Gantt scheduling-correctness benchmark

There's one Vite app per React Gantt. Each has the same 14 scheduling scenarios, and each scenario renders the same data in both products.

Structure

gantt-comparison/
├── bryntum/ — Bryntum React Gantt app
└── svar/    — SVAR Pro React Gantt app

The scenario data lives in shared/scenarios.js. Each app adapts that same data to its own Gantt API in src/scenarios/index.js.

Verifying the scenario code

Run the shared scenario checks from this folder:

npm test

The test verifies that:

• all 14 article scenarios exist in the expected order
• every scenario renders the project with child tasks
• dependency links point at real tasks and use valid link types
• the Bryntum adapter preserves the same rows and has date-pinning enabled
• the SVAR adapter preserves the same rows and has forward auto-scheduling enabled

The 14 scenarios

#	Scenario	Capability tested
01	SS dependency	Start-to-Start link
02	FF dependency	Finish-to-Finish link
03	SF dependency	Start-to-Finish link
04	FS with lag	Lag semantics (minimum gap vs exact gap)
05	must-start-on	Inflexible constraint mode
06	start-no-earlier-than	Semi-flexible constraint mode
07	Weekend calendar	Project-level working time
08	Resource calendar	Per-resource working time
09	Backward scheduling	ALAP / direction = Backward
10	Critical path	CPM correctness
11	Conflict handling	Circular / invalid link handling
12	Lunch break	Intra-day calendar intervals
13	Effort-driven scheduling	FixedEffort mode, effort field
14	Inactive tasks	Excluding a task from scheduling but keeping it in plan

Running both apps

Run the following commands in your terminal:

# Terminal 1
cd bryntum
npm install
npm run dev    # serves on http://localhost:5173

# Terminal 2
cd svar
npm install
npm run dev    # serves on http://localhost:5174

There's a dropdown menu in each app's header that let's you select the scheduling scenario.

Trial package note

Trial packages are used for both React Gantt libraries.