Boundary conditions¶
Commands of the boundary_conditions chapter of the database, 15 entries. The status words are defined once, on the compatibility matrix, which says what each rests on and what the script builder does with it. Read it before this page: broken and removed are refusals, so a row here can be telling you the emitter will not write that command for that version.
AIR_ALTITUDE¶
Phase setup, layout inline. Manual: SRC-003 p.328.
Sets the air properties from a standard-atmosphere altitude. THE 25.000 GRAMMAR TAKES A BARE VALUE AND THAT VALUE IS IN FEET. SRC-749 p.286 states it on the parameter row itself, altitude in feet units, with no units token to pass; SRC-748 p.302 introduces the second argument and the FEET or METERS choice with it. So the same call means different things on either side of that boundary, by a factor of 3.28, and the number alone cannot say which. Read the three broken rows beside that, and read them for what they actually record. All three say the same thing: the command ran and the expected density was not observed. ONE of them, 26.120, has a reading behind it, 1.056 against the 0.736 expected, which is the 5000 FOOT standard state and was seen twice on build 7012026 (RPT-014). On 26.100 and 26.101 the probe recorded only that the expected value was absent. An earlier version of this note said all three recorded 1.056. They do all CARRY that sentence, because it is the probe spec's static effect_note rather than a per-run reading, and the same sentence appears in 26.121's row under outcome verified, which is the proof that it is boilerplate. Whether those builds read every altitude in feet is a question about three solvers, and it now needs the density read back on 26.100 and 26.101 as well as 26.120 (PLN-20260810-2000).
| Argument | Specification |
|---|---|
value |
float |
units |
enum in |
| Version | Status | Evidence |
|---|---|---|
| 25.000 | documented | SRC-749 p.286, this edition's own grammar; see the note below |
| 25.100 | documented | SRC-748 p.302, same grammar as 26.120 |
| 26.000 | documented | SRC-747 p.301, same grammar as 26.120 |
| 26.100 | broken | reports/compat/CMP-26100_2026-08-08_full.yaml; the command ran (script processing continued past it) but its effect was not observed; expected: the settings dump reports the 5000 m [...] |
| 26.101 | broken | reports/compat/CMP-26101_2026-08-08_full.yaml; the command ran (script processing continued past it) but its effect was not observed; expected: the settings dump reports the 5000 m [...] |
| 26.120 | broken | reports/compat/CMP-26120_2026-08-08_full.yaml; the command ran (script processing continued past it) but its effect was not observed; expected: the settings dump reports the 5000 m [...] |
| 26.121 | verified | reports/compat/CMP-26121_2026-08-08_full.yaml |
| 26.122 | verified | reports/compat/CMP-26122_2026-08-11_full.yaml; Read that report with its erratum beside it, CMP-26122_2026-08-11_full_erratum_2026-08-11.md: the detail on this record carries the probe spec's static cross-build sentence, which is false on this build and which the record's own effect: true contradicts |
| 26.123 | verified | reports/compat/CMP-26123_2026-08-17_full-sim.yaml; SRC-751 p.333, unchanged from SRC-750 p.333 |
| 26.124 | verified | reports/compat/CMP-26124_2026-09-14_full-sim.yaml; SRC-752 p.333, the SRC-751 file, carried from 26.123 |
AUTO_DETECT_TRAILING_EDGES¶
Phase setup, layout bare. Manual: SRC-003 p.319.
Finds and marks all trailing edge nodes in the geometry. Trailing edge marking controls where the wake is shed, so it must precede solver initialization.
| Version | Status | Evidence |
|---|---|---|
| 26.000 | documented | SRC-747 p.292, same grammar as 26.120 |
| 26.100 | documented | SRC-741 p.314, same grammar as 26.120 |
| 26.101 | verified | reports/compat/CMP-26101_2026-08-08_full.yaml |
| 26.120 | verified | reports/compat/CMP-26120_2026-08-08_full.yaml |
| 26.121 | verified | reports/compat/CMP-26121_2026-08-08_full.yaml |
| 26.123 | documented | SRC-751 p.323, unchanged from SRC-750 p.323 |
| 26.124 | documented | SRC-752 p.323, the SRC-751 file, carried from 26.123; not run on 26.124 |
AUTO_DETECT_WAKE_TERMINATION_NODES¶
Phase setup, layout bare. Manual: SRC-003 p.321.
