Chat log from the meeting on 2023-11-14

From OpenSimulator

Jump to: navigation, search
[11:59 AM PT]  Andrew Hellershanks: Hello, everyone.
[12:00 PM PT]  Jagga Meredith: hey Clifford
[12:00 PM PT]  Andrew Hellershanks: I'm running with the latest (6.6.14) version of Firestorm.
[12:00 PM PT]  Vincent.Sylvester Welcome to the club
[12:00 PM PT]  Clifford.Hanger any issues yet?
[12:01 PM PT]  Ubit Umarov: don't remember major issues with 6.6.14
[12:02 PM PT]  Andrew Hellershanks: Clifford: No, but I haven't been inworld long enough with this viewer to notice any issues.
[12:02 PM PT]  Gavin.Hird that was the August update, right?
[12:02 PM PT]  Clifford.Hanger we have lotsa users on it already
[12:02 PM PT]  Andrew Hellershanks: I don't know when it was released.
[12:02 PM PT]  Ubit Umarov: think current version
[12:02 PM PT]  Vincent.Sylvester Same ole cache glitches swallowing textures or objects, but that's not on them, LL broke that fair and square
[12:02 PM PT]  Clifford.Hanger hey Joe
[12:02 PM PT]  Jagga Meredith: I'm running  Is there something newer?
[12:03 PM PT]  Joe Magarac: Hi.
[12:03 PM PT]  Ubit Umarov: well there is the beta
[12:03 PM PT]  Andrew Hellershanks: Yes, August 1 release. I just checked the dates on the file.
[12:03 PM PT]  Jagga Meredith: ag
[12:03 PM PT]  Gavin.Hird no newer release version
[12:03 PM PT]  Ubit Umarov: and the alpha :)
[12:04 PM PT]  Ubit Umarov: beta is 6.??.?? version adding inventory changes like thimbnails etc
[12:04 PM PT]  Ubit Umarov: alpha is 7.x.x  pbr viewer
[12:04 PM PT]  Ubit Umarov: thimb and thunb
[12:04 PM PT]  Gavin.Hird seems liek all visible LL viewer development is talled
[12:04 PM PT]  Gavin.Hird they have not posted mch at all for 2-3 weeks
[12:05 PM PT]  Gavin.Hird stalled*
[12:05 PM PT]  Ubit Umarov: well fs needs to follow it
[12:05 PM PT]  Ubit Umarov: ll servers will misbehave with older versions
[12:06 PM PT]  Ubit Umarov: like the beta did remove most lludp inventory for sl..   but seems to have it for us
[12:06 PM PT]  Ubit Umarov: we just see non working things abotu thumbnails
[12:06 PM PT]  Clifford.Hanger any progress on voice?
[12:06 PM PT]  Vincent.Sylvester The horror not having thumbnails will be hard to bear /s
[12:07 PM PT]  Gavin.Hird not a pip
[12:07 PM PT]  Ubit Umarov: in fact could not even support those since it does not have support for them on the lludp code paths
[12:07 PM PT]  Clifford.Hanger he asks hopefully?
[12:07 PM PT]  Clifford.Hanger hi Lyr
[12:07 PM PT]  Lyr Lobo: Hello *grins*
[12:07 PM PT]  Gavin.Hird Hi Lyr
[12:07 PM PT]  Andrew Hellershanks: Hello, Lyr
[12:07 PM PT]  Lyr Lobo: Great to see you
[12:08 PM PT]  Ubit Umarov: not that i kown clifford.Hanger
[12:08 PM PT]  Andrew Hellershanks: Hello, Jamie.
[12:08 PM PT]  Clifford.Hanger hi jamie
[12:09 PM PT]  Jamie.Jordan Hi everybody
[12:09 PM PT]  Lyr Lobo waves and grins
[12:09 PM PT]  Ubit Umarov: a spacewoman came traveling...
[12:09 PM PT]  Gavin.Hird Hi Jamie
[12:09 PM PT]  Ubit Umarov: on a ship for afar..
[12:09 PM PT]  Ubit Umarov: err coff coff
[12:09 PM PT]  Ubit Umarov: :)
[12:09 PM PT]  Clifford.Hanger i wonder if microsoft could be approached to create something generic?
[12:09 PM PT]  Andrew Hellershanks: Ubit, changing the gender of the song. :)
[12:10 PM PT]  Lyr Lobo chuckles
[12:10 PM PT]  Andrew Hellershanks: Clifford, possibly but only if they would agree to keep it open source.
