[11:07 AM] Andrew.Hellershanks @tanglegrid.net:8002: Hello, everyone.
[11:07 AM] Cuga.Rajal @rajal.org:9000: Hi all!
[11:07 AM] Lyr.Lobo @cc.opensimulator.org:8002: Hello *grins*
[11:07 AM] Andrew.Hellershanks @tanglegrid.net:8002: Good timing, Cuga. :)
[11:08 AM] Cuga.Rajal @rajal.org:9000: There are a couple of Mac issues going on
[11:08 AM] Gavin.Hird @grid.xmir.org:8002: I found why your dylib signature expires
[11:08 AM] Cuga.Rajal @rajal.org:9000: oh do tell
[11:08 AM] Jagga Meridith: logger?
[11:09 AM] Gavin.Hird @grid.xmir.org:8002: if you sign with the --timestamp option, it will check if the signing certificate was valid at the time when the dylib was signed
[11:09 AM] Gavin.Hird @grid.xmir.org:8002: without it will check against the latest cert
[11:09 AM] Cuga.Rajal @rajal.org:9000: ok, great to know
[11:09 AM] Cuga.Rajal @rajal.org:9000: I could re-sign all of them with that option
[11:09 AM] Gavin.Hird @grid.xmir.org:8002: so using --timestamp, the signature will never expire unless the certificate is actually revoked
[11:10 AM] Cuga.Rajal @rajal.org:9000: well thats the other problem
[11:10 AM] Gavin.Hird @grid.xmir.org:8002: yes, you should
[11:10 AM] Cuga.Rajal @rajal.org:9000: the certificates that Apple offers only last for a year
[11:10 AM] Gavin.Hird @grid.xmir.org:8002: yes, but use --timestamp
[11:10 AM] Cuga.Rajal @rajal.org:9000: ok, so when the certificate expires, the dylib will still load bc it was valid when it was signed?
[11:10 AM] Gavin.Hird @grid.xmir.org:8002: this is why my signed dylibs still work despite newer certificates
[11:11 AM] Gavin.Hird @grid.xmir.org:8002: correct
[11:11 AM] Cuga.Rajal @rajal.org:9000: ok *thanks Apple*
[11:11 AM] Vincent Sylvester: lol
[11:11 AM] Cuga.Rajal @rajal.org:9000: let me re-sign all 3 libs with that option
[11:11 AM] Gavin.Hird @grid.xmir.org:8002: good
[11:11 AM] Cuga.Rajal @rajal.org:9000: ther eare a few follow-on questions about that
[11:12 AM] Vincent Sylvester: What's become clear is we lack documentation on building these libs in the first place. We should create a wiki page for how they were created and signed simply for transparency, but also so they can be updated should they need to
[11:12 AM] Cuga.Rajal @rajal.org:9000: the 0.9,3.0 release has that problem. Would it make sense to re-package that with the newer libs?
[11:12 AM] Cuga.Rajal @rajal.org:9000: there are no dependencies with OS code in those unmanaged libs
[11:13 AM] Vincent Sylvester: It likely still fail on the part of it not being compatible somehow, so only thing it would fix is the cert
[11:13 AM] Cuga.Rajal @rajal.org:9000: Vincent, I agree, on the Mac side, the build process for all libs are documented at my github page
[11:13 AM] Ubit.Umarov @hg.osgrid.org:80: aren't all mac users but kent here?
[11:13 AM] Ubit.Umarov @hg.osgrid.org:80: :p
[11:13 AM] Cuga.Rajal @rajal.org:9000: fine with me if that info is moved to os.com
[11:13 AM] Gavin.Hird @grid.xmir.org:8002: :-)
[11:14 AM] Cuga.Rajal @rajal.org:9000: right.. the libsqlite segfaulting is a separate issue
[11:14 AM] Vincent Sylvester: We have to build new libraries for all platforms eventually anyways, so having the process documented would certainly help with testing
[11:14 AM] Ubit.Umarov @hg.osgrid.org:80: os.com?
