19:29:15 #startmeeting 19:29:16 Meeting started Wed Apr 18 19:29:15 2018 UTC. The chair is sumpfralle. Information about MeetBot at http://wiki.debian.org/MeetBot. 19:29:16 Useful Commands: #action #agreed #help #info #idea #link #topic. 19:29:35 #chair TheSnide dipohl h01ger be0rn chteuchteu 19:29:35 Current chairs: TheSnide be0rn chteuchteu dipohl h01ger sumpfralle 19:29:55 welcome to the weekly munin meeting :) 19:30:20 welcome back sumpfralle :) 19:31:18 thank you! 19:31:24 #topic review last meeting 19:31:37 http://meetbot.debian.net/munin/2018/munin.2018-04-04-19.30.html 19:32:34 There were two action items: (1) myself registers the munin website at web.archive.org and (2) TheSnide recovers a few removed guide pages for munin-2.0 19:32:56 (1): I failed to do this - and I am almost convinced now, that there is no need to "subscribe" there? 19:33:05 any other opinions on that matter? 19:33:53 for what purpose should we archive the old trac wiki? 19:34:25 it's nearly empty compared to the state of 2013.. 19:34:38 it was part of the discussion in the last meeting. If I remember correctly, we confused each other a bit with archive/backup/dump/export ... 19:34:49 yes, I think, we can omit that topic 19:34:54 +1 19:35:42 regarding the guide pages regarding munin-2.0: I did not take a close look - are they still lost in the git history? 19:36:41 * sumpfralle looks at the git history now 19:37:34 I looked in the Guide and there is no info about munin-graph and munin-html 19:37:35 http://guide.munin-monitoring.org/en/latest/master/index.html 19:38:31 compare it with the stable 2.0 Guide: http://guide.munin-monitoring.org/en/stable-2.0/master/index.html#components 19:38:46 indeed! 19:39:01 thus they were only added to the stable repository. Now I get it ... 19:39:06 I can move them over. 19:39:21 I don't thinkt they were added there 19:39:34 we only fed the master repo 19:39:51 so someone deleted the pages I suppose 19:40:10 we can check with "blame" in git 19:40:11 hm - but I took a quick look at the doc/ history in master and saw only the two of use adding commits in the last six month 19:40:32 master branch was devel before 19:40:39 and we fed to devel before 19:41:04 ah - here it is: 466ac02e46c389fa78da6cbe347a4c70f4b7255a 19:41:08 removed in 2015 by ssm 19:41:41 should we just revert this one - or do you suspect other missing details? 19:41:49 I don't know 19:42:14 I think it is worth to check ssm activities at the time 19:42:55 I thought TheSnide claimed that task in the last meeting 19:43:14 ok - so we leave it with him, until he complains :) 19:43:32 #action TheSnide will recover the few removed munin-2.0-pages of the munin-guide 19:43:49 #topic documentation 19:43:56 I hope that it will be done carefully.. 19:44:15 of course! 19:44:29 it is not funny to invest such an amount of time and work (as I did in the last years) and see vanishing the fruits.. 19:44:29 recently we talked about man pages and the doc directory 19:44:39 I understand 19:45:12 man/sphinx: I noticed that the manpages for the Debian package are generated from the doc directory. 19:45:23 by "make -C doc man" - just using sphinx 19:45:31 thus the magic is resolved? 19:46:06 so they are built from the guides pages already? 19:46:19 yes 19:46:26 just how it should be :) 19:46:46 at that moment I realized, that a bit of documentation of the build system would be nice 19:47:02 hmm, then we would have to care for this parts in the stable-2.0 repo.. 19:47:17 a task that I was not aware of since now.. 19:47:31 #action sumpfralle will add build system details to the development documentation of the guide (in case it is missing) 19:47:57 many thanks in advance for that!! :-) 19:48:14 indeed there are no manpages in the Debian package for 2.0.x 19:48:42 thus the bits are probably missing from stable-2.0 (expected - since we care only about the doc in master) 19:49:01 what the f.. 19:49:16 ? 19:49:49 or: they are just not built from that directory (this would be a thing for the Debian package) 19:49:50 I didn't expect such a mess.. 19:50:05 probably I got it wrong 19:50:16 there is indeed the doc directory in stable-2.0 19:50:21 I see man pages in epel munin 19:50:33 but the Debian package just did not use them - that's it 19:50:34 but they are different from the guides man pages 19:50:48 different from the old ones? 19:50:56 ("old" -> "stable-2.0" branch) 19:51:18 I didn't check the new Munin version 19:51:51 I am talking about the last epel package 2.0.33 if I see it right 19:52:42 I thought the man pages were built from "normal" files in the repo there not from the Guide.. 19:53:25 at least the Debian packaging uses just the source tarball of the release 19:53:29 TheSnide, h01ger: are you here? 19:53:35 thus (for 2.0.x): the branch "stable-2.0" 19:54:24 I think, it is fine like that - maybe a bit outdated compared to the master branch - but I do not think, that we should focus on that. 19:54:36 +1 19:54:47 anyway: I will take a look at the man pages for the current munin package in Debian. h01ger will probably like that ... 19:55:05 but there should be man pages also in the debian package of stable 2.0 19:55:18 #action sumpfralle will include manpages in the current (2.0.x) Debian package 19:55:27 that's the only point I have.. 19:55:32 thanks again! 