Meetings:Docs IRC log 20090312

From FedoraProject

Jump to: navigation, search
* quaid looks around 17:05
* jjmcd is here 17:05
* radsy here 17:05
* laubersm is sort of here 17:05
kirkz kirkz is her 17:05
-!- ryanlerch [n=rlerch@nat/redhat/x-5df7f47308d67817] has joined #fedora-meeting 17:05
-!- quaid changed the topic of #fedora-meeting to: Docs mtg 17:05
jjmcd GM ryanlerch 17:05
quaid <meeting id="Docs team"> 17:05
-!- mdious [n=mmcallis@nat/redhat/x-e94a24b95a20cf90] has joined #fedora-meeting 17:05
* ke4qqq pretends to be here 17:05
* quaid reloading browser, if someone else has a link to the agenda handy ... 17:06
* ryanlerch is here 17:07
mdious here 17:07
quaid ok, I'm agenda less 17:08
quaid I know release notes were a hot topic 17:08
kirkz https://fedoraproject.org/wiki/DocsProject/SteeringCommittee/Meetings#Agenda_for_Next_Meeting 17:08
quaid last night there was some solution thinking done to figure out how to get Fedora-legal packages from Publican 17:08
-!- rudi [n=rlandman@nat/redhat/x-0301dda83f4e389b] has joined #fedora-meeting 17:10
jjmcd I have a hack, now I need somebody who knows what he is doing to tell me if there is anything wrong 17:10
ianweller oh hi 17:10
jjmcd The resulting package installs ok, passes rpmlint, and meets the guidelines as best i can tell 17:10
jjmcd http://fedoraproject.org/wiki/User:Jjmcd/Drafts/Converting_Publican_RPM_for_Fedora 17:10
jjmcd But I know zip about packaging so there could be horrible monsters hiding there 17:11
quaid hmm 17:13
quaid this is one that is probably worth explaining in detail 17:13
quaid to the mailing list 17:13
quaid and seeking feedback that way. 17:13
jjmcd That might be a plan 17:13
quaid otherwise we'll spend this whole meeting in technojungle 17:13
jjmcd yah 17:13
-!- nman64 is now known as nman64_away 17:14
quaid ah, now I see agenda 17:15
quaid well, are there any questions on this item? 17:15
mdious radsy: does <http://fedoraproject.org/wiki/User:Jjmcd/Drafts/Converting_Publican_RPM_for_Fedora> help you? 17:16
radsy mdious, yes it looks good, but not knowing a lot about packaging, feedback from someone who works more closely with packaging would be valuable 17:17
laubersm quaid, As I understand it - Rel Notes have a deliverable soon - how can those of us not already involved help? 17:17
-!- themayor [n=jack@dsl081-200-011.nyc2.dsl.speakeasy.net] has quit [] 17:17
ryanlerch laubersm: we need writers 17:17
-!- stickster [n=npfrield@fedora/stickster] has quit [Read error: 110 (Connection timed out)] 17:18
quaid jjmcd: ok, can you send that email to the list? 17:18
jjmcd Yes I will 17:18
quaid hey, stickster, don't go now! 17:18
laubersm ryanlerch, in wiki? research can content or wordsmithing and proofreading 17:18
* laubersm expects both 17:18
-!- quaid changed the topic of #fedora-meeting to: Docs mtg -- Release notes: i) Beta deliverable discussion; ii) process discussion 17:18
jjmcd research more thanwordsmithing 17:19
quaid so, yeah, the Beta one-page is next due ... 17:19
jjmcd Especially it seems like the more technical beats - kernel, arch-specific, like that 17:19
quaid http://poelstra.fedorapeople.org/schedules/f-11/f-11-docs-tasks.html 17:19
jjmcd quaid: is that much more than an edit to the alpha? 17:20
quaid Mon 2009-03-23 17:20
quaid jjmcd: well .. 17:20
quaid take Alpha, make new page 17:20
quaid check features, see if there is content to mention/migrate 17:20
quaid check beats, see if there is content to mention/migrate 17:20
quaid post to f-devel-l and say, hey, the Beta page is open, put stuff there 17:20
quaid make sure it is edited by ~22 March :) 17:21
quaid so a repeat of the Alpha process, just easier since the page exists. 17:21
jjmcd hehe - there is one heck of a lot of content, whether it is worth mentioning depends on who you are 17:21
-!- stickster [n=pfrields@ip72-205-14-2.dc.dc.cox.net] has joined #fedora-meeting 17:22
jjmcd If you're a developer, gcc is a big deal, if you're not, who cares 17:23