[11:15 AM] Vincent Sylvester: Haven't gotten around to reading sqlite changelogs, that's first on the list for updating that lib to a newer version, cause running into a break change at runtime is not fun
[11:15 AM] Cuga.Rajal @rajal.org:9000: opensimulator.org I mean
[11:15 AM] Vincent Sylvester: I suspect there not really being anything that will break given the implementation of sqlite is fairly generic, but who knows
[11:15 AM] Vincent Sylvester: Given the issues currently makes sense to make sure it'll actually work
[11:15 AM] Cuga.Rajal @rajal.org:9000: I could start building different versions of libsqlite and just pop them in one at a time to see which breaks
[11:15 AM] Vincent Sylvester: Windows and Linux can probably benefit from newer versions too
[11:16 AM] Cuga.Rajal @rajal.org:9000: agree
[11:16 AM] Vincent Sylvester: The breaking might only happen in certain cases on specific queries so you'd have to test everything that somehow interacts with the database... not sure that's less work than reading their release notes
[11:16 AM] Ubit.Umarov @hg.osgrid.org:80: ( "if it isn't broken, don't fix it" comes to mind
[11:16 AM] Cuga.Rajal @rajal.org:9000: but what concerns me is that the lib that Gavin made was the same version and didn't segfault
[11:17 AM] Cuga.Rajal @rajal.org:9000: Ubit, under that reasoning, we could just use the libsqlite that comes with the system
[11:17 AM] Gavin.Hird @grid.xmir.org:8002: which version of xcode do you use?
[11:17 AM] Cuga.Rajal @rajal.org:9000: ther latest
[11:17 AM] Vincent Sylvester: Part of why it makes sense to document the process, cause I also would like to try and build it on the m4 and see if building it on hardware makes a difference
[11:17 AM] Gavin.Hird @grid.xmir.org:8002: did you set a deployment target?
[11:17 AM] Ubit.Umarov @hg.osgrid.org:80: isn't that one seg faulting for some also ?
[11:18 AM] Cuga.Rajal @rajal.org:9000: gavin yes, and tested, taget is macOS 10.15, and I have a test machine to check with of that OS
[11:18 AM] Cuga.Rajal @rajal.org:9000: actually re-checked this weekend
[11:18 AM] Vincent Sylvester: Far as I hear it has been faulting on pretty much all the arm based chips now including the pi
[11:18 AM] Cuga.Rajal @rajal.org:9000: 10.15 is till supported with brew, dotnet, etc
[11:18 AM] Gavin.Hird @grid.xmir.org:8002: try build it with target 10.14 which I am sure mine are built with
[11:18 AM] Vincent Sylvester: piusnoel reported on irc his attempts to use it failed and that was a pi I think
[11:19 AM] Vincent Sylvester: So definitely it seems to dislike arm
[11:19 AM] Cuga.Rajal @rajal.org:9000: well 10.15 is the earlest that dotnet 8 will install on
[11:19 AM] Cuga.Rajal @rajal.org:9000: I believe
[11:19 AM] Cuga.Rajal @rajal.org:9000: so that seemed like a good choice
[11:19 AM] Gavin.Hird @grid.xmir.org:8002: doesn't matter
[11:19 AM] Gavin.Hird @grid.xmir.org:8002: has to do with library support
[11:19 AM] Cuga.Rajal @rajal.org:9000: ok I can change it to 10.14
[11:19 AM] Gavin.Hird @grid.xmir.org:8002: squlite does not depend on dotnet anyway
[11:20 AM] Cuga.Rajal @rajal.org:9000: you mean rebuils all the libs like that or just libsqlite?
[11:20 AM] Gavin.Hird @grid.xmir.org:8002: and since the code is from 2011?
