06:30:57 #startmeeting i18n 06:30:57 Meeting started Wed Apr 11 06:30:57 2018 UTC. The chair is paragan. Information about MeetBot at http://wiki.debian.org/MeetBot. 06:30:57 Useful Commands: #action #agreed #halp #info #idea #link #topic. 06:30:57 The meeting name has been set to 'i18n' 06:30:57 #meetingname i18n 06:30:57 The meeting name has been set to 'i18n' 06:30:57 #topic agenda and roll call 06:30:57 #link https://fedoraproject.org/wiki/I18N/Meetings/2018-04-11 06:31:00 hi.. 06:31:05 hi 06:31:07 hi 06:31:13 hi 06:31:38 #chair suanand tagoh_ bhavin192 vishal_vvr epico fujiwarat juhp pravins 06:31:38 Current chairs: bhavin192 epico fujiwarat juhp paragan pravins suanand tagoh_ vishal_vvr 06:32:03 hi 06:32:35 hi 06:33:12 Hi all 06:33:17 hi 06:33:25 let's start our bi-weekly meeting 06:33:46 #topic Upcoming schedule 06:33:46 #info 2018-04-17 Final Freeze (*) 06:33:46 #info 2018-05-01 Fedora 28 Final Release (GA) (Preferred Target) 06:33:46 #info 2018-05-08 Fedora 28 Final Release (GA) (Target #1) 06:34:05 we will soon be in F28 Final Freeze 06:34:21 if anyone still got any bug to fix then please do so 06:34:54 bodhi will also take time to give auto push option 06:35:50 Has anyone got any issues to discuss that needs attention? 06:37:13 CJK fonts 06:37:52 do we have any corresponding bug for that? 06:38:18 #info the current blocker list is at https://qa.fedoraproject.org/blockerbugs/milestone/28/final/buglist 06:39:44 I think some discussion happened in last Workstation WG meeting 06:40:24 and one of the proposed solution needs some packaging changes 06:40:44 juhp, has sent that email on fonts list as well as packaging list 06:41:04 sorry 06:41:20 well we have a bug about subpackaging Noto CJK fonts 06:41:25 yes 06:41:31 #info https://lists.fedoraproject.org/archives/list/fonts@lists.fedoraproject.org/thread/3GTJURMJ5HJLKCYGYS6QETA2LRJDQLFJ/ 06:42:54 Overall the subpackaging idea looks difficult 06:43:01 #info for CJK fonts size issue reported bug is https://bugzilla.redhat.com/show_bug.cgi?id=1562315 06:43:09 to be accepted i mean - I already have a test copr 06:43:27 paragan: yes thanks 06:44:57 Note that https://bugzilla.redhat.com/show_bug.cgi?id=1561763 was accepted as a release blocker 06:45:24 aha 06:45:58 right its already accepted as blocker 06:47:04 I think work is going on on this issue by juhp and tagoh_ 06:47:41 hopefully we will get some fix for this issue this week only 06:47:55 So basically we are obliged to reduce the font footprint to around the level of f27 06:48:07 at least for the Live images I think 06:50:10 I thought it would be good to discuss here.... 06:51:23 assuming we can't subpackage or do nothing the main remaining options are: 06:51:40 (1) revert to f27 fonts 06:52:10 (2) use the OTC fonts 06:52:27 (3) remove weights from the Live media 06:53:41 tagoh_ and Owen think (2) should be left for F29, which is probably wise... 06:54:05 (1) is probably the conversative approach 06:54:43 conservative 06:54:48 yes, 2 is a bit risky at this moment. less time to test and make sure if ttc works well. 06:55:17 True, testing for Live apps would not be that hard 06:56:18 of course nothing to stop people from reverting font themselves - also updates will not get the new fonts 06:56:26 upgrades 06:56:50 3 is.. well, it sounds nice idea though, how can one get the full sets of fonts after installing live image? do we need to bump the release every releases? need to think about it more. this may be for f29+ 06:56:51 But I don't deny there may be some regressions 06:57:38 One would get the full set by updating 06:57:55 But I only think we should do that for F28 06:59:18 hi all 07:00:18 this will be happened only for live. so the packages got removed some files in live needs to just rebuilt without real changes. bump a release to let dnf detect an update. 07:00:36 actually the upgrade story is "interesting" since we are reducing the priority of the old default fonts but hopefully it won't impact too much 07:01:12 tagoh_: right 07:01:45 I think we have not tested upgrades yet? Anyone? 