quaid well, it's a one page 17:24
quaid not a preview of the big notes 17:24
jjmcd yep 17:24
stickster jjmcd: I've been hearing radio silence since about 0007 UTC -- but I think my server fell over without me knowing it. 17:24
quaid you can link to the beats in progress, for example 17:24
quaid stickster: yep 17:24
stickster If I missed stuff, can someone pastebin it for me? 17:24
quaid stickster: sure, I will, just a moment 17:25
radsy what's the last thing you saw? 17:25
stickster Nothing since 0000 UTC other than me yapping 17:25
quaid stickster: http://fpaste.org/paste/5905 17:25
stickster Thanks quaid 17:25
quaid ok, that cutoff the "I'm here" 17:26
* ianweller goes for dinner, sorry 17:26
quaid and I just discovered I need to do a child pick-up, zapped by the time change 17:26
quaid stickster: can you takeover with "relnotes process" talk? 17:26
quaid after you catch up :) 17:26
stickster OK 17:27
quaid ok, brb 17:27
ryanlerch where is Sparks? 17:27
jjmcd net probs 17:27
stickster *: OK, it looks like quaid explained the Beta relnotes one-page 17:27
stickster At this time, we should also be starting to make sure we know beat assignments, ask for developers and community people to help shore up beats that have no assigned persons 17:28
stickster I haven't seen the docs-list since yesterday -- did anyone read my schedule? 17:29
stickster s/my/the excerpted/ 17:29
jjmcd I scanned it and didn't see any surprises, i think we were all a little shell shocked after last night 17:30
stickster What happened last night? 17:30
jjmcd phone conf where we learned about publican probs 17:30
jjmcd the ones i'm trying to sort 17:30
stickster Uh 17:30
stickster Can I read about these somewhere? 17:31
jjmcd I don't think Sparks wrote it down, I posted a hack on the wiki basically 17:31
jjmcd publican makes rpms that won't pass review 17:31
stickster Ah, OK, that's been a continuing problem, and I recall that bug. 17:32
jjmcd Sparks can get a good rpm by hacking publican, but that is obvioulsy a prob 17:32
jjmcd so i am looking at a post process 17:32
mdious he could submit a patch and make a new option... "make fedora-srpm...." 17:32
jjmcd The prob is making it an option is hard 17:32
jjmcd and changing it breaks it for RH 17:32
stickster mdious: Making it an option should only be a Makefile change, right? 17:33
stickster It just needs a make rule? 17:33
mdious stickster: in the templates I think...if you use strace i believe the little makefile with each book pulls in the 'big ones' 17:34
mdious just a sec...bugging someone about 17:34
mdious it 17:34
stickster mdious: That's how I understand it. That's just how ours worked too. 17:34
mdious i think the issue is how ot make a desktop file that won't break on other distros 17:34
jjmcd I couldn't see much in the publican desktop file different from the packaging guidelines 17:35
mdious someone is thinking about the desktop file issue... 17:35
stickster Is the problem the .desktop file or the .spec file here? 17:35
jjmcd both 17:35
stickster OK, I see two problems here right now. 17:35
ke4qqq but the packaging guidelines say that if anything is changed in .desktop you need to create a new one 17:35
jjmcd the package gets the wrong name 17:36
mdious http://icanhascheezburger.files.wordpress.com/2008/08/funny-pictures-cat-is-pondering.jpg 17:36
* stickster steps back for a sec 17:36
stickster The last mention I see of this bug was quite a while ago on the docs list. 17:37
stickster If we want people to help or participate we need to see the discussion there 17:37
-!- fab [n=bellet@bellet.info] has quit ["Leaving"] 17:37
jjmcd I think Sparks thought someone would deal with it back in Jan, and was giving him some time 17:37
mdious perhaps since the second last comment is about a patch, someone is waiting for a patch? 17:38
ke4qqq last I heard in a meeting someone was tasked with finding out what the packaging guideline differences were and why RH and Fedora have different guidelines. at least that's the last I recall seeing. 17:38
