The space between authors in a chat is a preference

Two preferences in the Display group: "Author spacing" (auto, on, off;
default auto) and "Spacing from" (the smallest terminal, in rows, that gets
the blank line with auto; default 24). The test reads the real rows of the
terminal before each paint, so a small desktop window drops the line as the
Cardputer Zero does, and a change applies at the next paint. The transcript
cache stores the spacing, so a change draws the lines again. The CLI
manual's sample of `preferences show` is the real output again.

Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
This commit is contained in:
Jean-Pierre MartineauandClaude Opus 5.5 committed 2026-10-13 18:16:15 -04:00
1 parent 1c678d43f3
commit 0c3a8b5b5c
7 files changed
+134 -55

No files matched your search

+5 -2
View File
@@ -23,8 +23,11 @@ can change how things behave. It does not only add new things.
the node uses when it makes its own key. A firmware companion still gets the full 64-byte
key, because it accepts only that form.
- **In a chat, a blank line comes after the messages of each author.** Thus the
conversation is easier to follow when many people write. The Cardputer Zero has no
blank line there, because its screen has only 14 rows and each row shows a message.
conversation is easier to follow when many people write. By default, the line shows on
a terminal of 24 rows or more. A smaller terminal, and the 14 rows of the Cardputer Zero,
keep each row for a message. Two new preferences in the Display group control the line:
**Author spacing** (`auto`, `on`, or `off`) and **Spacing from** (the smallest terminal,
in rows, that gets the line with `auto`).
## [0.11.1] - 2026-10-13
+28 -25
View File
@@ -1727,31 +1727,34 @@ If you give no subcommand, it runs `show`.
```console
$ meshterm preferences show
PREFERENCE VALUE DEFAULT DESCRIPTION
set_clock_on_connect off off Set the device clock from this computer when it connects
trace_cooldown_s 5 5 Wait this long before transmitting again
flood_advert_cooldown_s 60 60 Extra wait before another mesh-wide advert
weekly_flood_advert off off Flood an advert after a week without one
advert_quiet_s 30 30 Silence to wait for before the weekly advert
direct_message_soft_retries 0 0 Times to resend a message that gets no reply
tx_opt_min 18 18 Weakest power it will try
tx_opt_max 28 28 Strongest power it will try
tx_snr_tolerance_db 1 1 Signals this close are a tie, so less power wins
watch_silence_hours 12 12 Warn when a watched node goes quiet this long
watch_alerts_kept 200 200 How many past alerts to keep
map_view_fraction 0.5 0.5 Share of nodes to fit on screen; 1 shows them all
basemap_tilejson_url https://tiles.openfreemap.org/planet https://tiles.openfreemap.org/planet Where map images come from (next launch)
history_days 365 365 Older ones are deleted; 0 keeps everything
chat_history_limit 200 200 Old messages to load when you open a chat
courier_history_kept 100 100 Finished outbox messages to keep
cli_time_format relative relative Whether the command line prints ages or timestamps
fast_render on on Faster screen updates; turn off if it looks wrong (next launch)
full_width auto auto Use the extra column some terminals hide
color_depth auto auto How many colours to send this terminal (next launch)
console_setup auto auto Move to a better terminal when this console can't draw MeshTerm
console_font 6x12 6x12 Bigger text, or more rows on the handheld's panel
name_overlap_days 7 7 Grey a name two nodes used within this many days
log_level WARNING WARNING How much MeshTerm writes to its log file (next launch)
PREFERENCE VALUE DEFAULT DESCRIPTION
set_clock_on_connect off off Set the device clock from this computer when it connects
trace_cooldown_s 5 5 Wait this long before transmitting again
flood_advert_cooldown_s 60 60 Extra wait before another mesh-wide advert
weekly_flood_advert off off Flood an advert after a week without one
advert_quiet_s 30 30 Silence to wait for before the weekly advert
direct_message_soft_retries 0 0 Times to resend a message that gets no reply
room_relogin_minutes 30 30 Opening a room this quiet logs in again
