Recover the record line of a display format without input field under INDARA #604

Manually merged
binjovi-bot merged 3 commits from rpg-9-d1-indds-20261022 into trunk 2026-10-22 16:19:24 +00:00
Owner

What

A display file with INDARA whose format has no input field now recovers.

  • rpg_specs.readRecord: with INDARA the record line of such a format loads no record buffer; the function keys follow the entry label at once. Before, the statement was ? unresolved statement, the input block had no record line (no_record_line), the place of the generated lines was not proved, and the rebuilt display member had no fields. Every statement that names a field of the file failed.
  • rpg_source_files: dcl-f NAME workstn indds(DS) with dcl-ds DS len(99); end-ds;.
  • Probes d4-ia-out (one format) and d4-ia-two (two formats).

Measured

  • Census 645/0/0.
  • Public emission against v0.1.310: worse none; better files-display-write, r7d-indds; newly clean d4-ia-out, d4-ia-two.
  • Twins: 0 regressions. All groups: 0 regressions.
  • Held-out exam, counts only: failure comments 71 to 29, emitted 84 to 120 of 148, display member rebuilt with 24 fields. The round trip still fails because 18 statements are unresolved.

Not done

r7d-indds stays different: the object keeps the structure as 99 bytes of '0', and the subfields of the source appear only as literals. A structure with the single subfield of the source (pos(3)) matches.

## What A display file with INDARA whose format has no input field now recovers. - `rpg_specs.readRecord`: with INDARA the record line of such a format loads no record buffer; the function keys follow the entry label at once. Before, the statement was `? unresolved statement`, the input block had no record line (`no_record_line`), the place of the generated lines was not proved, and the rebuilt display member had no fields. Every statement that names a field of the file failed. - `rpg_source_files`: `dcl-f NAME workstn indds(DS)` with `dcl-ds DS len(99); end-ds;`. - Probes `d4-ia-out` (one format) and `d4-ia-two` (two formats). ## Measured - Census 645/0/0. - Public emission against v0.1.310: worse none; better `files-display-write`, `r7d-indds`; newly clean `d4-ia-out`, `d4-ia-two`. - Twins: 0 regressions. All groups: 0 regressions. - Held-out exam, counts only: failure comments 71 to 29, emitted 84 to 120 of 148, display member rebuilt with 24 fields. The round trip still fails because 18 statements are unresolved. ## Not done `r7d-indds` stays `different`: the object keeps the structure as 99 bytes of `'0'`, and the subfields of the source appear only as literals. A structure with the single subfield of the source (`pos(3)`) matches.
dcl-f NAME workstn indds(DS) with dcl-ds DS len(99). The structure
compiles; the subfields of the source are not recovered yet, so the
r7d-indds twin stays different.
With INDARA the response indicators live in the indicator area, so the
record line of a format without input field loads no record buffer and the
function keys follow the entry at once. The statement was unresolved, the
place of the generated lines was not proved and the rebuilt display member
had no fields. Add the d4-ia-out and d4-ia-two probes.
Update the aggregate counts of the held-out program for the INDARA record line
All checks were successful
binjovi/ci Binjovi completed the frozen plan
6e497158a6
binjovi-bot manually merged commit 6e497158a6 into trunk 2026-10-22 16:19:24 +00:00
Sign in to join this conversation.
No reviewers
No labels
No milestone
No project
No assignees
1 participant
Notifications
Due date
The due date is invalid or out of range. Please use the format "yyyy-mm-dd".

No due date set.

Dependencies

No dependencies set

Reference
sean/sourcerer!604
No description provided.