Finds and marks all wake termination nodes in the geometry. RAN ON 26.124 (reports/RPT-069_three-routes-to-one-body-and-the-file-routes-root-node_2026-09-24.md, T07), on a twisted blade whose root end is its one wake-termination node. Emitted after AUTO_DETECT_TRAILING_EDGES, the saved simulation carries that node. Emitted right after IMPORT_WAKE_EDGES_FROM_FILE, which marked the same twelve edges, it marks nothing and the saved simulation carries no node. Emitted after INITIALIZE_SOLVER it marks the node, and the loads do not move, since the solver was initialised without it; a second INITIALIZE_SOLVER after it logs that the solution was cleared and the initialisation removed, initialises again, and the loads then equal those of the saved simulation and of the detection route to the printed digit. The log reports the trailing-edge groups only during an initialisation, which is the inferred reason: the detection finds the ends of groups the file route has not created yet. So a workflow importing its edges from a file emits it between two initialisations with the same settings, and a setup command after INITIALIZE_SOLVER passes the phase order only through Script.emit_after_initialization.
| Version | Status | Evidence |
|---|---|---|
| 26.000 | documented | SRC-747 p.294, same grammar as 26.120 |
| 26.100 | documented | SRC-741 p.316, same grammar as 26.120 |
| 26.101 | documented | SRC-725 p.320, same grammar as 26.120 |
| 26.120 | documented | SRC-003 p.321 |
| 26.121 | documented | SRC-740 p.324, same grammar as 26.120 |
| 26.123 | documented | SRC-751 p.325, unchanged from SRC-750 p.325 |
| 26.124 | documented | SRC-752 p.325, the SRC-751 file, carried from 26.123. Ran on 26.124 (reports/RPT-069_three-routes-to-one-body-and-the-file-routes-root-node_2026-09-24.md, T07): after IMPORT_WAKE_EDGES_FROM_FILE it marks nothing until the solver has initialised, and a repeated INITIALIZE_SOLVER clears the first initialisation and initialises again with the node it marked |
DETECT_LEADING_EDGES_WAKES_BY_SURFACE¶
Phase setup, layout keyword_block. Manual: SRC-750 p.323.
Runs wake detection on the LEADING edges of the listed surfaces, which is what a vortex-shedding wake needs and what the trailing-edge detection beside it does not do. First documented by the 26.122 edition. Its page describes the parameter as the count of surfaces checked for TRAILING edge detection, word for word the sentence on the trailing-edge command above it. Read as a copied sentence rather than as a statement about this command, since the function name and the section it sits in both say leading edges; nothing here rests on that sentence.
| Argument | Specification |
|---|---|
surfaces |
int |
surface_indices |
int_list |
| Version | Status | Evidence |
|---|---|---|
| 26.122 | documented | SRC-750 p.323, the first edition to document it |
| 26.123 | documented | SRC-751 p.323, unchanged from SRC-750 p.323 |
| 26.124 | documented | SRC-752 p.323, the SRC-751 file, carried from 26.123; not run on 26.124 |
DETECT_TRAILING_EDGES_BY_SURFACE¶
Phase setup, layout keyword_block. Manual: SRC-003 p.319.