stickster ke4qqq: Do you remember who was tasked with that? 17:38
mdious quaid 17:39
ke4qqq I at least remember quaid was asking the question - he may well have taken the task 17:39
* quaid sticks his head in from the community center 17:39
stickster We're kind of off in the weeds here, but quaid, take a look at the above. 17:40
stickster Let's go back to release notes now. 17:40
quaid nope, I didn't take that task. 17:40
stickster quaid: OK 17:40
mdious i think you did 17:40
stickster jjmcd: Can you take the task of (1) finding out who took that task, and (2) reporting it to the fedora-docs-list? 17:40
quaid what is happening, once again we approach zero hour and no one has cared enough to make Publican work for Fedora. 17:41
jjmcd roger 17:41
quaid mdious: hmm, well, if I did, it was probably a mistake :) 17:41
mdious quaid: I'll find the link...i want to confirm I'm not going insane ;) 17:41
stickster Let's leave the tooling aside for a bit. 17:42
stickster Let's talk about the release notes process. 17:42
stickster Tooling is important, I'm not denying that -- but we set this meeting up to help people understand this process and I want to make sure we hit that target if possible. 17:42
stickster I'm getting the feeling that people didn't have a chance to review the schedule highlights I sent. 17:43
stickster https://www.redhat.com/archives/fedora-docs-list/2009-March/msg00021.html 17:43
stickster Here's basically how things work. We have essentially two different "flavors" of release notes during a release cycle. First, there's the one-sheets. Second, there's the full-blown compendium. 17:44
stickster The one-sheets are used for the Alpha and Beta release. 17:44
stickster The compendium is created first for the Preview Release, then polished again for the GA, and then we prepare a zero-day update to catch things that didn't make it for the GA spin. 17:44
stickster Does this make sense so far? 17:45
radsy yep 17:45
jjmcd I assume for alpha and beta we also make an rpm to invlude in the package 17:45
kirkz what's GA? 17:45
stickster General Availability 17:45
stickster i.e. Final 17:45
stickster or "gold" 17:45
stickster That's the one that gets printed up, put on our download mirrors, torrents, etc. 17:46
stickster jjmcd: No, one sheets for Alpha and Beta are not packaged. 17:46
stickster They are only written on the wiki. 17:46
jjmcd OK... that simplifies it a bit 17:46
stickster The Alpha and Beta one-sheets are created there because things are too much in flux to bother with a big deal of RPM packaging, web publishing, translation, and other stuff 17:46
jjmcd At one point you talked about a diff on the wiki ... I assume you use the mediawiki export? 17:47
stickster jjmcd: Let's not talk about specific tools right now. 17:47
jjmcd ok 17:47
stickster I think it's more important to discuss the process first, then we can drill down 17:47
stickster All right, so about now, Docs team members should be emailing developers, including the f-devel-l, to make sure that devs know they can write content in the Beats pages, to be included in the "compendium" versions of the release notes. 17:48
stickster Actually, that should have already started. If it hasn't then Sparks should head up that work. 17:48
stickster I'll record it here for the summary: 17:49
stickster ACTION ITEM: Sparks -- kick off f-devel-l call for content in release notes beats on wiki 17:49
jjmcd I have poked a few developers where I knew stuff was happening that was interesting and could identify a victim 17:49
stickster Let's look at that schedule together to make this process flow easier to understand, herer's the link again: 17:49
stickster jjmcd: cool 17:49
stickster https://www.redhat.com/archives/fedora-docs-list/2009-March/msg00021.html 17:49
-!- mejla [n=mejla@75.195.broadband10.iol.cz] has quit [Read error: 60 (Operation timed out)] 17:49