07:02:18 well easy enough to test by removing the new fonts and installing the old ones... 07:02:54 not tested upgrades 07:03:17 I am stupid enough to be willing to do a PR for the kickstart change for f28 if people think it is acceptable and workable? 07:03:56 Otherwise we should mostly revert to the f28 package set or equivalent 07:04:03 erm F27! 07:04:38 As such we could keep the Noto fonts for Chinese but without Mono 07:05:24 epico: I was trying to ask you earlier about reverting to Noto or Adobe.... 07:06:06 pravins: any thoughts? dunno how much you have followed this discussion until now 07:06:08 fedora live media have integrity check, what does it check? 07:06:23 The checksum of the image? 07:06:27 sums? 07:06:36 juhp: maybe the same as F27 is easier than Noto. 07:06:43 yes 07:06:53 I mean in boot loader. 07:07:01 it can check the live media. 07:07:08 It's true though - removing the weights will give rpm -V warning on the packages of course 07:07:18 epico: nod 07:07:31 aha. right.. 07:07:51 but i think that is acceptable - or the least of our problems 07:07:59 * epico dunno whether will affect the media check in boot loader. 07:08:05 it won't 07:08:12 okay, I see 07:08:17 check is about image integrity 07:08:23 not afaik anyway 07:08:26 s/whether/whether it/ 07:09:07 ie whether the downloaded image is correct - nothing to do with rpm content 07:09:13 anyway 07:09:17 aha, I see 07:09:37 revert to Noto may create new bug... 07:09:38 there are other non-rpm changes on images anyway 07:09:43 okay 07:10:05 got it 07:12:50 https://pagure.io/fedora-kickstarts/blob/master/f/fedora-live-base.ks#_319 07:14:49 A little hard to test though, in the sense that harder to make Live image these days 07:16:57 let's move to next agenda if nothing more need to be discussed here. 07:17:00 ls /usr/share/fonts/google-noto-cjk/NotoS* | grep -v Regular | xargs rm -rf 07:17:09 would work I think 07:18:16 ah need test on the dir I guess 07:18:28 anyway should we make a decision now? 07:18:56 +1 for the above solution 07:19:26 How many variants we were shipping in F27. 07:19:38 It we only provide those many, size should not affect much, am i right? 07:19:46 7 variants for Chinese only. 07:19:57 given that proposing modifying .ks, who's responsible for that change to be accepted? how long do we need to wait to see go or no-go with it? 07:20:12 epico: and how much size for it? 07:20:33 tagoh_: given that is a blocker issue now - I guess we have a little time 07:20:51 (Not saying we should wait and do nothing :) 07:21:07 pravins: https://bugzilla.redhat.com/show_bug.cgi?id=1562315#c2 07:21:13 * pravins clicks 07:21:19 juhp: is it going to be accepted? 07:21:29 shrug maybe maybe not 07:21:39 the size information is in the link. 07:21:44 if that is the only easy option then why not try that? 07:22:00 there is no size breakdown by lang alas 07:22:13 but Chinese is by far larger I should think 07:22:33 we can add the numbers after 07:22:50 paragan: even if it works, it doesn't mean it will be accepted right? 07:22:59 pravins: the rough iso increase is probably around 260MB or something like that 07:23:11 excluding the glibc locales increase 07:23:24 pravins: I mean F27 -> F28 for CJK 07:23:24 well, reverting to f27's also takes a time to make changes back. just worrying about the worst case. 07:23:39 tagoh_: yep 07:23:44 We do have some Autobold features from freetype. 07:23:54 Yep 07:23:56 If size is issue, we can skip Bold variant. 07:24:08 and trust Freetype autobold for sometime. 07:24:09 Bold on Bold in Libreoffice is awesome ;-P 07:24:27 🤣 07:24:30 pravins: dropping bold will actually reduce the size 07:24:37 tagoh_, just propose that change to fedora-kickstarts project, discuss with upstream on why it is important and request to accept it. 07:24:43 there are more details in the WG ticket.... 07:24:44 pravins: emboldening is better than nothing but unlikely it is better than bold provided by upstream... 