politas changed the topic of #ckan to: CKAN - The Comprehensive Kerbal Archive Network - https://git.io/vgh4P | Code of Conduct: https://git.io/vgh4t | Waffle: https://waffle.io/KSP-CKAN/CKAN | Please upgrade to v1.24.0 Bruce - https://git.io/vgh4g | Indexing over 800 modules for 1.3.1! | Be _very_ patient in waiting for responses. :)
Lyneira has quit [Quit: Bye]
politas has joined #ckan
politas has quit [Ping timeout: 198 seconds]
politas has joined #ckan
blowfish has quit [Ping timeout: 198 seconds]
stibbons has joined #ckan
stibbons has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
stibbons has joined #ckan
stibbons has quit [Client Quit]
politas has quit [Ping timeout: 182 seconds]
politas has joined #ckan
m4v has quit [Ping timeout: 186 seconds]
m4v has joined #ckan
taniwha has quit [Ping timeout: 186 seconds]
taniwha has joined #ckan
politas has quit [Ping timeout: 198 seconds]
blowfish has joined #ckan
politas has joined #ckan
zengei__ has quit [Ping timeout: 198 seconds]
politas has quit [Ping timeout: 383 seconds]
tccki has quit [Quit: tccki]
tccki has joined #ckan
politas has joined #ckan
<ckanbot> [CKAN] phardy opened pull request #2341: Fix building on macOS (master...fix-macos-build) https://git.io/vA9x9
tccki has quit [Quit: tccki]
tccki has joined #ckan
<ckanbot> [CKAN] HebaruSan commented on issue #2341: Still works for me on Ubuntu. :+1: ... https://git.io/vA9pO
<ckanbot> [CKAN] phardy commented on issue #2341: Got a build environment running today. Expect some updates to the wiki shortly too. :)... https://git.io/vA9pc
<ckanbot> [CKAN] HebaruSan commented on issue #2341: Cool! Issues you might be fixing by doing that, for future reference: #760, #1334, #2272. https://git.io/vA9pu
<ckanbot> [CKAN] politas commented on issue #2337: So we are allowing mod authors to destroy the integrity of their release version numbers?How do our users know which revision of a given mod release they have installed? Could we force an extra version level when we do this somehow? https://git.io/vA9hy
<ckanbot> [CKAN] politas commented on issue #2337: So we are allowing mod authors to destroy the integrity of their release version numbers?How do our users know which revision of a given mod release they have installed? Could we force an extra version level when we do this (similar to epochs, but automatic) somehow? https://git.io/vA9hy
<ckanbot> [CKAN] politas pushed 2 new commits to master: https://git.io/vA9jm
<ckanbot> CKAN/master 6b14299 Paul Hebble: Don't assume path is always longer than root
<ckanbot> CKAN/master f45d014 Myk: Merge #2332 Fix ArgumentOutOfRangeException when removing files from game root
<ckanbot> [CKAN] politas closed pull request #2332: Fix ArgumentOutOfRangeException when removing files from game root (master...fix/game-root-remove) https://git.io/vAPsx
<ckanbot> [CKAN] politas pushed 2 new commits to master: https://git.io/vA9jg
<ckanbot> CKAN/master 99f0417 Paul Hebble: Check conflicts and depends versions in sanity checks
<ckanbot> CKAN/master 5853e3f Myk: Merge #2339 Obey version properties of conflicts and depends relationships
<ckanbot> [CKAN] politas closed issue #2325: Conflict checking ignores "version" property https://git.io/vAKhZ
Lyneira has joined #ckan
<ckanbot> [CKAN] politas pushed 2 new commits to master: https://git.io/vA9jQ
<ckanbot> CKAN/master 567fd17 Myk: Merge #2329 Improve CKAN.app launch script
<ckanbot> CKAN/master a523648 Paul Hebble: Simplify path handling, add Mono 5 path, alert on errors
Shuudoushi has quit [Ping timeout: 383 seconds]
Shuudoushi has joined #ckan
politas has quit [Ping timeout: 198 seconds]
politas has joined #ckan
blowfish has quit [Quit: Leaving]
Hypergolic_Skunk has joined #ckan
Technicalfool has quit [Read error: Connection reset by peer]
tccki_ has joined #ckan
Lyneira_ has joined #ckan
Lyneira has quit [Ping timeout: 383 seconds]
tccki has quit [Ping timeout: 383 seconds]
tivec has quit [Ping timeout: 383 seconds]
tccki_ is now known as tccki
tivec has joined #ckan
anaximander has joined #ckan
<ckanbot> [CKAN] politas commented on issue #2341: Builds fine under Windows, too. https://git.io/vAHGo
<anaximander> hi all
<politas> o/
<anaximander> quick question from a first-time contributor - is there anything special I need to do to get CKAN building?