Runs trailing-edge detection on a named set of boundaries rather than the whole geometry, using the sweep and bluntness thresholds. The count sits on the SURFACES keyword line and the indices follow comma separated, the same shape as SURFACE_ROTATE. The manual gives the payload row no parameter name, describing it in prose as the line following SURFACES; the name here is this database's.
| Argument | Specification |
|---|---|
surfaces |
int |
surface_indices |
int_list |
| Version | Status | Evidence |
|---|---|---|
| 25.000 | documented | SRC-749 p.276, same grammar as 26.120 |
| 25.100 | documented | SRC-748 p.293, same grammar as 26.120 |
| 26.000 | documented | SRC-747 p.292, same grammar as 26.120 |
| 26.100 | documented | SRC-741 p.314, same grammar as 26.120 |
| 26.101 | documented | SRC-725 p.318, same grammar as 26.120 |
| 26.120 | documented | SRC-003 p.319 |
| 26.121 | documented | SRC-740 p.322, same grammar as 26.120 |
| 26.123 | documented | SRC-751 p.323, unchanged from SRC-750 p.323 |
| 26.124 | documented | SRC-752 p.323, the SRC-751 file, carried from 26.123; not run on 26.124 |
DETECT_WAKE_TERMINATION_NODES_BY_SURFACE¶
Phase setup, layout inline. Manual: SRC-003 p.321.
Detects wake termination nodes on one boundary, where AUTO_DETECT_WAKE_TERMINATION_NODES does the whole geometry. One boundary per call and no all-form is stated.
| Argument | Specification |
|---|---|
surface_id |
int |
| Version | Status | Evidence |
|---|---|---|
| 25.000 | documented | SRC-749 p.278, same grammar as 26.120 |
| 25.100 | documented | SRC-748 p.295, same grammar as 26.120 |
| 26.000 | documented | SRC-747 p.294, same grammar as 26.120 |
| 26.100 | documented | SRC-741 p.316, same grammar as 26.120 |
| 26.101 | documented | SRC-725 p.320, same grammar as 26.120 |
| 26.120 | documented | SRC-003 p.321 |
| 26.121 | documented | SRC-740 p.324, same grammar as 26.120 |
| 26.123 | documented | SRC-751 p.325, unchanged from SRC-750 p.325 |
| 26.124 | documented | SRC-752 p.325, the SRC-751 file, carried from 26.123. Ran on 26.124 once (reports/RPT-069_three-routes-to-one-body-and-the-file-routes-root-node_2026-09-24.md, T07), right after IMPORT_WAKE_EDGES_FROM_FILE and before the solver was initialised, where it marked nothing, as AUTO_DETECT_WAKE_TERMINATION_NODES marks nothing there |
DISABLE_WAKE_NODES_ON_TRAILING_EDGE¶
Phase setup, layout inline. Manual: SRC-003 p.320.
| Argument | Specification |
|---|---|
te_index |
int |
| Version | Status | Evidence |
|---|---|---|
| 25.100 | documented | SRC-748 p.294, same grammar as 26.120 |
| 26.000 | documented | SRC-747 p.293, same grammar as 26.120 |
| 26.100 | documented | SRC-741 p.315, same grammar as 26.120 |
| 26.101 | documented | SRC-725 p.319, same grammar as 26.120 |
| 26.120 | documented | SRC-003 p.320 |
| 26.121 | verified | reports/compat/CMP-26121_2026-08-08_savedstate2.yaml |
| 26.122 | verified | reports/compat/CMP-26122_2026-08-11_full.yaml |
| 26.123 | verified | reports/compat/CMP-26123_2026-08-17_full-sim.yaml; SRC-751 p.323, unchanged from SRC-750 p.324 |
| 26.124 | verified | reports/compat/CMP-26124_2026-09-14_full-sim.yaml; SRC-752 p.323, the SRC-751 file, carried from 26.123 |
FLUID_PROPERTIES¶
Phase setup, layout keyword_block. Manual: SRC-003 p.328.
