PeopleSoft Studio

A VS Code extension for developing PeopleSoft definitions — PeopleCode, records, pages, components, App Engine programs and SQL — with the aim of replacing Application Designer.

Status

Early. The foundation is in place and builds: connection management, a provider abstraction with two backends, definitions surfaced as editable virtual files, a project and definition browser, PeopleCode syntax highlighting, and a read-only record editor. Most write paths and all the visual designers are still ahead — see docs/ROADMAP.md.

Two ways to reach your definitions

Direct database connects to the PeopleTools tables on Oracle and reads definitions live. It sees the whole environment and every project in it.

Project export file reads an Application Designer XML export from disk. It sees only what is in that project, but it needs no database credentials — and it is currently the accurate source for PeopleCode.

Both are the same to the rest of the extension; views and editors are written against one DefinitionProvider interface and never learn which is behind them. Capability differences (writability, global search, DDL) are declared by each provider so the UI can adapt rather than fail.

The PeopleCode problem

Oracle does not document how PeopleCode is stored. It is not source text: it is a tokenized byte stream in PSPCMPROG.PROGTXT, split across rows by PROGSEQ, with every identifier replaced by an index into a per-program PSPCMNAME table.

What is implemented and verified: chunk reassembly, gap detection, and name table resolution. What is not: the opcode table itself, which covers only the constructs confirmed so far. Everything else decodes to an explicit unknown token and is reported at the top of the rendered source.

This is deliberate. A decoder that guesses at an unrecognised opcode produces source that looks right and means something else. So:

• Unmapped opcodes are surfaced, never smoothed over.
• Saving PeopleCode to the database is refused, not attempted, and such documents open read-only so you find out before you type rather than after.
• peoplesoft.peoplecode.decoder: "raw" dumps the token stream and name table, which is how the opcode table gets extended from real programs.

Until the table is calibrated against a real database, read PeopleCode from a project export, where App Designer has already written plain source.

Coexisting with other PeopleSoft extensions

jatz.peoplesoft-tools contributes a language also called peoplecode, on scope source.peoplecode, claiming .pcode and .ppl. Two grammars on one scope means load order decides which wins.

So this extension uses psft-peoplecode on source.psft.peoplecode, and its virtual documents end in .peoplecode rather than .pcode. Both extensions can be installed together. richardwood.peoplesoft-datamover only claims .dms and .dmt, so it does not overlap at all.