Bug 201341 - databases/rocksdb: define BROKEN_DragonFly
Summary: databases/rocksdb: define BROKEN_DragonFly
Status: Closed FIXED
Alias: None
Product: Ports & Packages
Classification: Unclassified
Component: Individual Port(s) (show other bugs)
Version: Latest
Hardware: Any Any
: --- Affects Only Me
Assignee: Po-Chuan Hsieh
URL:
Keywords:
Depends on:
Blocks:
 
Reported: 2015-07-04 20:39 UTC by John Marino
Modified: 2015-07-09 18:55 UTC (History)
0 users

See Also:


Attachments

Note You need to log in before you can comment on or make changes to this bug.
Description John Marino freebsd_committer freebsd_triage 2015-07-04 20:39:26 UTC
I don't have time for these games:
https://lists.freebsd.org/pipermail/svn-ports-all/2015-July/098544.html

Three consecutive versions of this port have failed on gcc because it generates warnings on modern GCC and -Werror is set.  Clearly upstream does not test their product on modern GCC, so -Werror is even a worse idea than normal.

I don't want to get involved in a commit war or yet another proposal to strength the policy against horrible -Werror, so please just set "Broken_DragonFly= will not build on GCC with -Werror"

This s/w has broken 3 consecutive times so as long as -Werror is set (against recommendations) then let's just mask it to forget about it.
Comment 1 commit-hook freebsd_committer freebsd_triage 2015-07-09 15:49:30 UTC
A commit references this bug:

Author: sunpoet
Date: Thu Jul  9 15:48:59 UTC 2015
New revision: 391655
URL: https://svnweb.freebsd.org/changeset/ports/391655

Log:
  - Mark BROKEN on DragonFly BSD

  PR:		201341
  Submitted by:	marino

Changes:
  head/databases/rocksdb/Makefile
Comment 2 Po-Chuan Hsieh freebsd_committer freebsd_triage 2015-07-09 15:56:45 UTC
I think the problem is that you do not respect maintainer again and again. I'm fine with BROKEN_DragonFly.
Comment 3 John Marino freebsd_committer freebsd_triage 2015-07-09 16:03:20 UTC
In this case, the maintainer is violating general policy (and I hope soon-to-be hard, fast and documented policy) as well as good sense.

keeping -Werror on this port is a lousy technical decision.

Please understand the new world order.  Ports are not "owned" by maintainers, even if they created the port.  This was made clear within the last year when the just-fix-policy because official.

But since this port is probably crap anyway base on these gross errors, I'm going to let it go.
Comment 4 Po-Chuan Hsieh freebsd_committer freebsd_triage 2015-07-09 17:22:23 UTC
Please understand that DragonFly BSD is not FreeBSD. The just-fix-it blanket approval is not for breakage on other systems. I think portmgr clarified the support for extended systems in latest portmgr report.

I've never said that maintainer has ownership of a port. Being a committer, please do not forget to "Respect existing maintainers" [*]. You keep committing without any approval and this is the first time I reverted your commit.

If you want to fix this port on DragonFly BSD, you could submit PR or send mails just like others do.

[*] http://www.freebsd.org/doc/en/articles/committers-guide/rules.html
Comment 5 John Marino freebsd_committer freebsd_triage 2015-07-09 17:29:17 UTC
(In reply to Sunpoet Po-Chuan Hsieh from comment #4)
No one is talking about DragonFly.  -Werror affects GCC which FreeBSD has in ports.  You are supposed to support GCC and clang.  People override CC all the time.  This is why DF is valuable because it changes GCC failures.

Secondly, even if we were, this is in scope.  You did not understand the clarification.

You reverted a valid commit - period.   That is childish commit war stuff.  I'm not engaging when if you are going to do that.

-Werror is considered an ERROR, a MISTAKE.  "Just fix it" should cover it.  I will not push this further until -Werror policy is put on paper so I have something proving this.
Comment 6 John Marino freebsd_committer freebsd_triage 2015-07-09 17:29:57 UTC
*detects GCC failures
Comment 7 John Marino freebsd_committer freebsd_triage 2015-07-09 18:55:28 UTC
I just noticed you reverted the patch too.

I absolutely have a blanket permission to add patches like that which are no-op for FreeBSD. 

While the -Werror "fix" can be debated, removing that patch cannot.  I had explicit permission for it.