(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
[11:01 AM] Andrew Hellershanks: Hello, everyone.
[11:02 AM] Ubit Umarov: hi
[11:03 AM] Andrew Hellershanks: Some regulars aren't here yet. I'll wait a couple minutes before get started.
[11:03 AM] Joe Magarac: OK
[11:06 AM] Vincent.Sylvester @hg.zetaworlds.com: Isn't today a holiday as well?
[11:06 AM] Andrew Hellershanks: We might as well start although we are few in numbers this week.
[11:06 AM] Andrew Hellershanks: Is it?
[11:06 AM] Vincent.Sylvester @hg.zetaworlds.com: No clue, I don't live in the outside world, it's scary and cold
[11:07 AM] Andrew Hellershanks: It is a federal holiday where I am so goverment and the post office get the day off. The rest of us don't.
[11:07 AM] Andrew Hellershanks: Hello, Jagga.
[11:10 AM] Andrew Hellershanks: Not much to report on the code changes side but there were a few commits. The first one of this past week was regarding mantis #9219.
[11:10 AM] Vincent.Sylvester @hg.zetaworlds.com: sqlite continues to be annoying to deal with
[11:10 AM] Andrew Hellershanks: What are the main issues with it?
[11:10 AM] Ubit Umarov: wel actually that is a big change.. may break things
[11:10 AM] Andrew Hellershanks: Difficulty in handling database migrations?
[11:11 AM] Andrew Hellershanks: Hence the part of the commit message saying itneeds testing.
[11:11 AM] Ubit Umarov: replacing code fom 2007 by code of 2025..
[11:11 AM] Ubit Umarov: or mb 2015..
[11:12 AM] Ubit Umarov: well a bit old
[11:12 AM] Ubit Umarov: :)
[11:12 AM] Ubit Umarov: i did made a standalone, did load a oar, made avatar inventory changes, and saw no issues
[11:12 AM] Vincent.Sylvester @hg.zetaworlds.com: Oh right that's likely going to merge conflict with the new migration system, damn, well that'll be fun to fix
[11:13 AM] Andrew Hellershanks: :)
[11:13 AM] Ubit Umarov: ( ofc no one tells that most those code changes they made on the libs where more useless than cosmetics )
[11:13 AM] Andrew Hellershanks: oh, yeah. Merge conflicts are always fun... not. :P
[11:14 AM] Ubit Umarov: but tjhe older code we did use had hidden dependencies on .net 3.5
[11:14 AM] Ubit Umarov: that ms is removing
[11:14 AM] Ubit Umarov: as reported on that mantis
[11:15 AM] Ubit Umarov: sqlite is a strange mess
[11:15 AM] Vincent.Sylvester @hg.zetaworlds.com: On the one hand it's good to upgrade things once in a while to stay compatible, on the other it's hidden changes possibly breaking the connectors so needs testing
[11:16 AM] Ubit Umarov: but wel l what isnt this days
[11:16 AM] Vincent.Sylvester @hg.zetaworlds.com: I did look through their changelog a while back to see if there was anything breaking, thankfully apart from the case sensitivity there was not much that impacts code
[11:19 AM] Vincent.Sylvester @hg.zetaworlds.com: sqlite is something I question on a more general basis though. It is easiest to setup for users, but more difficult to maintain and ultimately not something for production use. So the question is does it make sense to even have it as option or should it, like so much other software, have a database install as basic requirement
[11:20 AM] Vincent.Sylvester @hg.zetaworlds.com: A user unable to figure out how to install a database likely won't be able to properly administrate
[11:21 AM] Ubit Umarov: sqlite should be fine and a nice option for standalones
[11:21 AM] Vincent.Sylvester @hg.zetaworlds.com: In the end it's not just the person using it, but also everyone else that may join a grid. I feel some responsibility to make sure those people are impacted negatively by bad administration
[11:21 AM] Ubit Umarov: the way .net world messes it, is the problem
[11:21 AM] Vincent.Sylvester @hg.zetaworlds.com: I agree, for standalones in isolation it is an okay option, but some do use it in production
[11:22 AM] Andrew Hellershanks: Vincent, I think you missed a "not" in that statement.
[11:22 AM] Ubit Umarov: Opensim is not just for grid owners on linux
[11:22 AM] Ubit Umarov: standalones are a big part..
[11:22 AM] Ubit Umarov: and many of those have no use for mysql or post...