[12:10 PM PT]  Gavin.Hird naw, they'll just stuff it full of cringe AI and will make it useless
[12:10 PM PT]  Vincent.Sylvester xD
[12:10 PM PT]  Andrew Hellershanks: hehe
[12:11 PM PT]  Clifford.Hanger if they are that bad, why are we using dotnet
[12:11 PM PT]  Ubit Umarov: bc they own mono and killed i
[12:11 PM PT]  Orbert.Tatham Because that is where we started
[12:11 PM PT]  Gavin.Hird beats me, but someone made the decision in the last century almost
[12:11 PM PT]  Orbert.Tatham They bought Mono and killed it, as expected
[12:12 PM PT]  Vincent.Sylvester Find an alternative that offers the same cross platform capability out the box along with the now critical features for how the entire OpenSim architecture works internally
[12:12 PM PT]  Orbert.Tatham We just didn't move before they killed it, and are stuck with their leavings
[12:12 PM PT]  Vincent.Sylvester But wait there is an option, let's rewrite it in Java, that's cross platform right xD
[12:12 PM PT]  Gavin.Hird they just dropped an incremental update to dotnet6 today
[12:12 PM PT]  Ubit Umarov: i was thinking cobol or pl/i
[12:12 PM PT]  Ubit Umarov: hmm ada?
[12:13 PM PT]  Gavin.Hird why not FORTRAN?
[12:13 PM PT]  Vincent.Sylvester Fortran 2 at least
[12:13 PM PT]  Orbert.Tatham on Vaxen, too
[12:13 PM PT]  Gavin.Hird :-)
[12:13 PM PT]  Clifford.Hanger well that would end one of our dody reliances
[12:13 PM PT]  Clifford.Hanger dodgy
[12:13 PM PT]  Orbert.Tatham All of those things cost money to do, Cliff
[12:13 PM PT]  Ubit Umarov: hmm opensim in fortran on a cray xmp
[12:13 PM PT]  Andrew Hellershanks: I used to do a bit of programming in WATFIVS.
[12:14 PM PT]  Orbert.Tatham and don't happen quickly
[12:14 PM PT]  Vincent.Sylvester If you rid yourself of all dependencies you just re-invent the wheel and make it much harder to keep everything in working order. The reason these frameworks and runtimes exist is to make it simpler to get things out the door
[12:14 PM PT]  Ubit Umarov: ( no i would not pay the cray power bill)
[12:14 PM PT]  Orbert.Tatham Vax ORTRAN at least had strings
[12:14 PM PT]  Gavin.Hird This should teach them:
[12:14 PM PT]  Clifford.Hanger that sounds like a call for an early start
[12:14 PM PT]  Gavin.Hird IBM 1401 runs Fortran 2
[12:15 PM PT]  Ubit Umarov: on that.. minor changes on code the "cosmetics" class of changes  during last week
[12:16 PM PT]  Ubit Umarov: hmm did add a null check on part.inventory that seems was been triggerrd
[12:16 PM PT]  Ubit Umarov: with bad consequences for the region ;)
[12:16 PM PT]  Gavin.Hird did you push that to bitbucket?
[12:16 PM PT]  Ubit Umarov: checking
[12:17 PM PT]  Gavin.Hird now you dud :-)
[12:18 PM PT]  Gavin.Hird did
[12:18 PM PT]  Ubit Umarov: fun,, did but dont see it there
[12:18 PM PT]  Ubit Umarov: ohh need to select the branch to see it.. odd
[12:18 PM PT]  Joe Magarac: What is the status of Mono? It's open source, so it can't just go away all at once.
[12:18 PM PT]  Ubit Umarov: ms killed it.. dead end
[12:18 PM PT]  Gavin.Hird yep
[12:18 PM PT]  Vincent.Sylvester You can ask on their discord how it is doing, the sentiment is grim
[12:19 PM PT]  Orbert.Tatham MS had staff maintaining it - they moved to dotnet
[12:19 PM PT]  Vincent.Sylvester MS killed it under their noses basically
[12:19 PM PT]  Ubit Umarov: they used it to make dotnet doe linux etc
[12:19 PM PT]  Ubit Umarov: then killed it
[12:19 PM PT]  Clifford.Hanger hi Arielle
[12:19 PM PT]  Gavin.Hird Hi Arielle
[12:19 PM PT]  Andrew Hellershanks: Hello, Arielle.
[12:19 PM PT]  Lyr Lobo: Hello Arielle!
