Created attachment 202963 [details] svn diff update to 0.207 Two svn patches added to update emulators/mame and emulators/mess to 0.207. Makefile required a bit of tinkering. Had to add an additional $REINPLACE_CMD to get all the 'util/png.h' -> 'util/localpng.h' correct. I tried to fix it in the grep above but it kept missing one file. This was the simplest and most stable way. pkg-plist needed some files added and removed. Patch files was a bit messy. svn remove files/patch-scripts_toolchain.lua svn remove files/patch-src_lib_netlist_plib_pparser.h svn remove files/patch-src_lib_netlist_plib_pparser.cpp svn move files/patch-src__osd__eigccx86.h files/patch-src_osd_eigccx86.h svn move files/patch-src__osd__eminline.h files/patch-src_osd_eminline.h svn remove files/patch-asio-libc++7 svn add files/patch-3rdparty_asio_include_asio_detail_config.hpp Removed: files/patch-scripts_toolchain.lua files/patch-src_lib_netlist_plib_pparser.cpp files/patch-src_lib_netlist_plib_pparser.h They were not needed any more. Renamed: files/patch-src__osd__eigccx86.h -> files/patch-src_osd_eigccx86.h files/patch-src__osd__eminline.h -> files/patch-src_osd_eminline.h Naming convention fix. Removed: files/patch-asio-libc++7 Added: files/patch-3rdparty_asio_include_asio_detail_config.hpp Patch file deals with the same file and basically does the same thing. But I had to redo it as it failed to patch the new source. The git commit reference at the top of the old file didn't make sense to me, it appeared to be pointing at something completely unrelated to this patch. New patch file named according to naming convention. Added also an update for the emulators/mess slave port. It had similar pkg-plist changes. Not sure why a separate mess port still exists, updated it because emulators/qmc2 depends on it.
Created attachment 202964 [details] svn diff for emulators/mess slave port
MAME 0.208 came and went. Now it's 0.209.
Created attachment 204098 [details] svn diff update to 0.209 MAME 0.208 came and went. I was testing a patch for it but 0.209 was released before I finished. Updated patch diff to get to 0.209. Subversion instructions remain the same.
Created attachment 204099 [details] svn diff for emulators/mess slave port Updated pkg-plist for emulators/mess.
testbuilds@work
Committed, thanks!
A commit references this bug: Author: pi Date: Tue Apr 30 05:32:17 UTC 2019 New revision: 500456 URL: https://svnweb.freebsd.org/changeset/ports/500456 Log: emulators/mame: upgrade 0.200 -> 0.209 emulators/mess: upgrade slave port pkg-plist PR: 236621 Submitted by: sirdice@gmail.com Relnotes: https://www.mamedev.org/releases/whatsnew_0209.txt https://mamedev.org/releases/whatsnew_0208.txt https://mamedev.org/releases/whatsnew_0207.txt ... Changes: head/emulators/mame/Makefile head/emulators/mame/distinfo head/emulators/mame/files/patch-3rdparty_asio_include_asio_detail_config.hpp head/emulators/mame/files/patch-asio-libc++7 head/emulators/mame/files/patch-scripts_toolchain.lua head/emulators/mame/files/patch-src__osd__eigccx86.h head/emulators/mame/files/patch-src__osd__eminline.h head/emulators/mame/files/patch-src_lib_netlist_plib_pparser.cpp head/emulators/mame/files/patch-src_lib_netlist_plib_pparser.h head/emulators/mame/pkg-plist head/emulators/mess/pkg-plist
Thanks for the time to commit it. But I believe a mistake has been made. Two patch files should have been renamed but are now deleted and are causing build failures. svn move files/patch-src__osd__eigccx86.h files/patch-src_osd_eigccx86.h svn move files/patch-src__osd__eminline.h files/patch-src_osd_eminline.h Restoring those two would be fine too. I only renamed them to make them conform to the naming convention.