stickster On April 1, we "freeze" the release notes beats on the wiki. 17:50
quaid stickster: just a process correction, shouldn't it be the relnotes lead who kicks off the f-devel-l call for content? 17:50
quaid caveat that all should be helping raise the visibility. 17:50
stickster quaid: Oops, I'm sorry, you're right, I had forgotten about the relnotes lead 17:50
stickster I apologize profusely!!! :-0 17:50
stickster OK, now I'll wipe the gravel off my knees and continue 17:51
stickster So, on 1 April we "freeze" the wiki. 17:51
ryanlerch so, are the one-sheets seperate documents? or just a summary of the current content in the proper release notes? 17:51
* ianweller isn't even sure what that means. just with release notes, right? 17:51
stickster ryanlerch: They are separate documents, only highlights. They also can be changed by the community itself as important issues are uncovered. 17:52
ryanlerch so in theory, there may be content in the one-sheets that never get over into the release notes 17:52
stickster ryanlerch: Thinking of them as a summary is probably not quite correct, since there's often not good content in the beats by that point. 17:52
stickster It would be nice if we had current Rawhide beats constantly, but that's another matter for another day. 17:53
ryanlerch maybe, that content should go into the release notes first then 17:53
stickster ryanlerch: It should go to both, yes. 17:53
ryanlerch just scanning the alpha notes, and there is content that is in there and on in the main beat pages 17:54
* quaid is about to drop offline again for a few to transit 17:54
jjmcd Some of the stuff in the one sheets seems a little expansive in the context of what else is happening in the release notes 17:54
ryanlerch i.e. the x changes 17:54
stickster Yes, from time to time the scope will probably be uneven because of the community editability. That's a price we pay for an open process. 17:54
stickster Since people use the Alpha and/or Beta for a significant amount of time after their release as "starting points" for testing (i.e. install Alpha, update to Rawhide, test), we do want to keep those flexible and editable for everyone 17:55
stickster OK, I need to make it a long way through this schedule and I'm running out of time here. 17:55
stickster Let's forge on. 17:55
jjmcd Yes, and there is good sales material there, too 17:55
stickster So, between 1-2 April we are porting the beats of the release notes to XML. 17:56
stickster This is an *intense* process because typically automated tools don't produce great DocBook XML from wiki text. 17:56
stickster It definitely gets you about 75% of the way there, though. 17:56
ryanlerch my current mockup is pretty up to date with the beats on the wiki ATM 17:56
ryanlerch http://ryanlerch.fedorapeople.org/Release_Notes/ 17:57
stickster ryanlerch: That's awesome -- if we get it into the git repo, then it will make the process much easier. 17:57
stickster So 17:57
ryanlerch stickster: that is my next task 17:57
mdious git makes things much easier? ;) ;) 17:57
ryanlerch i just have no idea with git 17:57
* rudi will be helping make that happen in the next couple of days 17:58
stickster Once we have good XML, one way or the other -- and we really need to make that happen within about 24 hours of the wiki freeze -- we generate POT files for the translators, notify the fedora-trans-list they're ready, and unfreeze the wiki. 17:58
Tsagadai ryanlerch, I'm converting my guide and notes to docbook after lunch 17:58
stickster The thing to remember about the release notes from the l10n perspective is, they're big. 17:58
ryanlerch okay, will work with Tsagadai and rudi on getting these in git 17:58
stickster For the Preview Release, we give the translators about 18-19 days 17:59
stickster On 21 April, we collect their PO (essentially this just means we look at the repo, where their submissions have already gone, we don't really have to "collect" anything), we tag the repository with a good tag like "10.93.0" (for F11 preview), and generate a SRPM for importing to the package build system. 18:00