[12:20 PM PT]  Gavin.Hird Joe, I compiled and ran your UI code on macOS
[12:20 PM PT]  Gavin.Hird not a hitch
[12:20 PM PT]  Arielle Popstar: hiyas
[12:20 PM PT]  Arielle Popstar: this region seems a mess
[12:21 PM PT]  Clifford.Hanger seems ok to me?
[12:21 PM PT]  Arielle Popstar: maybe the yengine thing?
[12:21 PM PT]  Ubit Umarov: most likely your glasses..  but details
[12:21 PM PT]  Ubit Umarov: :p
[12:21 PM PT]  Arielle Popstar: maybe my ao script is old
[12:22 PM PT]  Vincent.Sylvester Anyways, no news on xunit either. I been talking to someone more familiar with testing on how to approach setting them up. Probably will be a few more weeks to have a battle plan for that
[12:23 PM PT]  Arielle Popstar: so i was noticing the new inventory preview image in FS isnt supported in Osgrid sandbox.
[12:23 PM PT]  Arielle Popstar:
[12:23 PM PT]  Arielle Popstar: any plans to put that in?
[12:24 PM PT]  Vincent.Sylvester You don't read the meeting logs do you, no not planned at the moment
[12:24 PM PT]  Ubit Umarov: not in imediate future. FS does not have support for it on opensim either
[12:24 PM PT]  Arielle Popstar: I do on occasion but usually I come to the meeting and hear it first hand unless it cases someone stress
[12:24 PM PT]  Lyr Lobo: It was reviewed, Arielle and I think that until the viewer supports it, no
[12:24 PM PT]  Andrew Hellershanks: Regarding libraries ->   ;)
[12:25 PM PT]  Arielle Popstar: the viewer does support it though
[12:25 PM PT]  Lyr Lobo: in Opensimulator?
[12:25 PM PT]  Arielle Popstar: works in sl
[12:25 PM PT]  Ubit Umarov: uff yes.. does support it IN SL
[12:25 PM PT]  Arielle Popstar: yes
[12:25 PM PT]  Orbert.Tatham Find it in the OpenSim version and you get a cookie
[12:25 PM PT]  Arielle Popstar: i see it
[12:27 PM PT]  Arielle Popstar:
[12:28 PM PT]  Arielle Popstar: if you on the beta viewer r-click on a folder and you are given the option of taking a preview image
[12:28 PM PT]  Ubit Umarov: and then viewer does nothing with it on opensim..
[12:28 PM PT]  Arielle Popstar: this will be a huge boon to residents  with browsing their inventory
[12:29 PM PT]  Ubit Umarov: because it was not added to that API we use
[12:29 PM PT]  Arielle Popstar: well because it requires you adding some code?
[12:29 PM PT]  Ubit Umarov: no no point o adding code on our side ( and mess up inventory dbs with useless mess )
[12:29 PM PT]  Ubit Umarov: so no point..
[12:29 PM PT]  Arielle Popstar: no, it helps people with their inventory
[12:30 PM PT]  Vincent.Sylvester Didn't need it for a decade, why now, name your things properly to remember what they are
[12:30 PM PT]  Arielle Popstar: ok for people with only one outfit but for those of us with lots of clothes it will be a huge improvement
[12:30 PM PT]  Orbert.Tatham Arielle, as always, if you want the feature, write it and submit it. That is how this works, it is called open source.
[12:30 PM PT]  Ubit Umarov: yeah extra lag everytime viewers decide to cache it
[12:31 PM PT]  Arielle Popstar: there is also lag when one has to rezz out a whole clothing item to see what it is rather then a small preview image
[12:31 PM PT]  Lyr Lobo: I'm going to have to downsize my inventory if we offer it or become the lag beast
[12:31 PM PT]  Ubit Umarov: well as i said, current viewers do not have the code to update it on out inventory API
[12:31 PM PT]  Orbert.Tatham Yeah, but you chose to rezz it out
[12:31 PM PT]  Andrew Hellershanks: Vincent, naming things properly and using folders to help keep things organized.
[12:32 PM PT]  Vincent.Sylvester Something worth doing for anything, not just inventory
[12:32 PM PT]  Andrew Hellershanks: Very true
[12:32 PM PT]  Arielle Popstar: the viewer you are likely using Ubit has it
[12:32 PM PT]  Ubit Umarov: well im not going to repeat what i already told
[12:32 PM PT]  Clifford.Hanger hi cuga
[12:32 PM PT]  Arielle Popstar: shrugs men change once a year if that, for some of us it is multiple times a day
[12:32 PM PT]  Cuga.Rajal Hi, sorry late
[12:33 PM PT]  Andrew Hellershanks: Welcome, Cuga.