A commit references this bug: Author: pi Date: Tue Apr 30 16:14:06 UTC 2019 New revision: 500501 URL: https://svnweb.freebsd.org/changeset/ports/500501 Log: emulators/mame: revert r500456 and reapply with fixes PR: 236621 Reported by: Remko C. <sirdice@gmail.com> Pointy hat to: pi Changes: head/emulators/mame/Makefile head/emulators/mame/distinfo head/emulators/mame/files/patch-3rdparty_asio_include_asio_detail_config.hpp head/emulators/mame/files/patch-asio-libc++7 head/emulators/mame/files/patch-scripts_toolchain.lua head/emulators/mame/files/patch-src__osd__eigccx86.h head/emulators/mame/files/patch-src__osd__eminline.h head/emulators/mame/files/patch-src_lib_netlist_plib_pparser.cpp head/emulators/mame/files/patch-src_lib_netlist_plib_pparser.h head/emulators/mame/pkg-plist
Hmm, with the patches, the port fails on 11.2amd64: https://people.freebsd.org/~pi/logs/emulators__mame-112-1556641451.txt which surprises me. A second attempt suceeded. Strange.
A commit references this bug: Author: pi Date: Tue Apr 30 17:50:29 UTC 2019 New revision: 500513 URL: https://svnweb.freebsd.org/changeset/ports/500513 Log: emulators/mame: upgrade 0.200 -> 0.209 - 2nd attempt, with additional patches PR: 236621 Submitted by: sirdice@gmail.com Relnotes: https://www.mamedev.org/releases/whatsnew_0209.txt https://mamedev.org/releases/whatsnew_0208.txt https://mamedev.org/releases/whatsnew_0207.txt ... Changes: head/emulators/mame/Makefile head/emulators/mame/distinfo head/emulators/mame/files/patch-3rdparty_asio_include_asio_detail_config.hpp head/emulators/mame/files/patch-3rdparty_bx_include_bx_platform.h head/emulators/mame/files/patch-3rdparty_bx_src_mutex.cpp head/emulators/mame/files/patch-3rdparty_bx_src_thread.cpp head/emulators/mame/files/patch-asio-libc++7 head/emulators/mame/files/patch-makefile head/emulators/mame/files/patch-scripts_toolchain.lua head/emulators/mame/files/patch-src__osd__eigccx86.h head/emulators/mame/files/patch-src__osd__eminline.h head/emulators/mame/files/patch-src_lib_netlist_plib_pparser.cpp head/emulators/mame/files/patch-src_lib_netlist_plib_pparser.h head/emulators/mame/files/patch-src_osd_eigccx86.h head/emulators/mame/files/patch-src_osd_eminline.h head/emulators/mame/files/patch-src_osd_modules_file_posixptty.cpp head/emulators/mame/pkg-plist
(In reply to Kurt Jaeger from comment #10) This seems to be the culprit: In file included from ../../../../../src/mame/drivers/konamigq.cpp:76: ../../../../../src/devices/cpu/tms57002/tms57002.h:208:10: fatal error: '../../emu/cpu/tms57002/tms57002.hxx' file not found #include "../../emu/cpu/tms57002/tms57002.hxx" ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 1 error generated. gmake[3]: *** [konami.make:1745: ../../../../freebsd/obj/x64/Release/src/mame/drivers/konamigq.o] Error 1 The missing .hxx file appears to be generated dynamically at some point. Unfortunately, the whole build process is about as clear as mud to me. Make just fires off a whole bunch of LUA scripts that do all sorts of wonderful things. Then a bunch of Python scripts are used. > which surprises me. A second attempt suceeded. Strange. A fluke? I'm going to setup a bunch of test jails with different versions just in case. I really only looked at AMD64 and 12-STABLE. But my changes to get it working from 0.200 to 0.209 were quite minimal once I understood what the previous patch files tried to do. Thanks for commiting it, sorry about the mixup with the patch files. I really like MAME. I'm going to persevere trying to figure out how it all works :)
(In reply to sirdice from comment #12) So it's probably not consistently build in multi-core build runs. Do you have a server at hand that has lots of cores to test this ?
Closing this, new patches please to new PRs 8-)