Road to Loadout Items

[erri120]

Loadout Items

• [x] #1830
• [x] #1750
• [ ] Remove almost all references of "mod" from code :)
• [x] Rework installers, should integrate with library items
  • [x] #1785
  • [x] #1799
  • [x] #1801
  • [x] #1802
  • [x] FOMOD Installer (#1823)
  • [x] Advanced Installer (#1823)
• [x] Rework diagnostic emitters (only implementations, nothing else has to change)
  • [ ] #1784
  • [x] #1807
  • [x] #1809
• [x] Rework synchronizer
  • [x] #1805
  • [x] #1806
  • [x] #1826
• [ ] Rework VMs
  • [ ] Rework "Library" page (#1864)
  • [ ] Rework "My Mods" page (#1864)
  • [x] #1880
  • [x] Implement delete command in Loadout Grid (#1854)
  • [x] Replace IModInfoViewModel with ILoadoutItemViewModel for LoadoutItem instead of Mod (#1810)
  • [x] Replace IModFilesViewModel and ModFileTreeViewModel to work with LoadoutFile. (#1810)
  • [x] Implement IFileTreeViewModel for loadout items (#1810)
  • [x] Update text editor to accept LoadoutFileId instead of FileId (#1810)
  • [x] IFileOriginsPage use ILibraryService directly (#1828)
  • [x] IFileOriginEntryViewModel use LibraryItem instead of DownloadAnalysis (#1828)
  • [x] Update LoadoutCardViewModel to not use Loadout.Mods.Count for the mod count (#1830)
  • [ ] Complete rework of #1405, requires downloads and jobs
  • [x] Rework LoadoutGrid (#1810)
  • [x] Remove Category and Version column for now
  • [x] List LoadoutItem instead of Mod
  • [x] Rework columns, mostly just renaming from Mod to LoadoutItem for now
• [ ] Rework all tests
  • [ ] #1455
• [ ] Rework CLI verbs (or remove)
  • [ ] stress-test: uses IHttpDownloader, IFileOriginRegistry, and IArchiveInstaller
  • [ ] download-and-install-mod: uses IHttpDownloader, IFileOriginRegistry, and IArchiveInstaller
  • [ ] analyze-archive: uses IFileOriginRegistry
  • [ ] install-mod: uses IFileOriginRegistry and IArchiveInstaller
  • [x] hash-folder (#1843)

Library

• [x] #1710
• [x] #1725
• [x] #1749
• [x] #1790
• [ ] Replace "vanilla state loadout" and put vanilla state in library
• [x] Rethink how IFileStore is going to be used with the new library items. We also want new, or at least updated, trees.
• [ ] Remove IArchiveInstaller integrate its functionality into ILibraryService
• [x] Remove IFileHashCache (#1843)
• [ ] Replace IFileOriginRegistry with ILibraryService.
  • [x] AddLocalFile requires jobs
  • [x] AddDownload requires working downloads with jobs (#1835)

Downloads

• [x] #1725
• [x] #1749
• [x] Replace AdvancedHttpDownloader and SimpleHttpDownloader with an implementation that users Downloader for simplicity. (#1832)
• [x] Integrate new job system with downloads (#1857)
  • [x] #1857
  • [x] Rework NXM IPC and protocol handlers (#1857)
• [ ] Replace IFileOriginRegistry with ILibraryService
• [ ] Adapt UI to jobs: #1405 (make use of new ILibraryService and jobs)

Jobs

• [x] #1749
• [ ] Add progress to jobs, required for downloads
• [ ] Add child jobs to groups
• [x] Add persistent jobs, required for downloads (#1837)
  • [x] Add restarting of jobs on app restart
  • [x] Add persisting of jobs to the DB
• [x] Add global job list or per-service job list, required for downloads

• [x] #1794

[Sewer56]

Replace AdvancedHttpDownloader and SimpleHttpDownloader with an implementation that users Downloader for simplicity.

I have a question with respect to this. I actually suggested using Downloader library all the way back in I think February-ish 2023, during standup.

I don't quite remember the exact reason why we didn't go with that. Original Ticket:

• https://github.com/Nexus-Mods/NexusMods.App/issues/20

I remember two of the talking points for running a custom solution were:

• Crash Tolerance

  • Downloader AFAIK does not periodically save state, so killing the App process would lose the download.
  • But for regular crashes it may be possible to save/restore with https://github.com/bezzad/Downloader?tab=readme-ov-file#how-to-stop-and-resume-downloads-other-time
  • We could implement this though.

• Alternative Source/Failover

  • The guys wanted to be able to do failover on individual chunks.
  • So say you're a premium user and you have access to 5+ servers. If one server dies, the App would pick up downloading that same chunk from another server, rather than restarting the whole download.

In any case, I don't remember the exact reasons, but maybe @halgari remembers.

[erri120]

Downloader AFAIK does not periodically save state, so killing the App process would lose the download.

State would be handled by the job. Downloads will be persistent jobs, so we wouldn't lose the download on crash.

So say you're a premium user and you have access to 5+ servers. If one server dies, the App would pick up downloading that same chunk from another server, rather than restarting the whole download.

The library supports mirrors, haven't tested how well this works:

https://github.com/bezzad/Downloader/blob/10bff5e1ec0412a7292fdd37e5e557d146d2e21a/src/Downloader/IDownloadService.cs#L77-L84

[Sewer56]

Then at this point I'm not even sure why we've never used it. Unless those are more recent features. I wasn't making the call on that.