[12:33 PM PT]  Ubit Umarov: mayve dayturn will add the missing api functions ;)
[12:33 PM PT]  Ubit Umarov: well tyhe api is there.. not the extra fileds
[12:33 PM PT]  Gavin.Hird nobody in opensim have huge amounts of clothes
[12:34 PM PT]  Arielle Popstar: it looks to me like it has to be added server side?
[12:34 PM PT]  Ubit Umarov: fields
[12:34 PM PT]  Arielle Popstar: yeah right  Gavin
[12:34 PM PT]  Gavin.Hird have you seen anyone where chaning clothes the last 10 years?
[12:34 PM PT]  Ubit Umarov: yes it wil need a lot of server side code... and DBs changes..
[12:34 PM PT]  Ubit Umarov: usless if VIEWERS DO NOT USE IT!
[12:34 PM PT]  Arielle Popstar: but the viewer is capable
[12:34 PM PT]  Arielle Popstar: should I ask Beq?
[12:35 PM PT]  Orbert.Tatham Arielle, write it and submit it. We aren't going to, it is very simple
[12:35 PM PT]  Arielle Popstar: no, it is something that needs to be added to core
[12:35 PM PT]  Arielle Popstar: thats why i am here
[12:36 PM PT]  Arielle Popstar: which is where these things are discussed
[12:36 PM PT]  Vincent.Sylvester Well you be waiting a long time then.
[12:36 PM PT]  Orbert.Tatham And the core doesn't think we need to add it
[12:36 PM PT]  Orbert.Tatham Sorry about your luck (NOT)
[12:36 PM PT]  Arielle Popstar: well then is Opensim a dead project?
[12:36 PM PT]  Orbert.Tatham You are free to leave
[12:36 PM PT]  Arielle Popstar: I'll let everyone know
[12:36 PM PT]  Orbert.Tatham and call it dea
[12:36 PM PT]  Orbert.Tatham dead
[12:36 PM PT]  Orbert.Tatham but that doesn't make it dead
[12:36 PM PT]  Clifford.Hanger it maybe if the voice plug is pulled
[12:37 PM PT]  Lyr Lobo: Arielle, we will have a Viewer Panel at the OpenSimulator Community Conference to discuss the viewer  and the future
[12:37 PM PT]  Orbert.Tatham Cliff, we have several time told you what it would take to deal with that
[12:37 PM PT]  Arielle Popstar: i dont see any development from Unbits link in his profile
[12:37 PM PT]  Arielle Popstar: Ubits
[12:37 PM PT]  Clifford.Hanger i realise that
[12:37 PM PT]  Lyr Lobo: It is at 12:30 pm on Saturday, December 9
[12:37 PM PT]  Gavin.Hird The difference Arielle is that opensim does not have 20  salaried developers working on server side only
[12:37 PM PT]  Arielle Popstar: whose fault is that?
[12:37 PM PT]  Clifford.Hanger is why i try to think out of the box
[12:38 PM PT]  Orbert.Tatham Like we have told you, if you want it, write it and submit it. Just because you want it does not mean that the core developers think it is needed
[12:38 PM PT]  Arielle Popstar: isnt that because  core is closed to other developers?
[12:38 PM PT]  Orbert.Tatham Not at all
[12:38 PM PT]  Lyr Lobo: This is an opensource project
[12:38 PM PT]  Orbert.Tatham Core is not paid
[12:38 PM PT]  Arielle Popstar: are you core Orbert?
[12:38 PM PT]  Orbert.Tatham You are welcome if you can develop
[12:38 PM PT]  Orbert.Tatham No, I am not good enough, I am just learning
[12:39 PM PT]  Orbert.Tatham I will contribute if and when I come up with something that works
[12:39 PM PT]  Orbert.Tatham And it will be up to the others to decide
[12:39 PM PT]  Andrew Hellershanks: Arielle, You should have learned by now how development works and that it is being done by volunteers.
[12:39 PM PT]  Orbert.Tatham That is how this project works
[12:39 PM PT]  Arielle Popstar: good luck getting them to accept it
[12:40 PM PT]  Orbert.Tatham Arielle, first I have to figure out what is really needed and what is fluff dreamed up by former Micro$haft twits at Linden
[12:40 PM PT]  Arielle Popstar: so no new features for opensim?