[11:23 AM] Vincent.Sylvester @hg.zetaworlds.com: It is a question of how much as developers we feel responsible for the users of users
[11:23 AM] Andrew Hellershanks: /me nods
[11:24 AM] Vincent.Sylvester @hg.zetaworlds.com: I just feel a sense of dread when I think about that, cause it all too often is spun as negative reflection on code quality or design when a lot of it comes down to poor administration
[11:24 AM] Vincent.Sylvester @hg.zetaworlds.com: Blaming Ubit for code he didn't even write
[11:26 AM] Vincent.Sylvester @hg.zetaworlds.com: Perhaps I'm jaded cause I see so much negative impact of all that out in the wild along with the folks spewing nonsense as facts
[11:27 AM] Vincent.Sylvester @hg.zetaworlds.com: At least when I get around to finishing the schema files for it the new migrations can be merged in and it'll at least stop folks from making bad edits to their databases
[11:27 AM] Vincent.Sylvester @hg.zetaworlds.com: Fact there is no enforcement or warning of that is scary
[11:28 AM] Vincent.Sylvester @hg.zetaworlds.com: Best case service just fails to load, worst case data gets corrupted
[11:28 AM] Vincent.Sylvester @hg.zetaworlds.com: Bad charsets can cause prims not to store correctly for example
[11:29 AM] Vincent.Sylvester @hg.zetaworlds.com: Speaking of which, that's still in works as well, so hopefully some time this year those fixes go in... once Ubit has stopped cringing at my code :)
[11:30 AM] Vincent.Sylvester @hg.zetaworlds.com: Joe, you said you were looking to open source Sharpview right? Is that still the plan?
[11:32 AM] Joe Magarac: Yes, I still need to do that. I'm currently working on the back end for region impostors, and all that's open source.
[11:32 AM] Joe Magarac: Anyone want to help?
[11:33 AM] Vincent.Sylvester @hg.zetaworlds.com: Neil has shown interest in helping with it. He is looking to make it easier for new users to get inworld and have a more streamlined experience
[11:33 AM] Joe Magarac: Does he know Rust?
[11:33 AM] Vincent.Sylvester @hg.zetaworlds.com: Not sure, but given his resume he can likely learn it
[11:34 AM] Joe Magarac: (We should do this out of the meeting)
[11:35 AM] Vincent.Sylvester @hg.zetaworlds.com: His interests are more along the lines of web viewer things, which in my view isn't the best option, but I think something like Sharpview has the best chance of creating a better onboarding experience for users given it can still be changed
[11:35 AM] Vincent.Sylvester @hg.zetaworlds.com: FS and other LL viewers are pretty stuck in ancient ui structures
[11:37 AM] Andrew Hellershanks: There have been attempts to update, or otherwise, change the UI of a viewer but those experiments haven't gone very far.
[11:37 AM] Vincent.Sylvester @hg.zetaworlds.com: Have you seen the code in there? It's brainfcuk and a half
[11:38 AM] Vincent.Sylvester @hg.zetaworlds.com: Just trying to add the interests tab back into profiles I got lost pretty much immediately
[11:38 AM] Andrew Hellershanks: I did see some of it way back when I had my first attempts to compile the viewer code. Definitely not pretty.
[11:41 AM] Vincent.Sylvester @hg.zetaworlds.com: Make a small change, hit compile, everything explodes
[11:42 AM] Andrew Hellershanks: Not sure what else we have to talk about today.
[11:44 AM] Vincent.Sylvester @hg.zetaworlds.com: No news from Canonical on libgdiplus so far either. Still open bug report -.-
[11:45 AM] Vincent.Sylvester @hg.zetaworlds.com: *throws table*
[11:45 AM] Andrew Hellershanks: I still haven't been able to find a way to run a program that uses that library and get useful debugging information when the program crashes.
[11:46 AM] Andrew Hellershanks: Best bet may be to concentrate on getting it to pass its own unit tests.
[11:46 AM] Andrew Hellershanks: and to clean up the compiler warnings. There are some legitimate concerns behind a few of the warnings.
[11:47 AM] Andrew Hellershanks: I've still been too busy on some work stuff to pursue the issues.
[11:47 AM] Vincent.Sylvester @hg.zetaworlds.com: It also depends on other things like pango, where I think the issue actually sits, but compiling that...
[11:47 AM] Vincent.Sylvester @hg.zetaworlds.com: C is just not my thing
[11:49 AM] Andrew Hellershanks: It is mine. I've built GIMP from source along with various of its dependencies in the past (ie. glib, gtk, Cairo, and Pango).
[11:49 AM] Vincent.Sylvester @hg.zetaworlds.com: https://bugs.launchpad.net/ubuntu/+source/libgdiplus/+bug/2069473
[11:49 AM] Andrew Hellershanks: What I have seen of libgdiplus there are serious issues in its code base before needing to look at how it interacts with other libraries.
