Created attachment 147246 [details] patch to update 2.2.6 zabbix22-server: update to 2.2.6
If possible please also include the following QA results to promote quick resolution: * Attach successful poudriere testport, or redports.org build logs * Attach portlint -AC output (after addressing any outstanding issues)
# portlint -AC WARN: Makefile: [0]: possible direct use of command "cp" found. use ${CP} instead. WARN: Makefile: possible use of absolute pathname "/etc/zabbix". WARN: Makefile: for new port, make $FreeBSD$ tag in comment section empty, to make SVN happy. WARN: Makefile: new ports should not set PORTREVISION. WARN: Makefile: COMMENT exceeds 70 characters limit. WARN: Makefile: "LIB_DEPENDS" has to appear earlier. WARN: Consider to set DEVELOPER=yes in /etc/make.conf 0 fatal errors and 7 warnings found.
Koobs - this a perfect example of why moving a PR to "open" to ask for testlogs is a terrible idea. The submitter did what you wanted, rather promptly, but the PR fell in the cracks anyway, just like I said it would. But what do I know? "OPEN" without assignee == wasteland. I'd recommend stop sending PRs to wasteland so easily
and pakhom706, the portlint is squawking about problems. If somebody askes for portlint output, and portlint shows an issue, you are expected to fix the issue and make a revised submission. There are 7 warnings there, 5 of which you could probably address.
although, these portlint output is suspect. If it thinks zabbix is a new port, then you are testing it outside the tree or something weird like that, I suspect.
Created attachment 147904 [details] with fix "COMMENT exceeds 70 characters limit"
I create new patch with fix "COMMENT exceeds 70 characters limit". > WARN: Makefile: [0]: possible direct use of command "cp" found. use ${CP} instead. This line in "@${REINPLACE_CMD}" > WARN: Makefile: possible use of absolute pathname "/etc/zabbix". This line in "@${GREP} and REINPLACE_CMD" > WARN: Makefile: for new port, make $FreeBSD$ tag in comment section empty, to make SVN happy. > WARN: Makefile: new ports should not set PORTREVISION. This isn't new port > WARN: Makefile: COMMENT exceeds 70 characters limit. Fix > WARN: Makefile: "LIB_DEPENDS" has to appear earlier. LIB_DEPENDS depended for port options.
(In reply to pakhom706 from comment #7) > > WARN: Makefile: new ports should not set PORTREVISION. > > This isn't new port > But portlint thinks it is new, which leads me to believe the way you are running portlint is strange. E.g. You don't have your port in an existing port tree and you are testing it by itself outside a tree.
I looked at the patch, it's a basic as it gets. In reality, koobs didn't really want the portlint, he wanted the poudriere logs for the version upgrade which you didn't provide.
ok, I'll be compile poudriere and provide logs. But I don't understand why. Changed only minor version and successfully compile.
Because: The committer that takes the PR may not be familiar with the port. He/she has no way of knowing if it's currently in good shape (although likely at this point) and he has no way of knowing if the update maintains the good status. Note -- plenty of simply updates forget to update the pkg-plist. Without QA checks, the port will build but it will fail poudriere with extra checks (bulk -t or testport) So just saying, "This is simple update" is really proving nothing. Poudriere checks more than "successful compile". If you don't do it, the committer must (he should anyway, but it depends on the reputation of the submitter). And committers want to know there's a good chance of success before they take the PR. They don't like untested patches -- untested as in: not tested by poudriere with -t option.
The redports.org can build zabbix22-agent 2.2.6, zabbix22-frontend 2.2.6 i386, zabbix22-proxy 2.2.6. Port zabbix22-server 2.2.6 the redports.org can't build becouse dependency could not be build. Logs: https://redports.org/~Pakhom/20141007202500-15461-251240/oniguruma4-4.7.1_1.log https://redports.org/~Pakhom/20141007202500-15461-251238/zabbix22-agent-2.2.6.log https://redports.org/~Pakhom/20141007202500-15461-251241/zabbix22-frontend-2.2.6.log https://redports.org/~Pakhom/20141007202500-15461-251242/zabbix22-proxy-2.2.6.log https://redports.org/~Pakhom/20141007202500-15461-251244/gmake-lite-4.1.log
You might be unaware, but redports is no longer considered a substitute for poudriere. It's completely unsuitable for QA checks these days. I'm not sure it even works correctly with recent changes to ports. But koobs did offer that as an alternative, so I'll let me take the risk. :)
s/me/him/ Weird typo..
Created attachment 148613 [details] update port to 2.2.7 Hi! It is an update for our zabbix22-* ports to the latest 2.0.7 version released today. Poudriere logs is coming.
Created attachment 148614 [details] zabbix22-server poudriere log
Created attachment 148615 [details] zabbix22-agent poudriere log
Created attachment 148616 [details] zabbix22-frontend poudriere log
sorry, you need to remove all the @dirrm* lines from the pkg-plist. You can see this in the poudriere logs too.
Created attachment 148827 [details] patch to 2.2.7
Created attachment 148828 [details] zabbix22-server-2.2.7.log poudriere log
Created attachment 148829 [details] zabbix22-frontend-2.2.7.log poudriere log
Created attachment 148830 [details] zabbix22-agent-2.2.7.log poudriere log
Bravo! Moving to patch-ready now.
i got confused who the maintainer was. since pakhom didn't object, I assume he's okay with the *tested* patch and I'm proceeding.
A commit references this bug: Author: marino Date: Tue Nov 4 21:27:05 UTC 2014 New revision: 372173 URL: https://svnweb.freebsd.org/changeset/ports/372173 Log: net-mgmt/zabbix22*: Upgrade version 2.2.5 => 2.2.7 PR: 193582 Fix by: timp87 (gmail) Approved by: maintainer - pakhom706 (gmail) Changes: head/net-mgmt/zabbix22-server/Makefile head/net-mgmt/zabbix22-server/distinfo head/net-mgmt/zabbix22-server/pkg-plist head/net-mgmt/zabbix22-server/pkg-plist.agent head/net-mgmt/zabbix22-server/pkg-plist.frontend
This patch contains several errors. Please wait, I will correct them.
:(
I don't see any errors. What are the concerns? This PR has been in the final queue for 5 days!
I mistake. this is correct path. I approve. Thanks
great, thanks.