[12:41 PM PT]  Orbert.Tatham We aren't doing them just because you demand them
[12:41 PM PT]  Arielle Popstar: at least there is some new feature development
[12:41 PM PT]  Lyr Lobo: We are going to discuss the software and viewer at the OpenSimulator Community Conference
[12:41 PM PT]  Arielle Popstar: its supposed to be discussed here pprior to that
[12:41 PM PT]  Orbert.Tatham There are ways to get things done, Arielle. Demanding them isn't one of them
[12:42 PM PT]  Arielle Popstar: are core spokesperson Orbert?
[12:42 PM PT]  Andrew Hellershanks: Arielle, Even if there was it seems likely it wouldn't a feature you are interested in. If you really want a particular feature and think money is the way to get it you can try and hire someone to implement whatever feature you want to see added.
[12:42 PM PT]  Lyr Lobo: No, we discuss current changes, future directions, and the plans and needs of our viewer developers and the community
[12:42 PM PT]  Orbert.Tatham I am someone with a brain, Arielle
[12:42 PM PT]  Lyr Lobo: it is a little different. This is a public log for community involvement
[12:42 PM PT]  Arielle Popstar: well i am community :)
[12:43 PM PT]  Lyr Lobo: and yes
[12:43 PM PT]  Orbert.Tatham Then perhaps it would benefit you to try listening
[12:43 PM PT]  Arielle Popstar: I been coming here about 12 years  so have heard  it
[12:43 PM PT]  Orbert.Tatham Ok, then when are you going to understand it
[12:44 PM PT]  Orbert.Tatham ?
[12:44 PM PT]  Arielle Popstar: i already do and know how it works beyond what you are saying
[12:44 PM PT]  Vincent.Sylvester Evidently not
[12:44 PM PT]  Orbert.Tatham Quite obviously false statement
[12:45 PM PT]  Lyr Lobo: well it is good to know that you like the thumbnail feature in SL and find it useful
[12:45 PM PT]  Lyr Lobo: let us know whether you experience any changes in performance in congested areas, etc.
[12:45 PM PT]  Arielle Popstar: same ol same ol, the non core pople would like to dictate how it should work
[12:45 PM PT]  Lyr Lobo: that would be you, Arielle
[12:45 PM PT]  Orbert.Tatham Yes, you are trying to
[12:46 PM PT]  Arielle Popstar: saw loading in lbsa taking longer
[12:46 PM PT]  Lyr Lobo: Your information is invaluable as without testing, we would not know how well it works
[12:46 PM PT]  Arielle Popstar: but then so is sl
[12:46 PM PT]  Lyr Lobo: that's good to know
[12:46 PM PT]  Orbert.Tatham As expected
[12:46 PM PT]  Lyr Lobo: you discover very useful usability issues
[12:46 PM PT]  Lyr Lobo: like that log file from a few months back
[12:46 PM PT]  Arielle Popstar: well i did test the image preview function and it isnt working
[12:47 PM PT]  Gavin.Hird it is not working, but we already knew that
[12:47 PM PT]  Arielle Popstar: ok good so then we should expect changes to resolve it soon?
[12:47 PM PT]  Gavin.Hird no
[12:47 PM PT]  Orbert.Tatham Ask SL, they are the ones that care
[12:48 PM PT]  Gavin.Hird ...does it not need AIS to work?
[12:48 PM PT]  Andrew Hellershanks: Arielle, When LL comes up with some new thing it doesn't mean that the OpenSim code and viewers will immediately implement that latest feature. You should know that by now.
[12:48 PM PT]  Arielle Popstar: and Opensim chooses to follow sl's footsteps
[12:48 PM PT]  Cuga.Rajal no, there is a calculated evaluation first
[12:49 PM PT]  Ubit Umarov: yes, they only added it to AIS code paths
[12:49 PM PT]  Clifford.Hanger Orbert, I hear what you say about telling me what needs doing, my point is that that is a good starting point, but I worry thay it doen not seem to have a high prioity, we have residents worried about the future of voice?
[12:49 PM PT]  Arielle Popstar: to see if it is convenient for the devs and not too nmuch work?
[12:49 PM PT]  Ubit Umarov: not to the llUPD ones
[12:49 PM PT]  Gavin.Hird right, so we are not going to have AIS support - probalby ever Arielle
[12:50 PM PT]  Andrew Hellershanks: Can we drop this for now? We are near the top of the hour.
