Rationale: portmaster is a >4000 (!!!) line shell script. It's common sense that shell scripts are ok for small tasks but quickly become unmaintanable once they grow. Portmaster should simply do "cd /usr/ports && make help" Don't laugh I'm serious with this.
(In reply to torsten.eichstaedt from comment #0) > Rationale: portmaster is a >4000 (!!!) line shell script. It's common sense > that shell scripts are ok for small tasks but quickly become unmaintanable > once they grow. > > Portmaster should simply do "cd /usr/ports && make help" > > Don't laugh I'm serious with this. Correction; it's a collection of routines. Not unlike that of a program written in C, or any number of other languages. I might also add; it works. So what exactly is your beef, again? --Chris
(In reply to Chris Hutchinson from comment #1) > I might also add; it works. So what exactly is your beef, again? Maybe in your universe it works. Working ports don't have PR lists like this:http://portsmon.freebsd.org/portoverview.py?category=ports-mgmt&portname=portmaster&wildcard= This port should be banned. Now without a _REAL_ maintainer, it's well on it's way.
(In reply to John Marino from comment #2) > (In reply to Chris Hutchinson from comment #1) > > I might also add; it works. So what exactly is your beef, again? > > Maybe in your universe it works. For it's intended purpose, it works. At least as well as most "utilitarian" type ports -- it has it's issues, as do the others. I wish that the OP had combined this with: https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=199572 it would have made his case here more coherent. This feels like a rant, after his expressed experience in the other PR. The other PR looks like an install, as opposed to an upgrade; which is what ports-mgmt/portmaster's intended use-case is. > > Working ports don't have PR lists like > this:http://portsmon.freebsd.org/portoverview.py?category=ports- > mgmt&portname=portmaster&wildcard= Several of those are dupes, and most are at least a year old. > > > This port should be banned. Now without a _REAL_ maintainer, it's well on > it's way. bdrewery@ was maintaining it up to at least 4 weeks ago. Are you suggesting I maintain this port, that it might get the attention it needs. Or will that just make me the scourge of the committers? IMHO portmaster is a far better choice for upgrades, than pkg, it you're building from source. It affords you options *other* than default. Which pkg doesn't cater well to, if at all. --Chris
(In reply to torsten.eichstaedt from comment #0) > Rationale: portmaster is a >4000 (!!!) line shell script. It's common sense > that shell scripts are ok for small tasks but quickly become unmaintanable > once they grow. > > Portmaster should simply do "cd /usr/ports && make help" > > Don't laugh I'm serious with this. I'm laughing. Poudriere is what we use to build packages now and is extremely stable and efficient. It is 95% shell script. The language something is written in is not very relevant usually as long as it works. Portmaster has a lot of issues. I will agree that the structure of the codebase in portmaster is very unmaintainable. I don't see any reason to ban it unless it starts doing things wrong, like Tinderbox does now. Closing as this is not a constructive ticket.