Output formats: `--as`¶

Every disc command that produces tabular or hierarchical output accepts an --as flag selecting how that output is rendered.

--as display    pretty tables, headers, colour, Unicode glyphs
--as tsv        tab-separated values, one row per line
--as json       JSON object, one entry per report (see below)

Default per output stream¶

If you do not pass --as, disc chooses for you based on whether stdout is a terminal:

stdout	Default	Why
Terminal (interactive)	`display`	Headers, alignment, and styling make the output legible at a glance.
Pipe or redirect	`tsv`	One row per line, stable column order, no decorative chrome so downstream tools (`awk`, `cut`, `column -t`, …) get parseable input.

The detection happens at command start; nothing about your shell context needs configuring. Override it explicitly with --as when you want pretty output piped to a pager, or scriptable output to a terminal:

disc ls scsi0.dat | less                # pipe → tsv by default
disc ls scsi0.dat --as display | less   # force pretty
disc ls scsi0.dat --as json > catalog.json

The three formats in detail¶

display¶

The interactive format. Tables are rendered with column headers, aligned cells, and titles; trees use box-drawing characters; key/ value reports use a labelled two-column layout. Long values are truncated to fit the terminal width. Colour is applied where the terminal supports it.

Reach for display when a human is reading the output. Avoid it when scripting — the column widths and any truncation depend on terminal size and are not part of the stable contract.

tsv¶

Tab-separated values, one row per line. Column headers (if not suppressed) are emitted on the first line, prefixed with # so awk and friends can skip the line cheaply:

$ disc ls scsi0.dat --as tsv
#name        load    exec    length  attr
!BOOT        4294967295      4294967295      20      WR/R
ELITE        6400    32803   23552   LR/R
…

Addresses and lengths are emitted as plain decimal integers — the base is irrelevant to a consumer, so cut/awk get a number they can use directly rather than a hex string to re-parse.

Field separators are literal tab characters (\t); no quoting, no escaping. The column order, names, and value formats are part of the CLI’s stable contract — they will not change without a major version bump.

Reach for tsv when feeding output into shell pipelines or other column-aware tooling.

json¶

A JSON object, with one key per report the command produced. Each report is rendered as a list of objects (one per row) for tabular data, or as a nested tree of objects for hierarchical data:

$ disc stat scsi0.dat --as json
{
  "disc": {
    "geometry": "615 cylinders, 4 heads, 32 sectors/track",
    "size": 20152320
  },
  "partition_1": { … },
  "partition_2": { … }
}

Field names match the tsv column names. Numeric quantities — load and exec addresses, file lengths, disc and partition sizes, user quotas — are rendered as JSON numbers, not strings, so a consumer can use them arithmetically without parsing. The display format shows the same values human-first (addresses as 0x-prefixed hex, sizes and quotas in friendly units like 19.2 MiB); the machine formats (tsv, json) always carry the raw integer.

Reach for json when consuming output from a structured-data tool (jq, yq, a Python or Node script) — the structure is the same shape your script will want.

Multi-report commands¶

Some commands logically return more than one piece of information:

disc stat OUTER_PATH returns a disc block (physical geometry + total size) and one partition_N block per filesystem partition (partition_1 for DFS or ADFS, partition_2 for AFS on a dual-partition image).
disc afs plan returns several blocks (geometry, occupancy, suggested afs init invocation, …).

Each sub-report has a stable name. You can ask for just the parts you want:

disc stat scsi0.dat --report partition_2          # AFS only
disc stat scsi0.dat --report disc --report partition_1
disc stat scsi0.dat --no-reports                  # silence display

--report may be repeated; --no-reports suppresses every report (the command’s side effects still run, which matters for action commands whose reports are incidental). --all-reports forces every report on, in case the command’s default filter would hide some.

Header control¶

The --header / --no-header flags toggle the leading row in tsv and the column-header row in display. JSON output is shape-stable regardless and ignores both flags. Default is “include” — turn it off when you want a header-less stream for awk -F'\t' pipelines.

Detailed vs essential columns¶

Some commands carry both a brief and a verbose column set — disc ls for instance can include the raw access byte alongside the symbolic form (--access-byte). The --detailed / --essential flags pick between them when the command exposes both:

--essential — the minimum useful column set (the default for display and tsv).
--detailed — every column the command knows how to emit (the default for json, since downstream scripts can simply ignore fields they do not need).

Either flag overrides the format-driven default.

Output formats: --as¶