[12:50 PM PT]  Arielle Popstar: tAIS?
[12:50 PM PT]  Arielle Popstar: AIS?
[12:50 PM PT]  Orbert.Tatham Cliff, voice is something that you are worried about - the current server/backend venddor has chosen to raise prices beyond your pain threshold, but we can't fix that
[12:50 PM PT]  Gavin.Hird yesw AIS, LL's inventory interface API
[12:50 PM PT]  Andrew Hellershanks: I would like to have Lyr give us an update about the upcoming OSCC event that is now just a few weeks away.
[12:50 PM PT]  Clifford.Hanger I do need to heads out very soon, it is not a reflection on this meeting
[12:51 PM PT]  Andrew Hellershanks: If everyone will hold off on further comments for the moment. Go ahead Lyr.
[12:51 PM PT]  Lyr Lobo: Great! The community conference is Dec 9-10 with a starting party at noon on Friday, the 8th
[12:51 PM PT]  Andrew Hellershanks: (I'll brb. RL is calling)
[12:51 PM PT]  Lyr Lobo: we have 30 sessions from 7 am until 6 pm and dancing in the venings
[12:51 PM PT]  Lyr Lobo: thanks, Andrew
[12:51 PM PT]  Lyr Lobo: Core Dev speaks at 7 am on the 9th and we have a VIP Core Dev session at 11:45 on that saturday. The Viewer panel is right after it
[12:52 PM PT]  Arielle Popstar: whats the theme?
[12:52 PM PT]  Lyr Lobo: We have quite a few research, education, entertainment, and the arts sessions as well as a few tool integrations with other worlds and technologies
[12:52 PM PT]  Lyr Lobo: It is the OpenSimulator Community Conference, so all topics related to Opensim are welcome, including the future, ARielle
[12:53 PM PT]  Lyr Lobo: Unlike my other conferences, we do not limit it to a theme
[12:53 PM PT]  Arielle Popstar: in past there was a particular theme
[12:53 PM PT]  Lyr Lobo: in 2015, we did 5 tracks
[12:53 PM PT]  Arielle Popstar: ok
[12:53 PM PT]  Lyr Lobo: for business, technology, research, education, creativity and the arts
[12:53 PM PT]  Lyr Lobo: but they ran at the same time
[12:53 PM PT]  Lyr Lobo: this way, everything is visible to all from the 4 regions on the OSCC Keynotes
[12:54 PM PT]  Clifford.Hanger thanks everyone, i will read the logs to find more on the conf
[12:54 PM PT]  Clifford.Hanger ty Lyr
[12:54 PM PT]  Lyr Lobo: the program is being developed and will be online in a couple of weeks. Here's the link to register
[12:54 PM PT]  Clifford.Hanger thanks everyone
[12:54 PM PT]  Lyr Lobo:
[12:54 PM PT]  Lyr Lobo: come join us *grins*  It is not a community without you
[12:54 PM PT]  Clifford.Hanger bye
[12:54 PM PT]  Lyr Lobo: bye Clifford
[12:55 PM PT]  Lyr Lobo: This is our 11th year
[12:55 PM PT]  Lyr Lobo: our goal is to strengthen the community, see what everyone is doing and where we are growing
[12:56 PM PT]  Lyr Lobo: Maria will have a State of the Metaverse results, but we need more people to participate in it
[12:56 PM PT]  Lyr Lobo: from Hypergrid Business, she hosts a survey
[12:57 PM PT]  Cuga.Rajal Looking forward to that
[12:57 PM PT]  Lyr Lobo: We are excited to have everyone participate  *grins*
[12:57 PM PT]  Lyr Lobo: a few new releases will be unveiled at the conference. Will let the presenters share that
[12:59 PM PT]  Vincent.Sylvester Back on topic of weekly updates and things, we had a mantis ticket come in that seems to show migrations not running properly between updates
[12:59 PM PT]  Vincent.Sylvester This has me digging into that again thinking perhaps it might be a good idea to add a check to startup to verify table schema
[12:59 PM PT]  Vincent.Sylvester Didn't help that the wiki is out of date on the table schema, but that is an easy fix
[1:00 PM PT]  Ubit Umarov: the invalid cast issue?