<anaximander> VS2017 on Win10
<anaximander> any prereqs, manual steps etc?
<anaximander> or is it just clone and go?
<ckanbot> [CKAN] politas closed pull request #2341: Fix building on macOS (master...fix-macos-build) https://git.io/vA9x9
<politas> Run the ./build.ps1 script from PowerShell rather than using VS build
<anaximander> gotcha
<anaximander> every time, or just first time?
<politas> That'll pull in all the pre-reqs and create a build_ subfolder with all the build artifacts, that you can blow away if you have any issues building later.
<politas> Every time. That way, you use the same build script that GitHub uses when we push code to the master branch
<anaximander> ah, that makes sense
<anaximander> I noticed there's an issue from a few weeks back regarding a TLS error on startup
<anaximander> the issue is resolved but when I run the latest CKAN I still get that
<politas> There may be some way to get VS to kick off the build properly, but I use monodevelop undser Linux primarily (as does HebaruSan and several other CKAN devs over the years)
<politas> You're running 1.24.0?
<anaximander> yeah, I'll have a poke around in VS; if there's a way to make VS build run that script instead I'll put a wiki page in for it
<anaximander> yep, downloaded a fresh copy of 1.24 this morning
<politas> Have you done the "click continue and refresh" step that HebaruSan has been telling everyone to do?
<politas> I have had some serious home network issues over the last few days, so haven't really gotten a handle on the workarounds.
<politas> It was all I could do to tidy up the branch status and push out a release.
<anaximander> yeah, it just logs that the connection to the master repo could not be established, and then hangs, gets greyed out, and Windows asks if I want to end the process
<anaximander> I figured I'd try building it from code in case there's anything new in there that might help, and if not, I could offer to lend a hand fixing it
<politas> Oh, if you could figure that out and write a wiki, I'd appreciate it! I'd love to be able to make use of VS debugging tools.
<politas> There are a few changes that may help with that, but have a read through the last few pages of the CKAN forum thread, because there have been a few people posting issues that have had workarounds posted for.
<anaximander> I'm a C# software engineer in my day job, so I figured I might be able to work something out ;)
<politas> Well, we would certainly welcome your expertise if you feel like helping out!
<anaximander> so yeah, after clicking to continue at the exception dialog and then clicking Refresh, the error it (eventually) gives is a nullref
<politas> My coding history is almost entirely procedural languages
<anaximander> parameter name: collection
<anaximander> which means it's probably iterating or LINQ-ing something that isn't there
<anaximander> likely whatever's supposed to come back from this web call it's failing to make
<anaximander> so that's where I'll start digging
<politas> Ah, crap. Travis, what did you do? (The build errored)
<politas> Have you got any experience working on code to work with high latency connections, by any chance?
<politas> Righto, bed time for me!
<anaximander> high-latency?
<anaximander> a little, actually
<anaximander> my last job, I was working on software that managed content downloads on embedded devices across a network that was known for being slow and prone to disconnection
<anaximander> had to do some clever things with variable wait times, smart retry, circuit-breaker pattern, that sort of thing
<anaximander> there's some handy libs for it on NuGet
zengei has joined #ckan
<ckanbot> [CKAN] LadivItnas opened issue #2342: Could not find a part of the path 'C:\Users\name\Appdata\Local\Temp\CdFileMgr\02a3e06-bc35-40.json'. https://git.io/vAHzh
<ckanbot> [CKAN] HebaruSan commented on issue #2342: Interesting. Does this happen every time, or was it a one-off? Can you please check whether these folders exist?... https://git.io/vAHga
<ckanbot> [CKAN] LadivItnas commented on issue #2342: Yeah all of those folders exist https://git.io/vAHVI
<ckanbot> [CKAN] LadivItnas commented on issue #2342: Yeah all of those folders exist sorry for the late reply https://git.io/vAHVI
<ckanbot> [CKAN] HebaruSan commented on issue #2342: Thanks, what about C:\Users\name\Appdata\Local\Temp\CdFileMgr ? https://git.io/vAHV1
<ckanbot> [CKAN] LadivItnas commented on issue #2342: Oh no it's not there... https://git.io/vAHVA
Hypergolic_Skunk has quit [Quit: Connection closed for inactivity]
politas has quit [Ping timeout: 186 seconds]
<ckanbot> [CKAN] LadivItnas commented on issue #2342: How do I get it back there? https://git.io/vAHwM
politas has joined #ckan
<ckanbot> [CKAN] HebaruSan commented on issue #2342: That's quite strange! The Transactions library attempts to create that folder if it doesn't exist:... https://git.io/vAHwx
<ckanbot> [CKAN] LadivItnas commented on issue #2342: Creating the folder seems to have fixed it (even if it's empty)... https://git.io/vAHro
anaximander has quit [Quit: Leaving]
Wetmelon has joined #ckan
* Wetmelon grumbles
<Wetmelon> CKAN used to rebuild directory paths if they were missing. Has that been removed?
