of macro 'ALIAS_MATH_FUNC1' [exec] ALIAS_MATH_FUNC1(bool, isfinite); [exec] ^~~~~~~~~~~~~~~~ [exec] ./x10aux/math.h: In function 'int x10aux::math::signbit(float)': [exec] ./x10aux/math.h:55:48: error: '::signbitf' has not been declared [exec] inline type name(float v) { return ::name##f(v); } \ [exec] ^ [exec] ./x10aux/math.h:85:9: note: in expansion of macro 'ALIAS_MATH_FUNC1' [exec] ALIAS_MATH_FUNC1(int, signbit); [exec] ^~~~~~~~~~~~~~~~ [exec] x10aux/alloc.cc: At global scope: [exec] x10aux/alloc.cc:128:16: warning: 'ssize_t mygetline(char**, size_t*, FILE*)' defined but not used [-Wunused-function] [exec] static ssize_t mygetline (char **lineptr, size_t *sz, FILE *f) [exec] ^~~~~~~~~ [exec] gmake[1]: *** [Makefile:264: x10aux/alloc.o] Error 1 build log: http://sprunge.us/YJGj
Is this a real problem right now, i.e. is there a need to build with GCC 6? This port is way behind and I can't prioritize updating it right now. I would suggest reexamining this at a later date when the port is upgraded to the latest stable version.
(In reply to Jason Bacon from comment #1) > Is this a real problem right now, i.e. is there a need to build with GCC 6? Once bug 219275 lands (probably not soon due to other blockers) this port may end up marked BROKEN.
It is part of a general sweep to upgrade the default ports GCC away from 5. So it is not an immediate priority but still something that should be looked at sometime in the near future.
(In reply to Jason Bacon from comment #1) > Is this a real problem right now, i.e. is there a need to build with GCC 6? > > This port is way behind and I can't prioritize updating it right now. Yes, it's been three months now and this is one of a handful broken ports blocking the update of the default version of GCC. Any chance you can get to this, or should we mark it BROKEN?
I tried poking around the code but couldn't get it to build with GCC 6 (and codehaus.org is gone, so even fetching the port is a PITA). There were some changes upstream many years ago that dropped the port's own math.h altogether. The port could either force GCC 5, be updated or marked BROKEN.
I won't be able to prioritize this any time soon, so I would mark it broken. I don't think there's any point trying to fix this version since it's so far behind and upgrading to the latest is going to take considerable time and effort.
A commit references this bug: Author: gerald Date: Sun Aug 13 20:37:41 UTC 2017 New revision: 447914 URL: https://svnweb.freebsd.org/changeset/ports/447914 Log: Mark BROKEN due to failure to build with modern compilers (and it generally requires updating). PR: 219282 Approved by: bacon4000@gmail.com (maintainer) Changes: head/lang/x10/Makefile
Created attachment 191154 [details] svn-diff-lang-x10-math.h If I am right, this builds without problems with gcc6.
The port fails differently with the patch (in a 12-i386 poudriere jail): ===> Building for x10-2.2.1_7 Exception in thread "main" java.lang.UnsupportedClassVersionError: org/apache/tools/ant/launch/Launcher : Unsupported major.minor version 52.0 at java.lang.ClassLoader.defineClass1(Native Method) at java.lang.ClassLoader.defineClass(ClassLoader.java:803) at java.security.SecureClassLoader.defineClass(SecureClassLoader.java:142) at java.net.URLClassLoader.defineClass(URLClassLoader.java:442) at java.net.URLClassLoader.access$100(URLClassLoader.java:64) at java.net.URLClassLoader$1.run(URLClassLoader.java:354) at java.net.URLClassLoader$1.run(URLClassLoader.java:348) at java.security.AccessController.doPrivileged(Native Method) at java.net.URLClassLoader.findClass(URLClassLoader.java:347) at java.lang.ClassLoader.loadClass(ClassLoader.java:425) at sun.misc.Launcher$AppClassLoader.loadClass(Launcher.java:312) at java.lang.ClassLoader.loadClass(ClassLoader.java:358) at sun.launcher.LauncherHelper.checkAndLoadMain(LauncherHelper.java:482) *** Error code 1 Stop. make: stopped in /usr/ports/lang/x10
Comment on attachment 191154 [details] svn-diff-lang-x10-math.h Makes no sense to patch it, there are newer versions. http://x10-lang.org/releases/x10-release-list.html
(In reply to w.schwarzenfeld from comment #10) Walter / Jason: do you intend to update this port to the latest version or is it better to remove it?
I don't work with it. I only try to patch. I make a quick attempt to update it, but I doubt I can work it out.
I've suggested removing it before. It's a challenging port that I doubt anyone is using and I have many higher priorities right now. I created the port for a computer science class in parallel programming a few years ago that was never offered again. Most real world distributed parallel programs are written in C/C++/Fortran with MPI.
OK, I'll remove the port.