[1:00 PM PT]  Cuga.Rajal def good idea to check table schema on startup
[1:00 PM PT]  Vincent.Sylvester Yeah that
[1:01 PM PT]  Vincent.Sylvester Turns out the double to float part of that table migration never ran or somehow got missed
[1:01 PM PT]  Vincent.Sylvester Not sure how, but I been looking at migrations being a sort of fire and forget affair in the first place
[1:01 PM PT]  Vincent.Sylvester If you have a broken table when it tries to run it'll make a bigger mess
[1:02 PM PT]  Vincent.Sylvester I did look into ways in sql to "if field not exist add it" instead of just blindly adding it, but that's only really a bandaid me thinks
[1:02 PM PT]  Vincent.Sylvester Telling the user "hey the table doesn't match what I need" might prevent this
[1:03 PM PT]  Vincent.Sylvester Should not be too hard to implement either, run some sql to get table schema and compare it, error if mismatch
[1:04 PM PT]  Ubit Umarov: no idea.. those migrtations did work fine
[1:04 PM PT]  Vincent.Sylvester They usually do as well I agree
[1:04 PM PT]  Gavin.Hird did you test in MariaDB or "real" MySQL Vincent?
[1:05 PM PT]  Vincent.Sylvester I did not test this. It was what I found through the ticket and comparing what the migrations do that I found the field type mismatch
[1:05 PM PT]  Gavin.Hird ah, ok
[1:06 PM PT]  Vincent.Sylvester I been looking at migrations in the past thinking there has to be a better way to migrate things or at least have it check before adding fields to tables
[1:06 PM PT]  Gavin.Hird Has anyone tried MySQL 8.1 yet?
[1:06 PM PT]  Lyr Lobo: : There is a behind the scenes look at our events
[1:06 PM PT]  Lyr Lobo: Not the program, my planning view
[1:06 PM PT]  Vincent.Sylvester When I moved to dotnet for some regions I did have one fail because I had altered the table for another test and as a result the migrations failed and then all hell broke loose
[1:07 PM PT]  Vincent.Sylvester I did find some sql for "if not exist create" rather than just create
[1:07 PM PT]  Andrew Hellershanks: (I'm back)
[1:08 PM PT]  Cuga.Rajal would a schema check at startup simplify the process?
[1:09 PM PT]  Vincent.Sylvester What it would do is if there is no table at all just create it from a sql file directly, no need to do the whole migration thing. If the table schema doesn't match just stop startup and tell the user something is wrong
[1:09 PM PT]  Andrew Hellershanks: Vincent, there is also an INSERT IGNORE option for loading tables with data.
[1:10 PM PT]  Andrew Hellershanks: Vincent, that could be useful. One would certainly hope the database schema matches what is expected for the migration levels of the various tables.
[1:11 PM PT]  Cuga.Rajal it could avoid data corruption and Mantis tickets :)
[1:11 PM PT]  Vincent.Sylvester My ultimate plan for this would be a system that checks the "state" of a table and only run the migrations it needs and otherwise fail if it cannot correct the table to the proper structure.
[1:12 PM PT]  Vincent.Sylvester "Detected table version X, will run migrations..." "Could not update table schema" "Malformed table structure, please correct" type deal
[1:12 PM PT]  Gavin.Hird Isn't that what the migrations table is for. If a migration set is already at the latest level it does nothing?
[1:12 PM PT]  Vincent.Sylvester Yes
[1:12 PM PT]  Vincent.Sylvester Problem is it can blindly run even if the sql fails to execute and set migrations as done
[1:12 PM PT]  Orbert.Tatham I assume there is no field in a table somewhere that has the version of the database. Perhaps that would make a useful addition also?
[1:13 PM PT]  Vincent.Sylvester That's in the migrations table
[1:13 PM PT]  Orbert.Tatham So is it not showing the correct version that is causing this problem?
[1:14 PM PT]  Gavin.Hird the migration and setting the level per migration should be in a transaction then so if the transaction fails, the migration level is not incremented?
[1:14 PM PT]  Andrew Hellershanks: If it updates the migrations table even if the SQL statements doing the migrations fail that is a problem that needs to be fixed.
[1:15 PM PT]  Gavin.Hird but maybe setting the migration level is done in code and not SQL?
[1:15 PM PT]  Orbert.Tatham The migration also needs to be atomic, with a rollback if it fails part way through
[1:15 PM PT]  Cuga.Rajal do you run into the issue of different DBs (postgres, Maria, etc) requiring diff migration SQL?
[1:15 PM PT]  Orbert.Tatham All in one TRANSACTION
[1:15 PM PT]  Gavin.Hird exactly Orbert
[1:15 PM PT]  Andrew Hellershanks: That is how I would expect it to work, Orbert.
