Makefile of this port has hardcoded directories for integration with www/apache20. With this patch it is able to dynamically detect installed version of apache and configure accordingly. This adds support for www/apache21 (not tested) and www/apache22 (tested). How-To-Repeat: Try to use www/apache22 port with www/mod_perl2 shared module.
State Changed From-To: open->feedback Awaiting maintainers feedback
Hi, this appears to be OK, but note that I can only test mod_perl2 with apache20. If we put this patch in (and not just make mod_perl2 apache20-compatible only), how are we dealing with testing compatibility in the future? Lars -- Lars Eggert NEC Network Laboratories
Do you suggest making another port of mod_perl2 that works only with apache22? I don't agree with that. Other ports don't work that way either. For example, php5 port has the similar apache version detection. If the maintainer is not able to test some of the functionality then he should find someone else to do the testing for him. Apache 2.2 is stable, and I'm not the only one that needs mod_perl2 with it.
On Dec 13, 2005, at 10:18, =8Aimun Mikecin wrote: > Do you suggest making another port of mod_perl2 that works only with > apache22? I don't agree with that. apache20, actually. > Other ports don't work that way either. For example, php5 port has the > similar apache version detection. > If the maintainer is not able to test some of the functionality =20 > then he > should find someone else to do the testing for him. Well, do *you* volunteer to test apache21 and apache22 whenever I =20 make a change to the mod_perl2 port? > Apache 2.2 is stable, and I'm not the only one that needs mod_perl2 =20= > with > it. Lars -- Lars Eggert NEC Network Laboratories
Just like the maintainership of the port does not mean you will *always* take care of bugfixing, updating and resolving problems, I cannot be a tester whenever you make a change. Port maintanership is based on doing it on best-effort basis as time permits, so is my help with testing. Nobody is paid to do it. Beside maintaining a few ports, I do testing for other maintainers in the environment they cant do the testing themself. It is usually testing on different hardware or different architecture. It would be a very bad thing for them to have an attitude to refuse to support other architectures just because they can test it only on (for example) i386.
On Dec 13, 2005, at 12:51, =8Aimun Mikecin wrote: > Just like the maintainership of the port does not mean you will =20 > *always* > take care of bugfixing, updating and resolving problems, I cannot be a > tester whenever you make a change. Port maintanership is based on =20 > doing > it on best-effort basis as time permits, so is my help with testing. > Nobody is paid to do it. I am well aware of what port maintainership entails. > Beside maintaining a few ports, I do testing > for other maintainers in the environment they cant do the testing > themself. It is usually testing on different hardware or different > architecture. Hence my question if you were wiling to do the same for mod_perl2. =20 Apparently, you aren't, which is fine of course. > It would be a very bad thing for them to have an attitude > to refuse to support other architectures just because they can test it > only on (for example) i386. I'm not refusing anything. I'm pointing out that adding support for =20 configurations I can't test may create problems, because people will =20 expect those configurations to work when I cannot test them. I'm open for constructive suggestions on how to deal with this. One =20 way would be someone that can test these configurations and provide =20 feedback. Another - worse - way is to temporarily disable those =20 configurations, to avoid violating POLA ("well, it compiled, so it =20 must run correctly, too"). Lars -- Lars Eggert NEC Network Laboratories
Don't get me wrong: I'm willing to help as my time and environment permits me to. No more, no less. If in some configuration it doesn't work, you can always mark it BROKEN in that case until someone comes with a solution.
State Changed From-To: feedback->closed Superceded by ports/90655