[11:20 AM] Vincent Sylvester: The ubode and bullet one seems fine apart from that quarantine crap
[11:20 AM] Gavin.Hird @grid.xmir.org:8002: it probably needs support for some old lib
[11:20 AM] Cuga.Rajal @rajal.org:9000: in 2011 10.14 was a few years in the future. I think we were on 10.8 or something
[11:20 AM] Gavin.Hird @grid.xmir.org:8002: just libsqlite
[11:21 AM] Cuga.Rajal @rajal.org:9000: ok
[11:21 AM] Gavin.Hird @grid.xmir.org:8002: yes, irrelevant
[11:21 AM] Cuga.Rajal @rajal.org:9000: so just to recap
[11:21 AM] Gavin.Hird @grid.xmir.org:8002: the issue is when support were removed from macOS
[11:21 AM] Cuga.Rajal @rajal.org:9000: 1) re-sign the 3 unmanaged libs
[11:21 AM] Cuga.Rajal @rajal.org:9000: 2) rebuild libsqlite with 10.14 set to the target
[11:21 AM] Gavin.Hird @grid.xmir.org:8002: yep
[11:21 AM] Cuga.Rajal @rajal.org:9000: but with same old version
[11:21 AM] Gavin.Hird @grid.xmir.org:8002: yep
[11:21 AM] Vincent Sylvester: 3 Document the process on the wiki
[11:22 AM] Cuga.Rajal @rajal.org:9000: Document just sqlite or all the unmanged libs?
[11:22 AM] Cuga.Rajal @rajal.org:9000: for the others I can just copy my docs from my github page over
[11:23 AM] Vincent Sylvester: All of the preferably, but sqlite would be a start
[11:23 AM] Vincent Sylvester: I have to look at libgdi at some point since it seems we may have to start shipping that too
[11:23 AM] Cuga.Rajal @rajal.org:9000: sqlite is really nothing more than editing the make file to include 2 architectures, run configure and make
[11:23 AM] Cuga.Rajal @rajal.org:9000: thats all
[11:23 AM] Gavin.Hird @grid.xmir.org:8002: Remember 10.15 was the first macOS version that only support 64-bit libs, so they removed all kinds of old crap because of that
[11:24 AM] Cuga.Rajal @rajal.org:9000: 32 bit is not supported on current xCode
[11:24 AM] Cuga.Rajal @rajal.org:9000: hasnt been for a while
[11:24 AM] Gavin.Hird @grid.xmir.org:8002: I didn't say to compile for 32--bit
[11:24 AM] Cuga.Rajal @rajal.org:9000: although... macOS earlier than 10.15 did support 64 bit, just wasn;t 64-bit only
[11:24 AM] Gavin.Hird @grid.xmir.org:8002: they removed a lot of old functionality from the system in 10.15
[11:25 AM] Cuga.Rajal @rajal.org:9000: I wil try and see what happens
[11:25 AM] Gavin.Hird @grid.xmir.org:8002: I would try and see what happens
[11:25 AM] Cuga.Rajal @rajal.org:9000: but nobdy riunning opensim will be using anything older than 10.15, dotnet wont install on earlier
[11:26 AM] Cuga.Rajal @rajal.org:9000: ok
[11:26 AM] Gavin.Hird @grid.xmir.org:8002: I get that but the old code will most likely make old system calls that has been removed from 10.15 onwards
[11:26 AM] Cuga.Rajal @rajal.org:9000: possibly. But I think XCode dropped that too
[11:26 AM] Cuga.Rajal @rajal.org:9000: I will try
[11:26 AM] Gavin.Hird @grid.xmir.org:8002: so by building for target 10.14 the necessary code will be included with the lib
[11:27 AM] Ubit.Umarov @hg.osgrid.org:80: well about opensim code changes during last week.. i just did add LSL_Integer osListFindListNext(LSL_List src, LSL_List test, LSL_Integer lstart, LSL_Integer lend, LSL_Integer instance), like ll one but with search restricted to a sublist
[11:27 AM] Cuga.Rajal @rajal.org:9000: I have to say also that I compiled libsqlite and posted it to Mantis without fully testing it. I don;t think it should be in trunk until we test and confirm
[11:27 AM] Cuga.Rajal @rajal.org:9000: can we pull from trunk?
