The raised-error type: a structured error value the language owns. More...

#include "x-type.h"

Macros
#define	X_TYPE_ERR_NAME "ERR"

#define	x_obj_type_iserr(B, X) x_obj_is_type((B), (X), X_TYPE_ERR_NAME)

#define	x_err_code(X) x_firstobj((X))

#define	x_err_subject(X) x_restobj((X))

Functions
x_obj_t *	x_type_err_register (x_obj_t p_base, x_obj_t p_args)

x_obj_t *	x_type_err_struct (x_obj_t p_base, x_obj_t p_obj)

Detailed Description

The raised-error type: a structured error value the language owns.

Author: Jon Ruttan (jonru.nosp@m.ttan.nosp@m.@gmai.nosp@m.l.co.nosp@m.m)

Copyright: 2026 Jon Ruttan @license MIT No Attribution (MIT-0)

Macro Definition Documentation

◆ x_err_code

#define x_err_code ( X ) x_firstobj((X))

An ERR is a PAIR-shaped typed object: (code . obj).

The error's CODE: a static-string atom naming what went wrong.

The string is the C message literal from the raise site ("Unbound SYMBOL", "type: no + for"), repointed in place on each raise – the literals have static storage, so no copy and no allocation is needed to make one reachable from x-lang. The language maps the code to whatever tag vocabulary it likes; the engine does not name tags.

◆ x_err_subject

#define x_err_subject ( X ) x_restobj((X))

The error's SUBJECT: what the raise site was complaining about – the unbound symbol's name, the offending operand – as a static-string atom, repointed in place like the code. Empty string when the raise named no subject.

A STRING and not the object itself, deliberately. Raise sites build their subject on the C STACK (x-type/symbol.c:366 fills a local array and passes its address), and the longjmp out of x_eval_error destroys that frame – so retaining the pointer hands the language a dangling one. What IS stable is the string it points at: an interned symbol's name outlives any frame. The engine has always trusted exactly that much (x_eval_error extracts x_atomstr(p_obj) and keeps nothing else), and this slot is that same trust made reachable instead of flattened into a sentence.

◆ x_obj_type_iserr

#define x_obj_type_iserr	(	B,
		X
	)	x_obj_is_type((B), (X), X_TYPE_ERR_NAME)

Test whether object X is a raised error on base B.

◆ X_TYPE_ERR_NAME

#define X_TYPE_ERR_NAME "ERR"

Type-system symbol name

Function Documentation

◆ x_type_err_register()

x_obj_t * x_type_err_register	(	x_obj_t *	p_base,
		x_obj_t *	p_args
	)

Register or retrieve the ERR type on the base context.

Parameters

p_base	x_obj_t* – Base (execution context)
p_args	x_obj_t* – Unused

Returns: x_obj_t* – Registered type object

◆ x_type_err_struct()

x_obj_t * x_type_err_struct	(	x_obj_t *	p_base,
		x_obj_t *	p_obj
	)

Build the ERR type descriptor.

Name only. No make handler: an ERR is not constructed from x-lang – the base owns the single instance and the raise path fills it. No write or display handler either: those stacks are the language's extension point and boot empty on purpose, so an ERR renders as the bounded #<obj:ERR> form until x-lang's err-io.x pushes the prose.