On Sat, 2013-11-30 at 15:36 +1300, David Koontz wrote: > An mcode version is the same style of ghdl as currently fielded in Windows > and OS X. It's the jit version of ghdl which doesn't include the ability to > include foreign functions ( .o files) and doesn't operate as a gcc front end. > > I built one before answering, took about 5 minutes to build, including > fumbling, this was from memory (and I wrote the directions out elaborated a > bit in the OS X version's package installer to comply with the spirit of the > GPL). > > Directions: > > checkout revision r150 and move it to somewhere with a short path (apply patches, none of which affect the problem below)
> cd /opt/revision_150/translate/ghdldrv > If you look in the Makefile there you'll find the default target is linux and > i686, which is fine. I changed my target to x64... > make Here I see a failure: ----------------------------------- gcc-4.8 -c -I./ -I../../ortho/mcode -I../../ortho -gnaty3befhkmr -gnata -gnatwae -I../.. -I.. -I../../psl -I../grt -g -gnatf -gnat05 -I- /home/brian/Projects/ghdl/mcode/ghdl/canon.adb canon.adb:2432:16: warning: variable "El" is not modified in loop body canon.adb:2432:16: warning: possible infinite loop gnatmake: "/home/brian/Projects/ghdl/mcode/ghdl/canon.adb" compilation error ----------------------------------- Looking at the source, the loop is indeed infinite, but in an "if False" block... All I can assume is that one of the Makefile GNATFLAGS options elevates warnings to errors - but why does your build not see this error? perhaps Gnat 4.8 is better than it used to be at detecting warnings? I see you used gnat 4.6. I'm commenting out the erroneous code and moving on, but what would your recommended fix be? - Brian _______________________________________________ Ghdl-discuss mailing list [email protected] https://mail.gna.org/listinfo/ghdl-discuss