Sets the working fluid state as KEY VALUE lines. Sonic velocity is derived from temperature and specific heat ratio; it is no longer an independent input (SRC-003 p.328).
| Argument | Specification |
|---|---|
density |
float [kg/m^3] |
pressure |
float [Pa] |
temperature |
float [K] |
viscosity |
float [Pa s] |
specific_heat_ratio |
float |
| Version | Status | Evidence |
|---|---|---|
| 25.000 | documented | SRC-749 p.286, this edition's own grammar; see the note below |
| 25.100 | documented | SRC-748 p.302, this edition's own grammar; see the note below |
| 26.000 | documented | SRC-747 p.301, this edition's own grammar; see the note below |
| 26.100 | documented | SRC-741 p.324, same grammar as 26.120 |
| 26.101 | verified | reports/compat/CMP-26101_2026-08-08_full.yaml |
| 26.120 | verified | reports/compat/CMP-26120_2026-08-08_full.yaml |
| 26.121 | verified | reports/compat/CMP-26121_2026-08-08_full.yaml |
| 26.122 | verified | reports/compat/CMP-26122_2026-08-11_full.yaml |
| 26.123 | verified | reports/compat/CMP-26123_2026-08-17_full-sim.yaml; SRC-751 p.333, unchanged from SRC-750 p.333 |
| 26.124 | verified | reports/compat/CMP-26124_2026-09-14_full-sim.yaml; SRC-752 p.333, the SRC-751 file, carried from 26.123 |
IMPORT_WAKE_EDGES_FROM_FILE¶
Phase setup, layout inline. Manual: SRC-750 p.324.
Marks wake edges from an imported node list, matching an imported coordinate to a mesh edge whose mid-point is within TOLERANCE of it. First documented by the 26.122 edition. THE ARGUMENT ORDER IS THE SIGNATURE'S AND THE SAMPLE'S, not the parameter table's, which lists TOLERANCE first. The signature and the printed call agree with each other on type then tolerance, and a table row order is not a call order. The four types are the same four SET_TRAILING_EDGE_TYPE takes, so this sets for a whole imported file what that command sets for one edge. IT TAKES NO PATH, IN EITHER EDITION, AND THAT IS THE PAGE AND NOT A HOLE HERE. Both pages that document the command were read for this question, SRC-750 p.324 and SRC-751 p.323, and each prints a signature with two placeholders and a sample call with two values. Neither says where the node list comes from, on its own line, as a third argument, or through a prior command. The one command in this chapter that does take a path is TRAILING_EDGES_IMPORT, which carries it on the line after the command name, and its file format block is the one printed here; whether that file is also this command's input is exactly what neither edition states. Adding a file argument would be this database inventing a grammar, so the silence is recorded instead and the acceptance clause that asked for the argument goes back to the owning seat. THE TOLERANCE UNIT IS DERIVED AND PRINTED NOWHERE. The page describes a distance between a mesh edge mid-point and an imported node coordinate; both are expressed in the simulation's own length unit, so the tolerance can be in no other. Recorded on the precedent of SET_VERTEX_MERGE_TOLERANCE, whose unit is stated the same way and for the same reason. ITS DEFAULT IS THE VENDOR'S OWN NUMBER, 0.0001, which the parameter table states and the printed sample passes. It is recorded because a default without a stated basis is a number somebody later has to reverse engineer, and this one decides whether an edge is marked at all: too tight and edges go unmarked, too loose and the wrong edge is claimed. MEASURED ON 26.124 (RPT-061), where the build answers what the pages leave open, and only there. The two-value line the pages print is a syntax error that stops the script, the same line with the path after it is one too, and a path as the third token is read as the line after the command rather than as a file. The build reads the node list from the path on the NEXT line, and only when the command line carries a third token; the token is required, and METER and MILLIMETER marked the same edges, so it is not the unit of the coordinates. The file is the count, one coordinate triple the solver consumes and does not use, then the edge MID-POINTS in the simulation's length unit, with no unit line