[11:50 AM] Vincent.Sylvester @hg.zetaworlds.com: I mean it otherwise works, bad as it is
[11:51 AM] Vincent.Sylvester @hg.zetaworlds.com: Think it's something in how it calls up pango or in pango itself that just returns null
[11:51 AM] Andrew Hellershanks: It is prone to crashes.
[11:51 AM] Vincent.Sylvester @hg.zetaworlds.com: I couldn't break it any further, always came back to the pango call failing
[11:54 AM] Vincent.Sylvester @hg.zetaworlds.com: I tried finding the old version of pango used for the focal build, but getting that installed in latest ubuntu probably breaks all sorts of things
[11:54 AM] Andrew Hellershanks: Vincent, which version do you need?
[11:55 AM] Vincent.Sylvester @hg.zetaworlds.com: I'd have to look again, been a few weeks since I tried this. mono still has a version of libgdiplus for focal that works fine in ubuntu, it's the versions shipped with the distro that are broken
[11:55 AM] Vincent.Sylvester @hg.zetaworlds.com: I'm guessing because something in pango changed and no longer works with how it is called
[11:56 AM] Vincent.Sylvester @hg.zetaworlds.com: So I wanted to fetch older pango-dev and compile it so I could then point libgdiplus at that, much like how the focal version was built, which btw passes all its tests
[11:56 AM] Andrew Hellershanks: I have that situation with Banshee. The libraries it uses changed over time and the code needs to adapt to those changes.
[11:57 AM] Andrew Hellershanks: I have the source for 1.44.7 on my machine but that may be newer than you want. I don't think it would be too hard to find the older versions.
[11:58 AM] Vincent.Sylvester @hg.zetaworlds.com: To find no, but to get them onto a ubuntu noble compiled and working is more difficult cause it has so many small things it depends on
[11:58 AM] Vincent.Sylvester @hg.zetaworlds.com: I still hope Canonical gets their head out their ass and fixes it
[11:58 AM] Vincent.Sylvester @hg.zetaworlds.com: winehq version of libgdiplus, they know where the bug is, but have no fix for it either
[11:59 AM] Andrew Hellershanks: The gitlab repo for pango is https://gitlab.gnome.org/GNOME/pango and a collection of past versions can be seen at https://download.gnome.org/sources/pango/. There are versions going all the way back to 0.20
[12:00 PM] Vincent.Sylvester @hg.zetaworlds.com: I think I looked at some 1.4x version of it since there are minimal versions libgdiplus wants for that as well
[12:00 PM] Vincent.Sylvester @hg.zetaworlds.com: Trying the oldest minimal supported version to avoid whatever the problem with the current version is basically
[12:00 PM] Joe Magarac: Gotta go, next meeting
[12:01 PM] Andrew Hellershanks: ok, Joe. See you next week.
[12:01 PM] Vincent.Sylvester @hg.zetaworlds.com: If you could fix just that problem you could dunk on Canonical pretty hard :) hint hint nudge nudge
[12:01 PM] Andrew Hellershanks: :)
[12:02 PM] Andrew Hellershanks: I need it to be stable first
[12:02 PM] Vincent.Sylvester @hg.zetaworlds.com: It's apparently a single call to pango that fails, all other things seem to work fine, it's just that one call to get the compiled font collection that returns no data
[12:02 PM] Vincent.Sylvester @hg.zetaworlds.com: https://gitlab.winehq.org/mono/libgdiplus/-/issues/4
[12:03 PM] Vincent.Sylvester @hg.zetaworlds.com: Granted it only has a small impact in OS as this only breaks system fonts so you are basically limited to the few fonts it has hardcoded, but obviously other applications with custom fonts break entirely
[12:04 PM] Vincent.Sylvester @hg.zetaworlds.com: Keepass especially having a font so you can tell l apart from I in passwords is kinda important
[12:05 PM] Vincent.Sylvester @hg.zetaworlds.com: We'll see if next years release breaks it further at which point may have to bounty that
[12:06 PM] Andrew Hellershanks: I have grabbed copy of libgdi pointed to by winehq and will compare it to the version I have to see which one is the latest.