[1:16 PM PT]  Vincent.Sylvester If the migration level in migrations table is lower it runs the migrations to bring it up to date, that is the plan at least
[1:16 PM PT]  Vincent.Sylvester However if those migrations fail it just continues bootup regardless
[1:16 PM PT]  Vincent.Sylvester It doesn't check at the end if the table is now in the proper format
[1:16 PM PT]  Cuga.Rajal seems stopping startup on fail would be better
[1:16 PM PT]  Ubit Umarov: well he has several fields as DOUBLE  when there was a migration changing those to flOAT
[1:17 PM PT]  Gavin.Hird but each tmigration should be in a transaction where the last commit is updating the migration table. If anything in the transaction fails, it should be rolled back and the migration stopped
[1:17 PM PT]  Ubit Umarov: possible he had a crash during that migration, and it is flaged as done, and wasn't ?
[1:17 PM PT]  Orbert.Tatham If we fail to start up, we need to supply the users with some sort of tool to fix the problem, too
[1:17 PM PT]  Andrew Hellershanks: Cuga, there are different migration files for different database systems.
[1:17 PM PT]  Cuga.Rajal ah ok
[1:18 PM PT]  Vincent.Sylvester I wanted to include a message that states where the schema diverges and direct them to examine their database and manually make the adjustments according to a structure sql file
[1:18 PM PT]  Vincent.Sylvester Or well, output the proper schema vs what they have something like that
[1:19 PM PT]  Andrew Hellershanks: Orbert, Another thing I've been wanting to implement is to fail startup if it can't connect to the database when it tries to use the connection strings.
[1:19 PM PT]  Orbert.Tatham Yeah, that is what I was thinking
[1:19 PM PT]  Cuga.Rajal that sounds like a very good plan
[1:19 PM PT]  Vincent.Sylvester You know like the "you have an error in your sql syntax" error you get with queries heh
[1:19 PM PT]  Gavin.Hird It will fail at that Andrew with console spew
[1:19 PM PT]  Andrew Hellershanks: You would need to do an ALTER TABLE, or equivalent, as the table may already have data.
[1:19 PM PT]  Jamie.Jordan have a great week yall
[1:20 PM PT]  Lyr Lobo: bye Jamie
[1:20 PM PT]  Lyr Lobo: bye everyone *hugs*
[1:20 PM PT]  Gavin.Hird Bye Jamie
[1:20 PM PT]  Andrew Hellershanks: Bye, Jamie. Thanks for dropping by.
[1:20 PM PT]  Cuga.Rajal I have a couple of short tidbits
[1:20 PM PT]  Vincent.Sylvester I'll get started on the table check part then, work backwards through migrations from there
[1:20 PM PT]  Ubit Umarov: well and also has the old default DOUBLE NULL
[1:21 PM PT]  Ubit Umarov: that code never actually supported
[1:21 PM PT]  Andrew Hellershanks: Gavin, yes, there is console spew in the form of lots of red text but usually nothing that directly ties the red text to an error in the connection strings. It should immediately report a failure when it tries to establish a database connection.
[1:21 PM PT]  Ubit Umarov: current is FLOAT Default 0.0
[1:21 PM PT]  Gavin.Hird it is in there somewhere Andrew but you almost lose your will to live before you get to that part
[1:22 PM PT]  Lyr Lobo: hehe
[1:22 PM PT]  Cuga.Rajal brb
[1:22 PM PT]  Andrew Hellershanks: :)
[1:23 PM PT]  Andrew Hellershanks: If it is buried too deep the problem may be getting lost in propagating the error up the levels to where it can be reported in a sane fashion.
[1:23 PM PT]  Cuga.Rajal back
[1:23 PM PT]  Andrew Hellershanks: It wouldn't be the first "impossible" task I've tackled.
[1:24 PM PT]  Gavin.Hird :-)
[1:24 PM PT]  Gavin.Hird I need to hop
[1:24 PM PT]  Cuga.Rajal tc Gavin
[1:24 PM PT]  Orbert.Tatham Peace, Gavin
[1:24 PM PT]  Gavin.Hird have a great week all
[1:24 PM PT]  Andrew Hellershanks: ok, Gavin. Have a good week. See you next time.
[1:24 PM PT]  Andrew Hellershanks: It is almost half past. About time to wrap up todays meeting.
[1:25 PM PT]  Lyr Lobo: thank you! *waves*
[1:25 PM PT]  Andrew Hellershanks: Thank you all for coming. See you again next week.
Personal tools
About This Wiki