and no ids; a word on any line marks nothing, and so do the end vertices of the edges. Given the whole span of a straight wing it marks the edges angle detection marks, and a point farther than the tolerance from every edge mid-point marks nothing and logs nothing. So the 26.124 row states its own grammar, the two values plus the third token and the path on its own line, and 26.122 and 26.123 keep the printed line and are refused by the marking helper as unmeasured. VERIFIED ON 26.124 ONLY. The compat probe of 2026-09-24 (CMP-26124_2026-09-24_wake-edge-import) imported the sixteen trailing-edge mid-points of the qa wing through the 26.124 grammar and the solver logged sixteen imported edges; the saved state after it carries the sixteen and the state before it none. On 26.122 and 26.123 it stays documented: the two probe runs that reached those builds both NAME it and neither promotes it, each recording it unprobed for want of a probe specification, and the marking helper refuses the file route there as unmeasured. AUTO_DETECT_TRAILING_EDGES, the angle criterion it replaces for a geometry whose edges are not creases, carries verified rows on 26.101, 26.120 and 26.121, each with its own committed compat report. The reason for the trade is physical: on the one build where the route is offered it is now verified, and it is written down so a later reader does not take it for verified on the builds before. WHAT IT LEAVES TO THE INITIALISATION, MEASURED ON 26.124 (reports/RPT-069_three-routes-to-one-body-and-the-file-routes-root-node_2026-09-24.md, T07). A twisted blade marked by this command, by AUTO_DETECT_TRAILING_EDGES and from its saved simulation carried the same twelve edges, and only this route solved 1.8 % below the other two in induced drag: its saved simulation carried no wake-termination node, where the other two carry the blade's root end. A wake-termination detection emitted between this command and INITIALIZE_SOLVER marks nothing, automatic or by surface; emitted after INITIALIZE_SOLVER it marks the node, and the solver uses it once INITIALIZE_SOLVER is repeated, which logs that the solution was cleared and the initialisation removed and initialises again. The loads then equal the other two routes' to the printed digit. The log reports the trailing-edge groups only during an initialisation, which is the inferred reason. The node file's row order changed nothing.
| Argument | Specification |
|---|---|
type |
enum in |
tolerance |
float [simulation length units] |
| Version | Status | Evidence |
|---|---|---|
| 26.122 | documented | SRC-750 p.324, the first edition to document it |
| 26.123 | documented | SRC-751 p.323, unchanged from SRC-750 p.324 |
| 26.124 | verified | reports/compat/CMP-26124_2026-09-24_wake-edge-import.yaml; SRC-752 p.323 prints the two-value line of SRC-751; this build refuses that line as a syntax error, and reads the node list from the path on the next line only when the command line carries a third token (RPT-061). After it a wake-termination detection marks nothing until the solver has initialised, and a repeated INITIALIZE_SOLVER clears the first initialisation and initialises again (reports/RPT-069_three-routes-to-one-body-and-the-file-routes-root-node_2026-09-24.md, T07) |
MARK_WAKE_TERMINATION_NODES¶
Phase setup, layout bare. Manual: SRC-003 p.321.
Marks the currently SELECTED mesh nodes as wake termination nodes, which is the manual route where the two detection commands are the automatic ones. It takes nothing, so which nodes are marked comes entirely from the selection state left by SELECT_MESH_NODE. Absent from the February 2026 edition, so no 26.100 row.
| Version | Status | Evidence |
|---|---|---|
| 26.101 | documented | SRC-725 p.320, same grammar as 26.120 |
| 26.120 | documented | SRC-003 p.321 |
| 26.121 | documented | SRC-740 p.324, same grammar as 26.120 |
| 26.123 | documented | SRC-751 p.325, unchanged from SRC-750 p.325 |
| 26.124 | documented | SRC-752 p.325, the SRC-751 file, carried from 26.123; not run on 26.124 |
PHYSICS¶
Phase setup, layout keyword_block. Manual: SRC-748 p.293.
