Skip to content

FBCS v1.0.1

Latest

Choose a tag to compare

@ntnglz ntnglz released this 12 Nov 18:39

Corrections to the ECMA-116 claims. The interpreter is unchanged: FullBASIC v4 1.0.

The eight ECMA-116 claims that reported a "known deviation" were wrong. Every example concerned is accepted in the form the standard intends:

  • Examples that are bodies of statements. NUMERIC*5.2 A, B, ... (15.1.3) is the numeric-type of a DECLARE statement. WINDOW ..., VIEWPORT ..., CLIP ..., LINE STYLE ... and TEXT COLOR ... (13.1.3, 13.2.3) are set-objects of a SET statement.
  • Transcription losses. READ X, , Z (10.1.3) is READ X, Y, Z. In RECSIZE RS$ NUMCHARS (11.1.3), NUMCHARS is the numeric variable of RECSIZE.
  • Fixed-point sizes on parameters and functions (15.2.3). They are accepted when the program-unit specifies ARITHMETIC FIXED, as 15.2.2 requires. The old tests did not specify it.
  • SET #N and MAT REWRITE. SET #N with a channel expression works. MAT REWRITE of several arrays works when each array has its own template-element (ECMA-0249).

Changes to the suite:

  • The claims are rewritten.
  • ECMA-0071 and ECMA-0102 to ECMA-0104 now test the full forms.
  • ECMA-0072 uses SET #N.
  • ECMA-0249 is new.

Results: 2085 pass, 0 fail, 0 unsupported. The only remaining difference from ECMA-116 is its erratum in 11.3.2 (claim A-ECMA-errata-01).