<Wetmelon> Specifically $ksp/Ships/VAB
<ckanbot> [CKAN] HebaruSan commented on issue #2337: I don't think "this is a new version but I used the same version number" is what happened here. The use case is that the original upload is _in error_ (in GPP's case, the KSP-AVC file), and it is replaced to fix that error. The author wishes to retract the original version of the upload. And if they manage it before the netkan bot gets around to their mod, they are in fact able to do so.... https://git.io/
Hypergolic_Skunk has joined #ckan
blowfish has joined #ckan
<ckanbot> [CKAN] HebaruSan commented on issue #2337: I don't think "this is a new version but I used the same version number" is what happened here. The use case is that the original upload is _in error_ (in GPP's case, the KSP-AVC file), and it is replaced to fix that error. The author wishes to retract the original version of the upload. And if they manage it before the netkan bot gets around to their mod, they are in fact able to do so.... https://git.io/
<ckanbot> [CKAN] HebaruSan commented on issue #2337: I don't think "this is a new version but I used the same version number" is what happened here. The use case is that the original upload is _in error_ (in GPP's case, the KSP-AVC file), and it is replaced to fix that error. The author wishes to retract the original version of the upload. And if they manage it before the netkan bot gets around to their mod, they are in fact able to do so.... https://git.io/
<Wetmelon> wtf it works in my dev build lol
<Wetmelon> I literally just `git pull master` `cd CKAN; ./build
techman83 has quit [Remote host closed the connection]
m4v has quit [Ping timeout: 207 seconds]
techman83 has joined #ckan
m4v has joined #ckan
blowfish has quit [Quit: Leaving]
<politas> Hi Wetmelon, that sounds strange
<Wetmelon> Very :)
<Wetmelon> Are there any commits between the release and current master branch?
<Wetmelon> that could have affected that
Lyneira_ has quit [Quit: Bye]
<politas> Wetmelon: I don't believe so.
<politas> Sorry, been looking at https://github.com/KSP-CKAN/CKAN/pull/2337
<Qboid> [#2337] title: Invalidate stale cached files from GitHub in Netkan | ## Background... | https://github.com/KSP-CKAN/CKAN/issues/2337
<Wetmelon> yeah have fun with that :D
<ckanbot> [CKAN] politas commented on issue #2337: Yes, I agree that in this case, there would only be the new file being indexed, and no user would install the original zip file, but I'm thinking about future changes. In the past, we have tended to assume that a given URL is a specific file (and we negotiated with Sarbian in order to specifically achieve that for ModuleManager)... https://git.io/vAHSW
<Wetmelon> I have to admit I don't understand the problem
<politas> The problem is mod authors in-place modifying their release files.
<Wetmelon> And then you rehash, detect it, and move on?
<politas> But then how does the client know it needs to change it?
<politas> It's the same version, so it's not an upgrade
<Wetmelon> Doesn't matter, different hash.
<Wetmelon> I'd be inclined to simply BLOCK the download because it could be a compromised version
<Wetmelon> But other mod managers don't use versioning, they just use the hashes and assume it's legit
<politas> Which is what we do currently.
<Wetmelon> Where are the hashes stored / compared?
<politas> CKAN is more like software repository managers like apt than most mod managers
<politas> We store the hashes in the .ckan files.
<Wetmelon> politas: you need a "git push -force-with-lease" style upload lol
<ckanbot> [CKAN] HebaruSan commented on issue #2337: Users clearing their caches won't affect this; these changes affect netkan.exe only. The cache that we're purging is on the netkan bot server, which is the part of the architecture that's in charge of incorporating new files into the repository.... https://git.io/vAHSF
<Wetmelon> Actually I think that's my recommendation. Warn the author and reject the change
<Wetmelon> So users only see the originally indexed file
<Wetmelon> mm I guess that doesn't work if you're not caching locally though does it
<ckanbot> [CKAN] politas commented on issue #2337: Yes, I realise this PR is only handling the netkan side of things, but you're adding a new field into the .ckan files, which should be in the spec, and while this change is great for the specific case of an invalid download being replaced with a valid download, it's going to cause problems if a mod author changes a valid download to a new valid download. ... https://git.io/vAH93
<ckanbot> [CKAN] HebaruSan commented on issue #2337: I may have misunderstood; you're talking about possible future functionality to prompt users to reinstall the same version? I would picture this working by comparing a new `CkanModule.release_date` property to the already existing `InstalledModule.InstallTime` property, set when you install the module and not dependent on the cache.... https://git.io/vAH9C
<politas> Warning the author is unfortunately not possible
Hypergolic_Skunk has quit [Quit: Connection closed for inactivity]