tx_opt_min 18 18 Weakest power it will try
tx_opt_max 28 28 Strongest power it will try
tx_snr_tolerance_db 1 1 Signals this close are a tie, so less power wins
watch_silence_hours 12 12 Warn when a watched node goes quiet this long
watch_alerts_kept 200 200 How many past alerts to keep
map_view_fraction 0.5 0.5 Share of nodes to fit on screen; 1 shows them all
basemap_tilejson_url https://tiles.openfreemap.org/planet https://tiles.openfreemap.org/planet Where map images come from (next launch)
history_days 365 365 Older ones are deleted; 0 keeps everything
chat_history_limit 200 200 Old messages to load when you open a chat
courier_history_kept 100 100 Finished outbox messages to keep
cli_time_format relative relative Whether the command line prints ages or timestamps
fast_render on on Faster screen updates; turn off if it looks wrong (next launch)
full_width auto auto Use the extra column some terminals hide
color_depth auto auto How many colours to send this terminal (next launch)
console_setup auto auto Move to a better terminal when this console can't draw MeshTerm
console_font 6x12 6x12 Bigger text, or more rows on the handheld's panel
name_overlap_days 7 7 Grey a name two nodes used within this many days
chat_author_spacing auto auto A blank line after each author in a chat
chat_author_spacing_min_rows 24 24 With auto: the smallest terminal that gets it
log_level WARNING WARNING How much MeshTerm writes to its log file (next launch)
```
**`DESCRIPTION`** says what each preference does, because the name of a key does not give
+30
View File
@@ -173,6 +173,10 @@ _QUIET_CHOICES: dict[int, str] = {s: f"{s} s" for s in QUIET_CHOICES_S}
#: column, and for the cost when MeshTerm reclaims a column that is not hidden.
_WIDTH_CHOICES: dict[str, str] = {"auto": "auto", "yes": "yes", "no": "no"}
#: When a chat puts a blank line after the messages of each author. ``auto`` puts it on a
#: terminal that has at least ``chat_author_spacing_min_rows`` rows.
_SPACING_CHOICES: dict[str, str] = {"auto": "auto", "on": "on", "off": "off"}
#: How many colours to send to the terminal. ``auto`` is the honest default, and it is the
#: only value that is not a depth. It means "take prompt_toolkit's verdict unless we can
#: positively establish better". This rule is the difference between a repaired host and a
@@ -504,6 +508,32 @@ PREFERENCES: tuple[PrefSpec, ...] = (
maximum=3650,
unit="days",
),
PrefSpec(
key="chat_author_spacing",
label="Author spacing",
help="A blank line after each author in a chat",
group="Display",
value_type="enum",
# Auto, because the line costs one row for each group of messages. A tall terminal
# has the rows, and the 14 rows of the Cardputer Zero do not: there a blank line is
# a message less on the screen. The test is the real height of the terminal, so a
# small window on a desktop drops the line too (JP, 2026-10-13).
default="auto",
choices=_SPACING_CHOICES,
),
PrefSpec(
key="chat_author_spacing_min_rows",
label="Spacing from",
help="With auto: the smallest terminal that gets it",
group="Display",
value_type="int",
# A standard terminal (24 rows) and the PicoCalc (26) get the line. The Cardputer
# Zero (14) and a smaller window do not (JP, 2026-10-13).
default=24,
minimum=10,
maximum=100,
unit="rows",
),
# --- Diagnostics ---------------------------------------------------------------
PrefSpec(
key="log_level",
+32 -17
View File
@@ -25,7 +25,7 @@ from rich.cells import cell_len
from rich.console import Group, RenderableType
from rich.text import Text
from ..core import links
from ..core import links, preferences
from ..core.channels import MENTION, split_channel_sender
from ..core.connection import ContactNotOnDeviceError
from ..core.events import EventKind, MeshEvent
@@ -75,26 +75,21 @@ _URL_GUTTER = 2
_URL_CODES = False
#: Whether a blank line closes the messages of each author. It costs one row for each
#: group. The PicoCalc (26 rows) and a terminal (24) have the rows for it. The 14 rows of
#: the Cardputer Zero do not: there a blank line on each group is a message less on the
#: screen (JP, 2026-10-13).