stickster Keep in mind that, after we unfroze the wiki on 2 April, there are probably changes still happening as people discover little fixes here and there 18:00
stickster (or in some cases, a developer discovers they need a big chunk of advice somewhere) ;-) 18:01
stickster On 29 April, we basically do this routine again, but it gets easier. 18:01
stickster We freeze the wiki for the F11 Final release notes, but this time it's very easy 18:01
stickster We look for diffs between the last freeze and "now" (29 April) 18:02
stickster If there aren't any, that beat's already up to date. 18:02
stickster If there are, you port the diff to XML. In most cases, I found it easiest to do this by just copy-pasting text and applying tags as needed. 18:02
stickster *** Note: if ryanlerch's material is up to date with the wiki prior to either the Preview Release phase, or this Final phase, the job will become easier still. 18:03
stickster Just a side point. 18:03
stickster So again, on 30 April, having taken only about 24 hours to do wiki diffs and final porting, we generate POT files, notify the fedora-trans-list again, and unfreeze. 18:04
jjmcd can we watch a category and see changed pages in that category? 18:04
stickster jjmcd: I don't know, maybe ianweller does 18:04
ryanlerch i am watching all the individual pages... 18:04
stickster I prefer to put pages on my personal watch list. 18:04
ryanlerch but i may have missed soen 18:04
ryanlerch *some 18:04
jjmcd ryanlerch - that was my concern -- if a new one shows up there is some chance it will be in the beats category 18:05
stickster OK, returning to the schedule, notice that we don't give as much translation time for Final. That's because the vast quantity of material should already be done, and unchanged from the Preview Release. 18:05
stickster So the translators don't need as much time. 18:05
jjmcd And presumably the translators have the schedule too so they know what's coming 18:06
stickster jjmcd: correct! 18:06
stickster jjmcd: And the relnotes coordinator should be making sure they see it repeatedly along the way ;-) 18:06
ianweller hi! 18:06
stickster jjmcd: And if there are any conflicts, usually they're easily resolved by scootching a day or so this way or taht. 18:06
stickster *that. 18:06
ianweller jjmcd: at this moment no you can't 18:07
jjmcd so hopefully more than introductions on the l10n list 18:07
stickster ianweller: thanks 18:07
stickster jjmcd: Right. 18:07
ianweller and unless somebody wants to write some damn good PHP i'm nots eeing how that will be possible in the near future ;) 18:07
stickster jjmcd: Of course, our delivery of SRPM package is subject to when it's needed by release engineering. But this schedule was designed in cahoots with that schedule too, so it should work. 18:08
* jjmcd has been writing too much php past couple days 18:08
jjmcd yep 18:08
stickster All right, finally on 8 May, we again tag the repo and generate a SRPM package. 18:08
jjmcd Now, what happens once we have an srpm in out hot little hands 18:08
stickster "11.0.0" is a good one ;-) 18:08
stickster jjmcd: You can let me know and I'd import it into the fedora-release-notes package CVS 18:09
* quaid back and reading buffer 18:09
stickster Then it would be built and tagged by release engineering. 18:09
* laubersm has to go but will catch up in the morning 18:09
stickster G'night laubersm, thanks for coming 18:09
jjmcd OK, so once it is in the CVS then others do their magic 18:10
stickster If you look at the schedule for late May, you can see that we once again do this for a zero-day update. I won't try anyone's patience by repeating this process again :-) 18:10
stickster jjmcd: Exactly. 18:10
stickster Now... for the Preview Release you will probably need to consult with f13 to help you with the import 18:10