[11:27 AM] Ubit.Umarov @hg.osgrid.org:80: i did not added the ll one because vincent had work on it and may submit it :)
[11:28 AM] Cuga.Rajal @rajal.org:9000: the file can be posted and dropped into the file tree by hand for testing
[11:28 AM] Cuga.Rajal @rajal.org:9000: posted to Mantis
[11:28 AM] Cuga.Rajal @rajal.org:9000: Vincent is there any reason you want to keep that broken libsqlite in trunk?
[11:28 AM] Vincent Sylvester: Should get a load of lsl stuff once I got around to test it heh
[11:28 AM] Gavin.Hird @grid.xmir.org:8002: The current Dayturn viewer for macOS is built with target 10.14, while the version 3 beta is built for 10.15
[11:29 AM] Vincent Sylvester: It's trunk, expect bugs :)
[11:29 AM] Cuga.Rajal @rajal.org:9000: ok thats your decision
[11:29 AM] Ubit.Umarov @hg.osgrid.org:80: well commit to master is just extending the test group
[11:29 AM] Vincent Sylvester: It's being worked on and should resolve at some point soon I should think
[11:29 AM] Cuga.Rajal @rajal.org:9000: well from now on I won;t post to Mantis if it segfualts
[11:29 AM] Ubit.Umarov @hg.osgrid.org:80: can't be worse that it is now :)
[11:29 AM] Vincent Sylvester: exactly
[11:29 AM] Cuga.Rajal @rajal.org:9000: really there is no reason to keep it there
[11:29 AM] Cuga.Rajal @rajal.org:9000: can you back out please
[11:30 AM] Ubit.Umarov @hg.osgrid.org:80: unless there are other seg fault prettier than current ones :P
[11:30 AM] Vincent Sylvester: That might actually give a clue as to what is broken
[11:30 AM] Vincent Sylvester: If there is even issues with pi now
[11:30 AM] Cuga.Rajal @rajal.org:9000: it *couild* be dotnet's handing of the loading. There are no crash logs about that segfault anywhere
[11:31 AM] Cuga.Rajal @rajal.org:9000: looked deep into the system
[11:31 AM] Vincent Sylvester: It's a big overdue anyways to build some newer libraries. For everything really. mysql too Ubit fixed a while ago, but could also update it to newer code most likely
[11:31 AM] Vincent Sylvester: Ubit makes the decisions what goes into trunk. My main thing right now is documenting the process so that it can be transparent what is shipped with the binary
[11:32 AM] Cuga.Rajal @rajal.org:9000: ok
[11:32 AM] Vincent Sylvester: And to make it easier to build and test newer things and have the same underlying conditions
[11:32 AM] Cuga.Rajal @rajal.org:9000: Well, the 0.9.3.0 release is now "broken" for mac users. Do you want that?
[11:32 AM] Cuga.Rajal @rajal.org:9000: When I re-sign the libs will you re-package?
[11:32 AM] Vincent Sylvester: It worked fine on the Intel ones
[11:32 AM] Ubit.Umarov @hg.osgrid.org:80: brone bc signature?