[12:06 PM] Ubit Umarov: it is linux.. ofc things break
[12:06 PM] Andrew Hellershanks: :)
[12:06 PM] Vincent.Sylvester @hg.zetaworlds.com: You are mistaking it for Windows now xD
[12:07 PM] Ubit Umarov: and you are luck we dont try to display things :)
[12:07 PM] Vincent.Sylvester @hg.zetaworlds.com: Well the dynamic texture font rendering does kinda depend on this, it just means any system installed font fails to show up if they are specific types
[12:08 PM] Vincent.Sylvester @hg.zetaworlds.com: I think if they are just image fonts they work, it's the svg ones that break or something like that
[12:08 PM] Ubit Umarov: yeah but does not try to display it on the region machine
[12:08 PM] Ubit Umarov: bet that would just smoke
[12:08 PM] Andrew Hellershanks: I already had downloaded the library from winehq. I just pulled and most of the changes are related to tests.
[12:08 PM] Ubit Umarov: maybe ms does have a good reason to not try to maku a .net ui apoi for linux
[12:09 PM] Ubit Umarov: only win and mobiles
[12:09 PM] Ubit Umarov: (maui )
[12:09 PM] Vincent.Sylvester @hg.zetaworlds.com: Using ms and reason in the same sentence....
[12:09 PM] Ubit Umarov: wel in .net context
[12:10 PM] Ubit Umarov: ( if you ask them they will tell .net is not ms pff)
[12:10 PM] Vincent.Sylvester @hg.zetaworlds.com: I'd be shocked if half the management there finds their own parking spot at the end of the day
[12:10 PM] Vincent.Sylvester @hg.zetaworlds.com: They probably just wait for everyone else to leave
[12:10 PM] Ubit Umarov: even with the win10 issues, seems linux is losing users
[12:11 PM] Ubit Umarov: and most run linux to run wine
[12:11 PM] Andrew Hellershanks: Wine only gets you so far.
[12:12 PM] Vincent.Sylvester @hg.zetaworlds.com: I'm staying on 10 until 11 works properly, so for the next 4+ years at least
[12:12 PM] Ubit Umarov: yeah reason why many just went back buyin a box to run win11
[12:12 PM] Vincent.Sylvester @hg.zetaworlds.com: It's the productivity software that is most difficult to get running, most games outside of the ones you shouldn't be playing anyways, work fine
[12:12 PM] Ubit Umarov: sad.. but linux is just 2 broken for desktop
[12:13 PM] Vincent.Sylvester @hg.zetaworlds.com: Not more broken than Win11
[12:13 PM] Ubit Umarov: and libgdi is part of that mess
[12:13 PM] Ubit Umarov: nah win11 is a lot better
[12:13 PM] Ubit Umarov: at least stil is
[12:14 PM] Vincent.Sylvester @hg.zetaworlds.com: Having used it and mac os, nah 11 is ass
[12:14 PM] Ubit Umarov: i can plau nms on it :p
[12:14 PM] Ubit Umarov: play also
[12:15 PM] Ubit Umarov: can you test the sqlite changes on yr apple silicon thing?
[12:15 PM] Andrew Hellershanks: Hm... current libgdiplus from winehq compiled with a lot less nasty looking warnings vs when I last compiled it back in October. This version even passed all of its unit tests.
[12:16 PM] Vincent.Sylvester @hg.zetaworlds.com: Yeah I will at some point, best remind me of that again. I have to setup FS compile on it as well at some point
[12:16 PM] Ubit Umarov: fs compile is a bit harder than opensim :)
[12:16 PM] Vincent.Sylvester @hg.zetaworlds.com: heh "bit" heh...
[12:17 PM] Ubit Umarov: oops rl calls
[12:17 PM] Vincent.Sylvester @hg.zetaworlds.com: TJ did give me his compile setup so easier, but arm still be fun
[12:17 PM] Ubit Umarov: any other issue?
[12:17 PM] Ubit Umarov: i do compile fs...
[12:17 PM] Ubit Umarov: it makes my 13900 fail unless i cut threads usage
[12:18 PM] Ubit Umarov: pfff
[12:18 PM] Vincent.Sylvester @hg.zetaworlds.com: :)
[12:18 PM] Ubit Umarov: ok, see you here or there..
[12:18 PM] Andrew Hellershanks: Getting a viewer to compile on my machine was quite the task.
[12:18 PM] Andrew Hellershanks: ok, Ubit.
[12:19 PM] Ubit Umarov: fs instructions do wokr
[12:19 PM] Ubit Umarov: well i never did got the sound lib
[12:19 PM] Andrew Hellershanks: Good time to call this gathering to a close. I hope we will be back to regular strength next week.
[12:19 PM] Ubit Umarov: cya :)
[12:19 PM] Andrew Hellershanks: We are only about one month away from the OSCC now.
[12:19 PM] Andrew Hellershanks: Thank you all for coming. See you again next week.