stickster Because..... 18:10
stickster *drumroll* 18:10
stickster I will be on vacation from about 12-April to 21-April. 18:10
jjmcd is that allowed? 18:11
stickster Allowed, schmallowed. 18:11
stickster <cartman_voice>I do what I want! Whatevah!</cartman_voice> 18:11
* jjmcd will be going to FDIM shortly before GA but I see that the boys in Dayton were nice enough to schedule that when I'm waiting for translators 18:12
stickster They're wise like that. 18:12
stickster I'm sure quaid will be around to lend guidance where needed 18:12
quaid sho'nuff 18:13
stickster As far as the SRPM goes, we definitely need to have that problem solved by about 25-March. 18:13
jjmcd It seems to me that worst case, you take the Publican output and package it manually, but if we can get Publican to do the heavy lifting ... 18:14
-!- jwb [n=jwboyer@fedora/jwb] has quit [Remote closed the connection] 18:15
quaid that looks like a good overview of the process 18:15
stickster If Publican writes good tarballs, one could write a spec template of our own to use withit. 18:15
stickster s/withit/with it/ 18:15
jjmcd filled in some holes for me 18:15
stickster It's not as hard as it sounds, but you really have to track it as it goes. 18:16
jjmcd stickster: The link is on the fpaste, but the tar is named wrong, but otherwise ok 18:16
stickster Especially, you don't want to wait until the last minute to do a build of the whole complete translated group, and find out only then that you don't know how to figure out where a translation is breaking. 18:16
jjmcd The F10 spec file is about 10 times the size of the Publican spec file 18:17
stickster What happens often is that a translator has to turn '<para>a big bunch of technical stuff</para>' into '<para>bigger buncher techniker stoof</para>' 18:17
* quaid lol 18:18
stickster If they do '<para>bigger buncher technicker stoof<para>' (see the error there?), the build breaks. And *you* have to find it and fix it. 18:18
jjmcd and of course you are doing that at the last minute 18:18
stickster (hint: the misspelling isn't the error) 18:18
stickster (that's the sv-CHEF translation) 18:18
stickster or is it sk-CHEF? I forget. 18:19
stickster Anyway 18:19
jjmcd someone threatened MORE languages for F11? 18:19
stickster jjmcd: My point is, you want to leave enough time to fix these little problems 18:19
jjmcd yep 18:19
stickster Just a word to the wise from someone non-wise who stayed up many late hours because of bad planning on his part :-) 18:19
stickster However... there's always the chance that it will be perfect and you'll be able to go out and get ice cream instead! 18:20
jjmcd I like that plan better - let's do it that way 18:20
stickster hee hee 18:20
quaid first we need autobuilding of the translations along the way so we can put the onus to test on the l10n teams 18:21
quaid but there I go being practical again! 18:21
stickster quaid: One possible problem there is that expecting translators to identify XML errors may be out of scope 18:21
jjmcd Isn't it just make html_bg_IN ? 18:21
stickster We have many brilliant translators, so it *could* work 18:21
stickster jjmcd: Right, but we're talking about automating it in a way that doesn't make them have to have Fedora, Publican, command line skills, etc.... 18:22
* stickster really needs to run now 18:22
jjmcd ahh yes 18:22
jjmcd I forget there are people that like to click 18:22
stickster Can I ask someone to take log and summary duties? I have about 1400 emails waiting to address. 18:22
quaid there are ways to make it work even with XML errors, but that requires, uh, resources and stuff. 18:22
quaid ok, then, let's close this meeting 18:23
-!- mdious [n=mmcallis@nat/redhat/x-e94a24b95a20cf90] has left #fedora-meeting [] 18:23
quaid if anyone has any more, hop to the list with it, ok? 18:23
quaid ok, in 5 18:23
quaid 4 18:23
quaid 3 18:23
quaid 2 18:23
quaid 1 18:23
quaid </meeting> 18:23

Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!