Repository navigation
🍒 [v1.207] fix(uipath-maestro-bpmn): tidy a process built into the Studio Web template - #5002
Merged
rockymadden merged 1 commit intoNov 19, 2026
Conversation
…mplate (#4895) * fix(uipath-maestro-bpmn): tidy a process built into the Studio Web template Studio Web builds a new process by merging it into the template Process.bpmn. merge keeps the template's shapes and places new nodes around them, and the format in Studio Web's CLI is incremental: on a file whose every element has DI it changes nothing. The skill never named --relayout, so a new process kept the merge placement (the main line drifts down a row at each branch). Run `format --relayout` only when step 1 seeded a template (one start and one end event) or the user asked to replace the process; plain `format` stays the missing-DI repair; never --relayout a user's diagram. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * fix(uipath-maestro-bpmn): the Studio Web template may hold only a start event Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * fix(uipath-maestro-bpmn): split step 5, tie replace to Rule 12, align Rule 5 with --relayout Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * fix(uipath-maestro-bpmn): reset note in the loop, empty target in 5a, draft steps 2-4 Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * fix(uipath-maestro-bpmn): template merge in the remaining format-only-on-missing-DI lines Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * fix(uipath-maestro-bpmn): one draft-skeleton sentence for empty, template and replace Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * fix(uipath-maestro-bpmn): empty draft target skips step 4 before relayout Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * fix(uipath-maestro-bpmn): relayout keeps lanes, template rule stays Studio Web --relayout keeps pools, lanes and colors and only moves placement; say so. Terminal agents never merge into the init stub, so drop the template sentence from canonical Rule 5. Split the draft step-4 sentence so no "skip step 4" reads as skipping validate. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
github-actions
Bot
requested review from
a team,
astridstaicu,
dmetzgar and
tmatup
as code owners
November 19, 2026 15:31
dmetzgar
approved these changes
Nov 19, 2026
rockymadden
deleted the
fix-bpmn-format-relayout-template-CP-to-release-v1-207
branch
November 19, 2026 20:28
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.
Dogfood 2026-11-17, issue #12: the format/tidy step did not run.
Cause
In Studio Web a new process merges into the open template
Process.bpmn.mergekeeps the template's shapes and places the new nodes around them. Thenvalidatepasses, and theformatin Studio Web's CLI (cli-browser 1.207.0-dev.9857, uipcli #5637) is incremental: on a merged file it reportsMode: "Unchanged". The skill never names--relayout, so the process keeps the merge placement: the main line drops a row after each branch and never returns.Fix
draft-skeleton.md): step 5 is split in two.format --relayoutruns only when step 1 seeded a template (a start event and at most one end event: Studio Web's template holds only the start event,init's adds the end), the target was empty, or the user chose replace in Critical Rule 12 step 2.formatstays the missing-DI repair. The commands table says which form does what.formatrejects--relayout, run it with no flag.bpmn.brownfield.format:--relayouton a merged user diagram.--relayout, since a newer plainformatkeeps the layout.bpmn.brownfield.formatonly: a merge into a template, or into a process the user chose to replace, builds a new process and takesformat --relayout.--relayoutkeeps pools, lanes and colors and moves placement.initstub, so its path is unchanged.This matches the #13 design: a brownfield edit runs
merge, and nothing relays out an existing diagram.Evidence (Studio Web alpha, Haiku 5.5, new BPMN project per run)
All four runs used the same prompt: "We need a process for customer refund requests. Look up the order first. If the refund is over 200 a supervisor has to approve it, otherwise refund it automatically. Either way email the customer at the end." When the systems questions came up they were dismissed, with the reply "We haven't picked the systems yet, just put placeholder steps in for now and build it." Each run's path is read from the chat and the downloaded file.
4300148e-0959-447f-eda8-08df44e5aeb9ebab1105-91b2-41f5-da45-08df45053912BPMNLabel)format(Generated)a10de1bb-73c7-4a20-f72f-08df455df8a5format Process.bpmn --relayout→Mode: "Relayout", 12 shapes and 13 edges (chat)9d7b95fa-d3a9-4c45-55ea-08df456fbb18SKILL.mdhashfb5e629da78bmatched the build. Fix:SKILL.md66bad0478789andreferences/draft-skeleton.md0a69fc78c103matchbuild/skills/studioweb.getBundledPluginReleases()returned1.0.0-64b639a2e656, equal toshasum -a 256 dist/index.json(64b639a2…ca28).up.sh --no-chrome, and each run used a new chat.validaterejects; that question was dismissed and the canvas changes were accepted. Stock 2 was the earlier, invalid "after" run: its iframe still served the stock release.daa3a428-9628-4640-5f44-08df458daddf, in-iframe release1.0.0-b8eb8b81f53f=shasum -a 256 dist/index.json.compilethenmerge … --original Process.bpmn --baseline …(Placed 8) into the open template, thenformat Process.bpmn --relayout→Mode: "Relayout".bpmn.brownfield.formattext as main, so the dogfood report reproduces on main.@uipath/cli1.207.0-preview.314, merge into theinittemplate: plainformat→Unchanged;format --relayout→Relayout,validateValid.init→ decompile stub → compile straight over →formatGenerated→ validate Valid. Unaffected.Not in scope
validaterejects (Fix B). The untouched template fails the same way.🤖 Generated with Claude Code
Automatically cherry-picked from #4895 to
release/v1.207by @rockymadden.