Opens a block that launches the geometry autodetection tools, trailing edges and wake termination nodes, and closes on END. Both are presence keywords: writing the keyword runs that tool, omitting it runs nothing. THE HANDOVER IS EXACT AND IT EXPLAINS A WORKFLOW GAP. This block is on 25.000 and 25.100 and on no later edition, while AUTO_DETECT_TRAILING_EDGES and AUTO_DETECT_WAKE_TERMINATION_NODES are on 26.000 and every edition after it and on neither before. The two 25 builds are therefore not missing the capability, which is what their empty column looked like before this entry: they spell it as one block where the later builds spell it as two commands. A single successor is recorded because the field takes one name, and the wake-node command is the other half.
| Argument | Specification |
|---|---|
auto_trail_edges |
bool |
auto_wake_nodes |
bool |
| Version | Status | Evidence |
|---|---|---|
| 25.000 | documented | SRC-749 p.276, this edition documents it |
| 25.100 | documented | SRC-748 p.293, this edition documents it |
| 26.000 | removed | SRC-747 pp.270-346, the whole scripting reference of the 26.000 edition, does not print this command. That is a fact about the document rather than a measurement of the solver, which is the weakest of the three ways a removal is recorded here; no probe has asked this build whether it still answers the name.; Successor: AUTO_DETECT_TRAILING_EDGES |
| 26.100 | removed | SRC-741 pp.275-369, the whole scripting reference of the 26.100 edition, does not print this command. That is a fact about the document rather than a measurement of the solver, which is the weakest of the three ways a removal is recorded here; no probe has asked this build whether it still answers the name.; Successor: AUTO_DETECT_TRAILING_EDGES |
| 26.101 | removed | SRC-725 pp.280-375, the whole scripting reference of the 26.101 edition, does not print this command. That is a fact about the document rather than a measurement of the solver, which is the weakest of the three ways a removal is recorded here; no probe has asked this build whether it still answers the name.; Successor: AUTO_DETECT_TRAILING_EDGES |
| 26.120 | removed | SRC-003 pp.281-376, the whole scripting reference of the 26.120 edition, does not print this command. That is a fact about the document rather than a measurement of the solver, which is the weakest of the three ways a removal is recorded here; no probe has asked this build whether it still answers the name.; Successor: AUTO_DETECT_TRAILING_EDGES |
| 26.121 | removed | SRC-740 pp.284-379, the whole scripting reference of the 26.121 edition, does not print this command. That is a fact about the document rather than a measurement of the solver, which is the weakest of the three ways a removal is recorded here; no probe has asked this build whether it still answers the name.; Successor: AUTO_DETECT_TRAILING_EDGES |
SET_FREESTREAM¶
Phase setup, layout inline. Manual: SRC-003 p.322.
CONSTANT needs no further input. CUSTOM appends a FILETYPE token (STRUCTURED or UNSTRUCTURED) and takes the profile file path on the next line. ROTATION appends frame index, axis (X, Y or Z) and angular velocity in RPM. Which optional arguments each type takes is conditional; the free_stream helper enforces the valid combinations (SRC-003 p.322).
| Argument | Specification |
|---|---|
type |
enum in |
filetype |
enum in {STRUCTURED, UNSTRUCTURED} (optional) |
frame |
int (optional) |
axis |
enum in {X, Y, Z} (optional) |
angular_velocity |
float [rev/min] (optional) |
filename |
path (optional) |
| Version | Status | Evidence |
|---|---|---|
| 25.000 | documented | SRC-749 p.279, same grammar as 26.120 |
| 25.100 | documented | SRC-748 p.296, same grammar as 26.120 |
| 26.000 | documented | SRC-747 p.295, same grammar as 26.120 |
| 26.100 | documented | SRC-741 p.317, same grammar as 26.120 |
| 26.101 | documented | SRC-725 p.321, same grammar as 26.120 |
| 26.120 | documented | SRC-003 p.322 |
| 26.121 | documented | SRC-740 p.325, same grammar as 26.120 |
| 26.123 | documented | SRC-751 p.326, unchanged from SRC-750 p.326 |
| 26.124 | documented | SRC-752 p.326, the SRC-751 file, carried from 26.123; the ROTATION form ran on 26.124 in reports/RPT-052_the-sense-of-a-rotating-free-stream_2026-09-15.md (TWO ROTATION points against one CONSTANT baseline, all three CONVERGED), which measured its SENSE and is not an acceptance of the other two forms; the CUSTOM STRUCTURED form ran on 26.124 under the licensed probe T14 (RPT-071): a uniform field of the row's own speed loaded as the CONSTANT free stream at 0 deg, a field sheared in z moved the lift, and SOLVER_SET_AOA did not turn the field at 4 deg; the CUSTOM UNSTRUCTURED form ran on 26.124 under the licensed probe T16 (RPT-077) and read the same rows as the STRUCTURED one, a field did not reach beyond its grid (closest to the free stream there), a field carrying the incidence loaded the body as that incidence with CL and CDi printed in the stated angle's axes, and a saved simulation kept its field until SET_FREESTREAM was given again |
SET_TRAILING_EDGE_TYPE¶
Phase setup, layout inline. Manual: SRC-003 pp.319-320.