07:24:59 also inform this in WG ticket as well 07:25:16 Yes I am planning to add a note to the WG ticket 07:25:23 tagoh_: i agree, Bold provided by upstream is best choice. Since we have found issues for Lohit with emboldening. 07:25:33 https://pagure.io/fedora-workstation/issue/46#comment-504344 07:25:51 paragan: discuss with upstream - in github?? 07:25:58 and, i also had one concern. If i am installing in en locale, do i need all CJK fonts? Are we installing it? 07:26:15 pravins: as you know 07:26:50 yes we are installing 07:26:59 since ? :-) 07:27:10 Our goal was to provide at least one font to cover Unicode BMP. 07:27:14 not many :) 07:27:20 need well that is another discussion not for today 07:27:33 pravins: ah valid 07:27:45 sorry misread "all" hmm well 07:27:48 sub-packaging and not installing Bold in default install is good option. AFAI think. 07:27:53 well I also proposed to drop Serif 07:28:00 lets user install Bold with langpacks. 07:28:05 pravins: we can't subpackage... 07:28:23 my bad, i missed that thread :() 07:28:23 that is the sad problem - well unless we get an exception 07:28:29 np 07:28:39 https://pagure.io/fedora-kickstarts 07:28:47 juhp, I mean https://pagure.io/fedora-kickstarts 07:28:55 ah 07:29:00 that upstream ;) 07:29:02 hehe 07:29:04 okay 07:29:06 :) 07:29:20 I thought you meant Google Noto 07:29:24 no no 07:29:27 did anyone replied to juhp email on fonts SIG? 07:29:38 nope I don't think 07:29:43 feel free ;o) 07:30:14 anyway nim-nim already expressed he does not think it is a good idea here 07:30:19 yeah, considering the CJK fonts size. I think exception should be allowed. 07:30:25 on devel list 07:30:36 yeah we may get it later ;o) 07:30:43 who knows or not 07:31:14 Shall we wait to Friday if the PR is accepted or not? 07:31:23 for kickstarts 07:32:00 i will reply on fonts sig list 07:32:33 that still leaves at least 1 week to revert 07:33:56 just wonder if kickstarts people will ponder what happened.. we may need to provide pros and cons about it perhaps. 07:33:57 well it's a little tight but seems possible 07:34:09 Alternatively we start to revert now 07:34:21 sure we will in the PR 07:34:34 or sharing PR with more people will get wider audience? 07:34:35 needs good reason to do this 07:34:39 hmm 07:35:08 We could open a bug and propose it as a blocker 07:41:56 okay 07:42:34 I suppose we can continue to discuss this later on 07:42:56 let's go thru' tickets and close the meeting 07:42:58 #topic No assignee issues 07:42:59 #info #83: Tracker for Fedora 28 Changes (pnemade) 07:42:59 #link https://pagure.io/i18n/issue/83 07:43:34 I think there is nothing to discuss here and probably I will not pick this again 07:43:44 I will close this once Fedora 28 is released 07:43:53 #info #78: package Liberation Classic Font in Fedora (pnemade) 07:43:54 #link https://pagure.io/i18n/issue/78 07:44:20 I see pravins is doing some progress here 07:44:52 so let's wait for him to finish his work and start ours of packaging it 07:45:14 #info #85: Docs Beats for Fedora 28 (pnemade) 07:45:15 #link https://pagure.io/i18n/issue/85 07:45:33 anyone want to provide text documentation here? 07:46:13 otherwise we can skip this for Fedora 28 07:47:55 I will mark this closed in next week 07:48:09 #info #95: change this meeting to another day (petersen) 07:48:10 #link https://pagure.io/i18n/issue/95 07:48:26 I have not seen any more responses submitted here. 07:48:28 #chair 07:48:28 Current chairs: bhavin192 epico fujiwarat juhp paragan pravins suanand tagoh_ vishal_vvr 07:48:46 please submit your responses to the link given in the ticket before next meeting 07:49:47 #info next meeting discuss the new meeting timing based on responses submitted for whenisgood event 07:50:11 #info #84: Fedora 26 Bug triaging (pnemade) 07:50:11 #link https://pagure.io/i18n/issue/84 07:50:30 #info good to see less bugs here 07:51:36 #topic Open Floor 07:51:59 I think we are overtime today 07:52:07 thanks all for coming to meeting 07:52:12 #endmeeting