_GROUP_GAP = True
#: The fewest rows (``Platform.readable_rows``) for which a transcript has that blank line.
_GROUP_GAP_MIN_ROWS = 20
#: The rows of the terminal when the session cannot say (a stand-in session with no
#: ``terminal_rows``): the design minimum of the platform (``Platform.readable_rows``).
_FALLBACK_ROWS = 24
@on_platform
def _bind(platform: Platform) -> None:
"""Bind the transcript to the platform (now and at each switch).
The platform says whether codes hang under URLs, and whether it has the rows for a
blank line after each author.
The platform says whether codes hang under URLs, and how many rows to assume when the
session cannot tell the height of the terminal.
"""
global _URL_CODES, _GROUP_GAP
global _URL_CODES, _FALLBACK_ROWS
_URL_CODES = platform.url_codes
_GROUP_GAP = platform.readable_rows >= _GROUP_GAP_MIN_ROWS
_FALLBACK_ROWS = platform.readable_rows
#: The maximum width of a footer hint, in cells (refer to the UX standards).
@@ -583,7 +578,8 @@ class ChatScreen(Screen):
messages = self._messages
total = len(messages)
cache = self._render_cache
if cache is not None and cache["width"] == width:
spacing = self._author_spacing()
if cache is not None and cache["width"] == width and cache["spacing"] == spacing:
cached_ids = cache["ids"]
cached_acked = cache["acked"]
cached_selected = cache["selected"]
@@ -628,10 +624,10 @@ class ChatScreen(Screen):
and cached_ids[idx] == id(message)
and cached_acked[idx] == message.acked
)
# A blank line closes the messages of an author, where the platform has the rows
# for it (:data:`_GROUP_GAP`). It is the first line of the slice of the message
# A blank line closes the messages of an author, where the preferences ask for
# it (:meth:`_author_spacing`). It is the first line of the slice of the message
# that starts the next group, so a cached slice brings it back with it.
gap = _GROUP_GAP and bool(lines) and (new_day or group != prev_group)
gap = spacing and bool(lines) and (new_day or group != prev_group)
if reusable:
start = cached_boundaries[idx - 1] if idx > 0 else 0
slice_lines = cached_lines[start : cached_boundaries[idx]]
@@ -675,6 +671,7 @@ class ChatScreen(Screen):
self._render_cache = {
"width": width,
"spacing": spacing,
"ids": ids,
"acked": ackeds,
"selected": selecteds,
@@ -683,6 +680,24 @@ class ChatScreen(Screen):
}
return lines
def _author_spacing(self) -> bool:
"""Whether a blank line closes the messages of each author, at this paint.
The ``chat_author_spacing`` preference decides: ``on`` and ``off`` are fixed, and
``auto`` puts the line on a terminal of ``chat_author_spacing_min_rows`` rows or
more. The line costs one row for each group of messages, so a short terminal (the
Cardputer Zero, or a small window) keeps the transcript dense. The test reads the
rows of the terminal before the paint, and reads the preferences at each paint,
so a change on the Preferences page applies when the user comes back.
"""
prefs = preferences.current()
mode = prefs.chat_author_spacing
if mode != "auto":
return mode == "on"
rows = getattr(self._session, "terminal_rows", None)
height = rows() if callable(rows) else _FALLBACK_ROWS
return height >= prefs.chat_author_spacing_min_rows
def _sender_and_body(self, message: ChatMessage) -> tuple[str, str]:
"""Return the display sender and the clean body of a message.
+9
View File
@@ -2056,6 +2056,15 @@ class TuiSession:
size = self._app.output.get_size() # type: ignore[union-attr]
return max(20, size.columns), max(6, size.rows)
def terminal_rows(self) -> int:
"""The rows of the terminal now. A screen reads it before its paint, never after.
The chat uses it for the space between authors (``chat_author_spacing``). The rows
that the frame records after a paint come one paint late, so the first paint of a
chat would come out without the space and then jump.
"""
return self._size()[1]
def base_body_size(self) -> tuple[int, int]:
"""Return the ``(width, height)`` in cells that the body of the base screen can use.