STANDARD sheds the full net circulation (Kutta factor 1.0); RELAXED sheds half, approximating body carryover at non-lifting trailing edges such as wing-fuselage junctions (SRC-003 p.164). THE RELAXED TRAILING EDGE IS ALSO SPECIFIED OUTSIDE THE SCRIPT, AND THE NEWEST EDITION ADDS A FIELD THERE. Paraphrasing SRC-751 p.85: relaxed-Kutta trailing edges can also be declared as a COMPONENT parameter, written where a component is defined rather than in a script, taking semicolon-separated fields; that edition gives it a fifth field where it had four, an integer direction of the relaxed wake shedding parameter line, 0 for the axial direction and the default, 1 for the azimuth direction. The first four fields are unchanged, a chordwise or radial location and two spanwise or axial bounds, so the four-field form stays valid and the change is additive. SRC-750 p.85 prints the four-field form, so this is a 26.123 addition. The azimuth value is new control over the direction the relaxed wake sheds, which is what makes it a rotor fact and not a compatibility footnote. NO SCRIPTING COMMAND ON ANY REGISTERED BUILD TAKES THAT DIRECTION, and this paragraph exists so a reader does not go looking for a fourth argument here or for a new command elsewhere. This command takes two arguments on every build, an edge index and a type; IMPORT_WAKE_EDGES_FROM_FILE takes the same type vocabulary for a whole imported file; neither carries a direction. The direction is a field of the component specification, and this package writes no component file, so nothing here emits it. A writer for that specification is not implemented (PFS-2026.06). THE CITATION SITS IN THESE FREE-TEXT NOTES ON PURPOSE. Page 85 lies outside the registered scripting-reference range, which is where an evidence citation must point, so a manual_ref or a per-version note carrying it would be wrong twice: wrong about where the fact is, and wrong about what kind of fact it is. No status, no manual_ref and no version row moved for this reading, because nothing in it is a claim about the solver.
| Argument | Specification |
|---|---|
te_index |
int |
type |
enum in |
| Version | Status | Evidence |
|---|---|---|
| 26.000 | documented | SRC-747 p.292, same grammar as 26.120 |
| 26.100 | documented | SRC-741 p.314. THREE types, not four: this edition closes the value set at JET_OUTFLOW and does not carry VORTEX_SHEDDING, which first appears as a type in the May edition. Checked against the wrapped-line failure mode that would fake it, with a positive control on 26.101 and a sweep of every page of this edition finding the token nowhere. |
| 26.101 | documented | SRC-725 p.318, same grammar as 26.120 |
| 26.120 | documented | SRC-003 pp.319-320 |
| 26.121 | verified | reports/compat/CMP-26121_2026-08-08_savedstate2.yaml |
| 26.122 | verified | reports/compat/CMP-26122_2026-08-11_full.yaml |
| 26.123 | verified | reports/compat/CMP-26123_2026-08-17_full-sim.yaml; SRC-751 p.323, unchanged from SRC-750 p.324 |
| 26.124 | verified | reports/compat/CMP-26124_2026-09-14_full-sim.yaml; SRC-752 p.323, the SRC-751 file, carried from 26.123 |
SONIC_VELOCITY¶
Phase setup, layout inline. Manual: SRC-003 p.328.
| Argument | Specification |
|---|---|
value |
float [m/s] |
| Version | Status | Evidence |
|---|---|---|
| 26.101 | removed | Already deprecated in 26.1: the solver warns and ignores the value (SRC-725 p.327). |
| 26.120 | removed | No longer supported; the solver throws a deprecation warning and ignores the value. Sonic velocity now follows from temperature and specific heat ratio (SRC-003 p.328). |
TRAILING_EDGES_IMPORT¶
Phase setup, layout inline. Manual: SRC-003 p.319.
