Skip to content

html_feature_list.md

Latest commit

 

History

History
118 lines (83 loc) · 3.32 KB

html_feature_list.md

File metadata and controls

118 lines (83 loc) · 3.32 KB
Raw

Below is a list of features currently implemented and some known limitations of the HTML app (the initial version of the list was compiled by a LLM).

Build tab

Use the Build tab to:

  • add modules manually,
  • add subroutines manually,
  • parse pasted USE statements, including USE ... ONLY: imports,
  • handle multiline USE statements with continuation markers and ignore inline ! comments,
  • import JSON,
  • export JSON,
  • export DOT,
  • clear cached graph data.

Browse tab

Use the Browse tab to:

  • search nodes,
  • select a node,
  • hide or restore nodes,
  • identify modules, procedures, and external references.

Inspect tab

Use the Inspect tab to:

  • inspect incoming and outgoing dependencies,
  • view relationship labels such as USE, ONLY: symbol, and containment links even in Minimal mode,
  • rename nodes,
  • hide or delete nodes,
  • add or remove the selected node from the Quick Jump list,
  • reset a customized Quick Jump list back to the automatic default,
  • add per-node notes,
  • add global notes,
  • enable focus mode.

Sidebar sizing

On desktop, the right sidebar can be resized by dragging its left-edge handle. This is useful when working with long module or subroutine names that would otherwise wrap or truncate in the Browse, Inspect, or Quick Jump sections.

Layouts

FortGraph supports several Cytoscape layouts:

  • Dagre left-to-right
  • Dagre top-to-bottom
  • Force-directed
  • Breadth-first
  • Circle
  • Concentric
  • Grid

Path selection

Path-selection mode lets you select several nodes and highlight edges between them.

Local storage

The viewer stores the following data in browser local storage:

  • graph state,
  • node positions,
  • notes,
  • hidden nodes,
  • theme,
  • display settings.

Clearing browser storage or changing browser profiles will remove that state. Export the graph as JSON when you need a portable copy.

Additive JSON import

Imported JSON is merged with the current graph.

FortGraph:

  • preserves existing nodes and edges,
  • matches nodes by case-insensitive title,
  • adds previously unseen nodes,
  • remaps imported edge endpoints,
  • skips duplicate edges with the same source, target, and label,
  • adjusts conflicting IDs,
  • offsets imported positions to reduce overlap.

Because matching is case-insensitive, solver_mod and SOLVER_MOD are treated as the same node during import.

Circular dependencies

The parser and viewer allow circular dependencies.

For example:

procedure_a → procedure_b
procedure_b → procedure_a

Cycles are preserved in the JSON and displayed in the graph.

Parsing limitations

FortGraph uses a lightweight parser rather than a complete Fortran compiler frontend.

Known limitations include:

  • Preprocessor branches are not evaluated.
  • Generic interfaces may not resolve to a specific implementation.
  • Procedure bindings may not resolve completely.
  • Function-call inference is heuristic.
  • Calls through procedure pointers may not resolve.
  • Type-bound procedure calls may not resolve completely.
  • Renamed imports are simplified to their remote symbol.
  • Include files are not expanded automatically.
  • Conditional compilation may produce inactive dependencies.
  • Procedures with identical names in different scopes may be merged.

For compiler-grade semantic analysis, a full Fortran parser or compiler frontend would be required.