2 changes for this port: Toplevel Makefile: change USE_GMAKE= yes to USES= gmake toplevel files/patch-Makefile change the line + CFLAGS+=-Wall ${PTHREAD_CFLAGS} -DPREFIX='"${PREFIX}"' -DRUNDIR='"${RUNDIR}"' to + CFLAGS+=-Wall ${PTHREAD_CFLAGS} -DPREFIX='"${PREFIX}"' -DRUNDIR='"${RUNDIR}"' -Wno-pointer-sign And of course, change the Makefile to NOT show this port as broken!
1) normally revivals are shown with shar files or patch sets, not descriptions of changes made 2) I know this isn't right because pkg-plist wasn't touched, which would be required based on recent changes to port tree. 3) revival ports need to be accompanied with proof it builds cleanly without error. Ideally with poudriere testport but at least with "make check-plist", "make stage-qa" output, along with portlint output. So right now, this is a no-go. As a side remark, for some reason I can't see SVN linked from freshports but it should be there since it was deleted in sept 2014.
I forgot 4) ports are not revived without a maintainer, which was not identified either.
I presumed the original port maintainer would have been automatically emailed and might have eventually got back to it once he saw the change was so trivial. I guess when the port was marked broken the original maintainer's email address was struck from the port. OK whatever. I didn't know we were doing that now. This one built fine under 9.2 6 months ago and has worked ever since. Back in the "olden days" in FreeBSD we had a principle that if you made trivial unnecessary changes in the code that broke other stuff that you fixed what you broke. Changing USE_GMAKE variable to the USES variable must have broken hundreds if not thousands of ports and there seems no justification for it other than someone's sense of aesthetics. I suppose whoever did it is real proud of themselves as it has obviously killed many orphaned ports that otherwise worked fine but were on autopilot with maintainers out to lunch. Good Job! No wonder people go to Linux. So be it. Filing this was purely a hand out to whoever might have just updated to a new system and had this break. You might consider that syndrome is in the ports distro included with the 9.3-RELEASE ISO so anyone installing Ports off that instead of sucking the current ports tree down is going to get this port and when they try to build it they will run into this nonsense. It would have saved me an hour if this bug report had been in the system. If someone else wants to adopt this orphan at least they will know what is needed to fix it. I am not interested in going through the formal process of reviving and adopting this just to make a single line change. It worked fine for me when it was on autopilot and it works fine today on 9.3. The next time I build a server that needs it I'll just build it manually and screw fooling with ports. You guys are too busy breaking things in there anyway. Incidentally, it makes no difference if -Wno-pointer-sign is there or not since that just suppresses a warning anyway, and the compiler does the Right Thing in either case.
(In reply to tedm from comment #3) > I presumed the original port maintainer would have been automatically > emailed and might have eventually got back to it once he saw the change was > so trivial. I guess when the port was marked broken the original > maintainer's email address was struck from the port. No, the port was reset in 2012. Freshport history shows this clearly. It was deleted because it wasn't staged, along with 600 other ports. > Back in the "olden days" in FreeBSD we had a principle that if you made > trivial unnecessary changes in the code that broke other stuff that you > fixed what you broke. Changing USE_GMAKE variable to the USES variable must > have broken hundreds if not thousands of ports and there seems no > justification for it other than someone's sense of aesthetics. Where are you getting this crap? USE_GMAKE wasn't removed until all the *CURRENT* ports had been transitioned. It should be painfully obvious that any port removed before that happened would still have the knob because it had never been switched. > I suppose > whoever did it is real proud of themselves as it has obviously killed many > orphaned ports that otherwise worked fine but were on autopilot with > maintainers out to lunch. Good Job! No wonder people go to Linux. It killed exactly zero ports. Nice troll though. > So be it. Filing this was purely a hand out to whoever might have just > updated to a new system and had this break. You might consider that > syndrome is in the ports distro included with the 9.3-RELEASE ISO so anyone > installing Ports off that instead of sucking the current ports tree down is > going to get this port and when they try to build it they will run into this > nonsense. I have no idea what you are talking about. This port was pruned because it was unmaintained and nobody else cared enough to stage it before it was was removed despite have 14 months warning. > It would have saved me an hour if this bug report had been in the system. > If someone else wants to adopt this orphan at least they will know what is > needed to fix it. I am not interested in going through the formal process > of reviving and adopting this just to make a single line change. It worked > fine for me when it was on autopilot and it works fine today on 9.3. The > next time I build a server that needs it I'll just build it manually and > screw fooling with ports. You guys are too busy breaking things in there > anyway. Okay, so bottom line: You are withdrawing the PR. That's fine by me. Closing as requested.