Repository navigation
Release 0.26.0 - #116
Merged
Merged
Release 0.26.0#116
Conversation
Remove the loops that damaged exported fields to see whether other tests notice, and six controller figures another test already checks. Every direct expected-answer test stays. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Preparation chooses the unit q once, as the largest power of two that meets the input bound and the split-count bound, from counts it already knows; there is no candidate loop and no exact-fraction propagation per MSD probe or failure case, and the misses of later questions are gone. Preparation keeps one no-traffic validation of each capacity and count question. The units report records which bound set q; docs/concepts.md states what is promised. On NetLab's DC-BB scenarios preparation and run are 15 to 30 times faster. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The unit q is chosen once from a static bound and the per-case precision machinery is gone, which makes scenarios with many ECMP combine sources prepare and run 15 to 30 times faster; the changelog is dated today and the shipped skills move to their 0.26.0 version (networmix/skills 473c45b). Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Codex Review SummaryThis comment shows the latest Codex review activity on this pull request.
ℹ️ About Codex in GitHubYour team has set up Codex to review pull requests in this repo. Reviews are triggered when you
Codex reacts with 👀 while any review is running, comments if it has suggestions, and reacts with 👍 once all reviews finish with no findings. |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
NetGraph 0.26.0: preparation chooses the bandwidth unit
qonce, from a static bound, and the per-case precision machinery is gone.qis the largest power of two withq ≤ 2p·v_minandq·Γ ≤ p·c_min(Γcounts links, nodes, equal-split demands placed together and tunnels); no candidate loop, no exact-fraction propagation per MSD probe or failure case, no misses. Preparation keeps one no-traffic validation of each capacity and count question.units_reportrecords which bound setq(limited_by,achieved_precision,input_bound,split_bound,range_limit);docs/concepts.mdstates what is promised and where no bound is promised.combinesources prepare and run 15 to 30 times faster; placement uses itsworkersagain.86814c8(export tests compare with hand-derived values only) fromstreamline.Checked: full gate (5,843 passed, 1 skipped), the design's examples (210/210 against an independent reference), A/B/A timing on 13 workloads against 0.25.0, two independent code reviews.
🤖 Generated with Claude Code