+27 -11
View File
@@ -41,12 +41,17 @@ from tests.conftest import plain as _strip_ansi # the only strip-and-join scree
class _StubSession:
"""A session stand-in that counts the requests for a paint."""
def __init__(self) -> None:
def __init__(self, rows: int = 24) -> None:
self.invalidations = 0
self.rows = rows
def invalidate(self) -> None:
self.invalidations += 1
def terminal_rows(self) -> int:
"""The rows of the terminal, for the space between authors."""
return self.rows
def run_detached(self, work): # noqa: ANN001, ANN201
"""Start the flow of a key handler as a task, as ``TuiSession.run_detached`` does."""
return asyncio.ensure_future(work)
@@ -1545,17 +1550,20 @@ def test_channel_own_messages_do_not_merge_with_remote_namesake() -> None:
assert rendered.count("you") == 1 and rendered.count("★ Me") == 1
def test_a_blank_line_closes_each_author_where_the_rows_allow() -> None:
"""A terminal and the PicoCalc put a blank line after the messages of each author.
def test_a_blank_line_closes_each_author_where_the_rows_allow(
monkeypatch: pytest.MonkeyPatch,
) -> None:
"""With ``auto``, a terminal of 24 rows or more puts a blank line after each author.
The blank line also comes before the divider of a new day, and never before the first
line. The 14 rows of the Cardputer Zero have no row to spare for it. A second paint
reads the cached lines, and the divider of the second day still pins at its own line,
not at the blank line before it.
line. A terminal of fewer rows (the 14 of the Cardputer Zero) has no row to spare for
it. A second paint reads the cached lines, and the divider of the second day still pins
at its own line, not at the blank line before it. ``on`` and ``off`` ignore the rows,
and a change applies at the next paint.
"""
from datetime import datetime, timedelta, timezone
from meshterm.platforms import CARDPUTER_ZERO, REGULAR, set_platform
from meshterm.core import preferences
base = datetime(2026, 7, 5, 14, 24, tzinfo=timezone.utc)
@@ -1570,8 +1578,9 @@ def test_a_blank_line_closes_each_author_where_the_rows_allow() -> None:
ChatMessage(text="Bob: later", is_channel=True, channel_idx=0, created_at=at(1440)),
]
set_platform(REGULAR)
screen = _channel_screen(transcript())
prefs = preferences.Preferences()
monkeypatch.setattr(preferences, "current", lambda: prefs)
screen = _channel_screen(transcript(), session=_StubSession(rows=24))
lines = [_strip_ansi(line) for line in screen._render_grouped(53)]
assert lines[0].startswith("──") and lines[1] == "Alice"
assert lines[4] == "" and lines[5] == "Bob" # after the two messages of Alice
@@ -1581,9 +1590,16 @@ def test_a_blank_line_closes_each_author_where_the_rows_allow() -> None:
screen._render_grouped(53) # the cached lines
assert screen._sticky_headers == first and [at for at, _ in first] == [0, 8]
set_platform(CARDPUTER_ZERO)
screen = _channel_screen(transcript())
short = _channel_screen(transcript(), session=_StubSession(rows=14))
assert "" not in [_strip_ansi(line) for line in short._render_grouped(53)]
prefs.set("chat_author_spacing", "on") # always, also on 14 rows
assert "" in [_strip_ansi(line) for line in short._render_grouped(53)]
prefs.set("chat_author_spacing", "off") # never, and the cached lines change too
assert "" not in [_strip_ansi(line) for line in screen._render_grouped(53)]
prefs.set("chat_author_spacing", "auto")
prefs.set("chat_author_spacing_min_rows", 14)
assert "" in [_strip_ansi(line) for line in short._render_grouped(53)]
def test_a_mention_of_our_name_takes_our_chip() -> None:
+3
View File
@@ -147,6 +147,9 @@ class _GallerySession:
def base_body_size(self) -> tuple[int, int]:
return self._cols, self._rows
def terminal_rows(self) -> int:
return self._rows
async def button_dialog(self, prompt, buttons, **kwargs): # noqa: ANN001
return None