[11:33 AM] Cuga.Rajal @rajal.org:9000: no, the gatekepper issue is also with Intel
[11:33 AM] Cuga.Rajal @rajal.org:9000: trunk is fine, 0.9.3.0 throws gatekeepwr issue
[11:33 AM] Cuga.Rajal @rajal.org:9000: re-packaging with re-signed liubs will fix
[11:34 AM] Ubit.Umarov @hg.osgrid.org:80: or i can release 0.9.3.1 :p
[11:34 AM] Cuga.Rajal @rajal.org:9000: thats probably even better
[11:34 AM] Vincent Sylvester: Have to fix the libs first though before doing that
[11:34 AM] Dahlia.Trimble @grid.wwweb3d.net:8002: hi
[11:34 AM] Cuga.Rajal @rajal.org:9000: the lib re-signing is quick and easy, can do next couple days
[11:35 AM] Cuga.Rajal @rajal.org:9000: the sqlite issue is something else
[11:35 AM] Ubit.Umarov @hg.osgrid.org:80: yeah would be nice to have mac arm sqlite working also
[11:35 AM] Cuga.Rajal @rajal.org:9000: I'm going to try building just for arm and see if that runs
[11:36 AM] Cuga.Rajal @rajal.org:9000: and if the 2011 one still segfaults I will try rebuilding other versions, if Vincent can help determine which
[11:36 AM] Ubit.Umarov @hg.osgrid.org:80: ofc win version is very old... but that is the one for mono sqlite
[11:36 AM] Cuga.Rajal @rajal.org:9000: we kno9w the current libsqlite that comes with the OS is fine
[11:36 AM] Cuga.Rajal @rajal.org:9000: with macOS
[11:36 AM] Cuga.Rajal @rajal.org:9000: so sometime between 2011 and 2025
[11:36 AM] Cuga.Rajal @rajal.org:9000: :)
[11:36 AM] Gavin.Hird @grid.xmir.org:8002: I don't think you can even build with target 10.15 on arm and certainly not 10.14
[11:37 AM] Cuga.Rajal @rajal.org:9000: Yes you can and I do
[11:37 AM] Gavin.Hird @grid.xmir.org:8002: not 10.14 I am pretty sure
[11:37 AM] Vincent Sylvester: Lots of changelogs to read
[11:37 AM] Cuga.Rajal @rajal.org:9000: the unmanaged livs built on my aMac run both arches and run on 10.15
[11:37 AM] Cuga.Rajal @rajal.org:9000: and on 15.4
[11:37 AM] Vincent Sylvester: Have to look at the code first to see what's in there, what custom stuff it does use and then see if any of that has changed
[11:38 AM] Dahlia.Trimble @grid.wwweb3d.net:8002: I wonder if the devs publish binaries? they often do for popular libraries
[11:38 AM] Vincent Sylvester: I don't think there will be much if anything at all, but at this point I'd rather make sure than deal with the database corruption later on
[11:38 AM] Ubit.Umarov @hg.osgrid.org:80: https://sqlite.org/download.html
[11:39 AM] Ubit.Umarov @hg.osgrid.org:80: like that dahlia.Trimble ?
[11:39 AM] Ubit.Umarov @hg.osgrid.org:80: seems no arm there
[11:39 AM] Vincent Sylvester: There might also be unit tests for these libs that we could explore, perhaps they provide some information on the seg fault
[11:40 AM] Ubit.Umarov @hg.osgrid.org:80: yeah dream on :p
[11:40 AM] Ubit.Umarov @hg.osgrid.org:80: seg fault is 1 hard for any tests to see..
[11:40 AM] Ubit.Umarov @hg.osgrid.org:80: 2 hard
[11:41 AM] Vincent Sylvester: seg fault itself is memory. Given these libs are C or C++ code that can happen if some calls have changed. I just find it odd that it compiles but fails at runtime. Modern compilers do have some checks to stop you doing dangerous things
[11:41 AM] Dahlia.Trimble @grid.wwweb3d.net:8002: Ubit, yeah like that. But when I built the tarball it didn't have arm as an arch
[11:41 AM] Dahlia.Trimble @grid.wwweb3d.net:8002: could have just been me missing it
[11:41 AM] Ubit.Umarov @hg.osgrid.org:80: guess not there at all..
[11:42 AM] Ubit.Umarov @hg.osgrid.org:80: ( mb arm on android.. )
[11:43 AM] Vincent Sylvester: It failing on pi arm it should be loading the .so not the .dylib so I wonder why that blows up as well
[11:43 AM] Dahlia.Trimble @grid.wwweb3d.net:8002: it was a generated makefile when I ran the configure script, so macos specific I'd think
[11:43 AM] Vincent Sylvester: It must be that it is so old something just isn't made for the new arm archs? But then why does it compile at all
[11:43 AM] Cuga.Rajal @rajal.org:9000: well I just researched my signing process and I'm already using "--timestamp" in my signing
[11:43 AM] Cuga.Rajal @rajal.org:9000: so nothing to re-sign. Gatekeeper issue is something else
[11:44 AM] Vincent Sylvester: Bit of a weird thing, but generally updating that library to a newer version should help with things either way
[11:44 AM] Vincent Sylvester: What was that gatekeeper issue again?