19:56:09 we will see, if my next week will be more productive than the last one :) 19:56:20 I think it is ok to maintain the man pages for 2.0 in other files 19:56:51 we should start the /new/ method with 2.999 19:57:13 they are there in the doc directory - ready to be used (at least I assume that) 19:57:53 make -C doc man 19:57:57 writing... munin-async.1 { } munin-asyncd.1 { } munin-cgi-graph.1 { } munin-cgi-html.1 { } munin-check.1 { } munin-cron.1 { } munin-graph.1 { } munin-html.1 { } munin-limits.1 { } munin-node.1 { } munin-run.1 { } munin-update.1 { } munin.conf.5 { } munin-node.conf.5 19:58:06 this should be sufficient 19:58:15 (in stable-2.0) 19:58:27 and we need to clearly address outdated features, so care for multi-version in the new method 19:58:40 yes 19:58:45 as in the guide we want to explain both versions 19:58:48 do we already know how we can tag them? 19:59:08 but the man pages should contain only valid 2.999 info 19:59:29 tagging: no 19:59:38 nobody answered to your call on the ML 20:00:11 I repeated it now - we will see ... 20:00:35 "but the man pages should contain only valid 2.999 info" - this will be tricky 20:00:52 probably we will need to live with textual markup ("only 3.x" ...) 20:01:15 perhaps we can integrate sort of exclusion marks 20:01:24 (I am not a sphinx expert - just guessing) 20:02:02 or add some awk magic to the man page build 20:03:15 the "only 3.x" will not be the problem, when we only use the new method for 2.999/3.x 20:03:31 but to throw out outdated infos from stable 2.0 20:03:42 in the man pages for 2.999 20:04:49 or we maintain 2 separate pages in the guide only for this content (man pages) 20:05:08 with "only 3.0" markers, I meant markers for 3.0 and 2.0 - I could live with that, if sphinx does not offer something out of the box 20:05:13 we cannot do this for all the content, but only for the man pages will be possilbel 20:05:23 possible 20:05:50 personally I do not think, it is worth the effort - adding hints "only 2.0" and "only 3.0" would be sufficient for me 20:06:20 do you know of other software that has multi-version man pages? I don't 20:06:37 I will take a look, if sphinx offers a feature like that 20:06:39 And I wouldn't like it 20:06:51 ressources, ressources :) 20:07:13 we are talking about 4 or 5 pages if I see it right 20:08:02 I would prefer the sphinx / make -C doc man magic 20:08:03 I have no real opinion - and I could live with any option 20:08:24 but if not possible I vote for separate pages 20:08:28 for both versions 20:09:30 I can live with a multi-version guide and see also advantages as we show the users the differences between the 2 versions, so the progress of the new version 20:09:48 but a man page should address the version that I installed 20:10:04 and not outdated info 20:10:12 we can have this anyway: the files are separate in the two branches 20:10:36 thus we do not apply the "multi-version" policy to the manpage source files, right? 20:11:18 that is: everything below doc/reference 20:11:33 Let's see if we find sphinx / make doc magic to break down fitting man pages for both versions 20:11:50 that would be the best in my view 20:12:00 but remember: manpages for 2.0.x will always be built from the stable-2.0 branch, anyway 20:12:11 as I said before 20:12:19 (at least it is the Debian way of generating manpages from the source tarball) 20:12:23 let us build the man pages for stable 2.0 in the old way 20:12:37 there is no old way - I think, we have a misunderstanding here 20:12:43 there is only the sphinx way 20:12:48 I am not aware of another one 20:12:50 so not from the guide but from separate files in the source tarball 20:13:18 maybe I used improper wording: I always refered to the "doc/" directory 20:13:34 there are the manpage sources below doc/reference - see doc/conf.py 20:13:44 for master as well as for stable-2.0 20:13:55 https://github.com/munin-monitoring/munin/tree/stable-2.0/master/doc 20:14:11 ^that is not the guide! 20:14:27 thus it was just my confusing wording - sorry 20:15:08 so the new method to fetch the man pages from the guide is not implented yet also for the master branch? 20:16:14 let me rephrase it: we can run "make -C doc man" both in master and in stable-2.0. This was always possible. The resulting files are currently not used by the Debian package. I will take a look at that. 20:16:22 From my point of view, there is nothing to be done or changed? 20:16:38 https://github.com/munin-monitoring/munin/tree/stable-2.0/node/doc 20:17:00 this could be added, as well 20:17:23 there should be another man page for the node: "munindoc" 20:17:34 which seems to be missing at the time 20:17:47 and there is master/doc, too 20:19:39 the munindoc' manpage could be generated via perldoc - indeed 20:19:59 this should cover it all: doc/reference, node/doc, master/doc, node/bin/munindoc 20:20:27 build from doc/reference is the /new/ method 20:20:30 the Makefile refers to all of them, I think 20:20:34 yes 20:20:59 let us stop and pick up this thread again when you have documented the build process and h01ger and TheSnide are also on board, ok? 20:21:04 yes 20:22:21 thus let us close this meeting? 20:23:39 one point left on my side 20:23:49 #topic vetted plugins 20:23:52 http://guide.munin-monitoring.org/en/latest/develop/plugins/requirements-vetted.html 20:24:02 did you see this? 20:24:36 yes, I was happy to see another one disappearing :) 20:25:38 I would highly appreciate if you add your /rules/ from contrib therein 20:26:21 it would fit there - but I am not sure, if we should remove them from contrib/plugins/README in this case? I would not like to care for two copies? 20:27:19 I think of only the subject "plugin code" related issues 20:28:27 especially the thoughts about care for security are important as I see it 20:29:03 yes, indeed 20:29:20 which specific paragraphs would you think, should be moved from the README to the guide? 20:29:46 I have not a concrete concept at the time.. 20:29:58 let's talk again next week? 20:30:02 yes 20:30:09 #endmeeting