Marks trailing edges at vertices listed in a CSV file, the path on the next line. This is the exact route rather than the detection thresholds: the file names the vertices. The manual describes the file format, a vertex count, a unit token and then one line per vertex with an id and three coordinates. That is the content of a file this package does not write, so it is paraphrased here and nothing validates it. The unit line matters: the coordinates are read in the unit the file declares, not in the simulation's. THE NEWEST EDITION CONTRADICTS ITSELF ABOUT THIS COMMAND AND NO ROW IS WRITTEN FOR IT. SRC-751 p.323 is where the block used to open and the whole block is gone: that page carries the surviving neighbours, trailing-edge auto detection, detection by surface, the leading-edge wake detection and the wake-edge import, and prints this command nowhere. SRC-751 p.390, the Script Index, still lists it under the same chapter heading. So the chapter body and the index disagree, and this is the first recorded instance of the index OVER-reporting; the caution recorded until now ran the other way, that the index under-reports, which is why the surface is read from the body. The rule does not change and the body is still what is read. WHY THE ROW IS ABSENT RATHER THAN REMOVED.
removedwould be the nearest of the four statuses and it says more than the evidence does: no run has asked this build whether it still answers to the name, and an edition dropping a page is a fact about a document. So 26.123 answers ABSENT, which the committed enumeration tests/tier1_offline/goldens/absent_on_26123.txt already records, and the reading lives here. THAT ENCODING IS NOW DECIDED rather than pending (decision of 2026-08-19, PFS-2026.04): no version row, plus this note and the committed enumeration, and NO fifth member of the status set. A fifth member would be more expressive and would touch every consumer of the status, the census tables and the derived support levels included, for one command; and the state it would name, a page the vendor dropped, is a fact about a DOCUMENT rather than about the solver, which the four statuses are all about. The absence carries that distinction correctly and this paragraph carries the reason. 26.124 WAS ASKED, AND ITS ROW IS REMOVED (RPT-061). Its edition is the SRC-751 file, so the document says what it says for 26.123; the run of 2026-09-23 is what differs, and the solver answered the name as an unrecognized command with a file present and absent. That is a fact about the solver, so the row is removed and cites the run. 26.123 stays absent, because no run has asked that build.
| Argument | Specification |
|---|---|
file |
path |
| Version | Status | Evidence |
|---|---|---|
| 25.000 | documented | SRC-749 p.276, same grammar as 26.120 |
| 25.100 | documented | SRC-748 p.293, same grammar as 26.120 |
| 26.000 | documented | SRC-747 p.292, same grammar as 26.120 |
| 26.100 | documented | SRC-741 p.314, same grammar as 26.120 |
| 26.101 | documented | SRC-725 p.318, same grammar as 26.120 |
| 26.120 | documented | SRC-003 p.319 |
| 26.121 | documented | SRC-740 p.322, same grammar as 26.120 |
| 26.124 | removed | reports/RPT-061_where-the-wake-edge-import-reads-its-nodes_2026-09-23.md; Measured 2026-09-23: the 26.124 solver answers the name as an unrecognized command, with its file present and with it absent, and with end vertices or edge mid-points in the file; its edition stops printing the command, as SRC-751 did. Trailing edges are marked from a file on this build by IMPORT_WAKE_EDGES_FROM_FILE, whose file names edge mid-points rather than vertices (RPT-061); Successor: IMPORT_WAKE_EDGES_FROM_FILE |