[11:45 AM] Ubit.Umarov @hg.osgrid.org:80: yeah ive no idea that that gatekeeper issue is also
[11:45 AM] Dahlia.Trimble @grid.wwweb3d.net:8002: seems strange that it would segfault on a signing failure
[11:46 AM] Cuga.Rajal @rajal.org:9000: Gavin and I are discussing it
[11:46 AM] Cuga.Rajal @rajal.org:9000: it looks like Gavin and I have been using exactly the same signing commands
[11:47 AM] Ubit.Umarov @hg.osgrid.org:80: ( jezz 133 source files.. on something as "basic" as sqlite"
[11:47 AM] Dahlia.Trimble @grid.wwweb3d.net:8002: lol
[11:47 AM] Ubit.Umarov @hg.osgrid.org:80: ( guess some of those coders where payed per line of code .. )
[11:47 AM] Vincent Sylvester: Also been adding and changing things over the years given what's on their forum
[11:48 AM] Dahlia.Trimble @grid.wwweb3d.net:8002: maybe they have one of those rules like no more than 1000 lines in a source file
[11:48 AM] Vincent Sylvester: The development history is colorful
[11:51 AM] Dahlia.Trimble @grid.wwweb3d.net:8002: does anyone know if there's a fork of libomv somewhere that''s up to date with logins and teleports and whatever else may have changed in the last 10 years (for client use)
[11:51 AM] Ubit.Umarov @hg.osgrid.org:80: we use our own fork
[11:52 AM] Dahlia.Trimble @grid.wwweb3d.net:8002: ya but do you maintain client side?
[11:52 AM] Ubit.Umarov @hg.osgrid.org:80: but only gets code on parts needed by server
[11:52 AM] Ubit.Umarov @hg.osgrid.org:80: so yeah no care for client side
[11:53 AM] Ubit.Umarov @hg.osgrid.org:80: https://bitbucket.org/opensimulator/workspace/repositories/
[11:54 AM] Ubit.Umarov @hg.osgrid.org:80: evenqueue client code will need a fix for recent sl
[11:54 AM] Ubit.Umarov @hg.osgrid.org:80: it does need to handle event id and retries as sl wants
[11:55 AM] Ubit.Umarov @hg.osgrid.org:80: think ll told that radagast is brokne bc of that
[11:55 AM] Ubit.Umarov @hg.osgrid.org:80: well now all use cinder version of it.. so..
[11:55 AM] Dahlia.Trimble @grid.wwweb3d.net:8002: not worried about sl at the moment, just trying to hypergrid around and I get disconnected and I'm trying tu debug
[11:55 AM] Dahlia.Trimble @grid.wwweb3d.net:8002: but my code is ancient
[11:55 AM] Ubit.Umarov @hg.osgrid.org:80: ofc there is also cnder fork..
[11:56 AM] Ubit.Umarov @hg.osgrid.org:80: libremetaverse
[11:56 AM] Dahlia.Trimble @grid.wwweb3d.net:8002: oh yeah she took over radegast
[11:56 AM] Dahlia.Trimble @grid.wwweb3d.net:8002: ty for reminding me
[11:56 AM] Ubit.Umarov @hg.osgrid.org:80: np
[12:00 PM] Andrew.Hellershanks @tanglegrid.net:8002: Any other items for today?
[12:01 PM] Andrew.Hellershanks @tanglegrid.net:8002: I don't see anyone in the middle of typing anything so I take that as a no.
[12:01 PM] Lyr.Lobo @cc.opensimulator.org:8002: Thanks, everyone! Have a great week!
[12:01 PM] Andrew.Hellershanks @tanglegrid.net:8002: That will do it for another week. Thank you all for coming. See you again next week.
