Skip to content

Keep ME Requester crafting plans off the server thread - #76

Closed
Chronic-Reflexes wants to merge 1 commit into
AlmostReliable:1.21.1from
Chronic-Reflexes:async-requester-plans
Closed

Chronic-Reflexes wants to merge 1 commit into
AlmostReliable:1.21.1from
Chronic-Reflexes:async-requester-plans

Conversation

@Chronic-Reflexes

Copy link
Copy Markdown

Summary

  • Requester crafting plans no longer park the server thread in AE2's simulateFor wait. They keep running on the calculator thread and pause on AE2's normal time budget.
  • Pattern provider updates and the end-of-tick craftable-key refresh wait until those plans are paused, so they do not read or write the pattern map on the same tick.
  • Failed plans wait before they are calculated again, and only a few plans run at once. In-flight plans are cancelled when a request changes or the block is removed.

The same fix has been made for the latest 26.1 branch. It is not included here because that version cannot be tested at the moment.

Test plan

  • Minecraft 1.21.1 with NeoForge and Applied Energistics 2 19.2.19, in singleplayer and on a dedicated server
  • Requesters still plan jobs, submit them, and restock
  • No server hitching during planning, and no crash while pattern providers update during a plan
  • 26.1 build was not run

Made with Cursor

AE2 parks the server tick while a requester plan runs. These plans now keep working on the calculator thread, and pattern-map updates wait until they are paused.

Co-authored-by: Cursor <cursoragent@cursor.com>
@rlnt

rlnt commented Nov 17, 2026

Copy link
Copy Markdown
Member

We do not accept AI slop.

@rlnt rlnt closed this Nov 17, 2026
@Chronic-Reflexes

Copy link
Copy Markdown
Author

ai "slop" or not there is a large performance increase with the changes that have been made, its unfortunate you'd put your personal distaste of ai over the performance of the mod

@rlnt

rlnt commented Nov 17, 2026

Copy link
Copy Markdown
Member

I have nothing against AI.

However, I do have a problem with entirely AI generated code where the author has no idea what they are doing, especially when they expect me to spend my free time reviewing code they haven't invested a single minute in understanding.

You mixin into Applied Energistics internals without questioning it (as agents often do, they like hacky mixins) and then presented me with this collection of hacks. On top of that, you provided no technical explanation of why this would be the best solution, how exactly the code works, or why it would actually resolve the performance issues.

Posting an additional explanation on Discord is not particularly helpful. That information belongs in this post: profiling reports, explanations of your approach, and references to specific changes. Instead, you essentially say, "Here you go, deal with it," shifting all the time and effort onto me. You don't respect my time, so I don't respect your contribution.

In my response to your Discord post, I also made it clear that I would look into the performance issues, but that I would definitely not accept your code. Now that the issue has been brought to my attention, I can work on it, but deciding which code I accept remains my decision.

Since you seem unable to understand this simple concept and have instead resorted to a personal attack, you are hereby blocked from the organization.

@AlmostReliable AlmostReliable locked as resolved and limited conversation to collaborators Nov 17, 2026
Sign up for free to subscribe to this conversation on GitHub. Already have an account? Sign in.

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants