# .tlv file generated by https://github.com/akkartik/teliva
# You may edit it if you are careful; however, you may see cryptic errors if you
# violate Teliva's assumptions.
#
# .tlv files are representations of Teliva programs. Teliva programs consist of
# sequences of definitions. Each definition is a table of key/value pairs. Keys
# and values are both strings.
#
# Lines in .tlv files always follow exactly one of the following forms:
# - comment lines at the top of the file starting with '#' at column 0
# - beginnings of definitions starting with '- ' at column 0, followed by a
# key/value pair
# - key/value pairs consisting of ' ' at column 0, containing either a
# spaceless value on the same line, or a multi-line value
# - multiline values indented by more than 2 spaces, starting with a '>'
#
# If these constraints are violated, Teliva may unceremoniously crash. Please
# report bugs at http://akkartik.name/contact
- __teliva_timestamp: original
str_helpers:
>-- some string helpers from http://lua-users.org/wiki/StringIndexing
>
>-- index characters using []
>getmetatable('').__index = function(str,i)
> if type(i) == 'number' then
> return str:sub(i,i)
> else
> return string[i]
> end
>end
>
>-- ranges using (), selected bytes using {}
>getmetatable('').__call = function(str,i,j)
> if type(i)~='table' then
> return str:sub(i,j)
> else
> local t={}
> for k,v in ipairs(i) do
> t[k]=str:sub(v,v)
> end
> return table.concat(t)
> end
>end
>
>-- iterate over an ordered sequence
>function q(x)
> if type(x) == 'string' then
> return x:gmatch('.')
> else
> return ipairs(x)
> end
>end
>
>-- insert within string
>function string.insert(str1, str2, pos)
> return str1:sub(1,pos)..str2..str1:sub(pos+1)
>end
>
>function string.remove(s, pos)
> return s:sub(1,pos-1)..s:sub(pos+1)
>end
>
>-- TODO: backport utf-8 support from Lua 5.3
- __teliva_timestamp: original
debugy:
>debugy = 5
- __teliva_timestamp: original
dbg:
>-- helper for debug by print; overlay debug information towards the right
>-- reset debugy every time you refresh screen
>function dbg(window, s)
> local oldy = 0
> local oldx = 0
> oldy, oldx = window:getyx()
> window:mvaddstr(debugy, 60, s)
> debugy = debugy+1
> window:mvaddstr(oldy, oldx, '')
>end
- __teliva_timestamp: original
check_eq:
>function check_eq(x, expected, msg)
> if eq(x, expected) then
> Window:addch('.')
> else
> print('F - '..msg)
> print(' expected '..str(expected)..' but got '..str(x))
> teliva_num_test_failures = teliva_num_test_failures + 1
> -- overlay first test failure on editors
> if teliva_first_failure == nil then
> teliva_first_failure = msg
> end
> end
>end
- __teliva_timestamp: original
eq:
>function eq(a, b)
> if type(a) ~= type(b) then return false end
> if type(a) == 'table' then
> if #a ~= #b then return false end
> for k, v in pairs(a) do
> if b[k] ~= v then
> return false
> end
> return true
> end
> end
> return a == b
>end
- __teliva_timestamp: original
str:
>-- smarter tostring
>-- slow; used only for debugging
>function str(x)
> if type(x) == 'table' then
> local result = ''
> result = result..#x..'{'
> for k, v in pairs(x) do
> result = result..str(k)..'='..str(v)..', '
> end
> result = result..'}'
> return result
> end
> return tostring(x)
>end
- __teliva_timestamp: original
map:
>-- only for arrays
>function map(l, f)
> result = {}
> for _, x in ipairs(l) do
> table.insert(result, f(x))
> end
> return result
>end
- __teliva_timestamp: original
reduce:
>-- only for arrays
>function reduce(l, f, init)
> result = init
> for _, x in ipairs(l) do
> result = f(result, x)
> end
> return result
>end
- __teliva_timestamp: original
filter:
>-- only for arrays
>function filter(l, f)
> result = {}
> for _, x in ipairs(l) do
> if f(x) then
> table.insert(result, x)
> end
> end
> return result
>end
- __teliva_timestamp: original
find_index:
>function find_index(arr, x)
> for n, y in ipairs(arr) do
> if x == y then
> return n
> end
> end
>end
- __teliva_timestamp: original
trim:
>function trim(s)
> return s:gsub('^%s*', ''):gsub('%s*$', '')
>end
- __teliva_timestamp: original
split:
>function split(s, d)
> result = {}
> for match in (s..d):gmatch("(.-)"..d) do
> table.insert(result, match);
> end
> return result
>end
- __teliva_timestamp: original
menu:
>-- To show app-specific hotkeys in the menu bar, add hotkey/command
>-- arrays of strings to the menu array.
>menu = {
> {'^k', 'clear'},
> {'^w', 'write prose to file "toot" (edit hotkey does NOT save)'},
>}
- __teliva_timestamp: original
Window:
>Window = curses.stdscr()
>curses.curs_set(0) -- we'll simulate our own cursor
- __teliva_timestamp: original
main:
>function main()
> init_colors()
>
> while true do
> render(Window)
> update(Window)
> end
>end
- __teliva_timestamp: original
init_colors:
>function init_colors()
> for i=0,7 do
> curses.init_pair(i, i, -1)
> end
> curses.init_pair(8, 7, 0)
> curses.init_pair(9, 7, 1)
> curses.init_pair(10, 7, 2)
> curses.init_pair(11, 7, 3)
> curses.init_pair(12, 7, 4)
> curses.init_pair(13, 7, 5)
> curses.init_pair(14, 7, 6)
> curses.init_pair(15, -1, 15)
>end
- __teliva_timestamp: original
prose:
>prose = ''
- __teliva_timestamp: original
cursor:
>cursor = 1
- __teliva_timestamp: original
render:
>function render(window)
> window:clear()
> debugy = 5
> local toots = split(prose, '\n\n===\n\n')
> pos = 1
> for i, toot in ipairs(toots) do
> if i > 1 then
> pos = render_delimiter(window, '\n\n===\n\n', pos, cursor)
> end
> pos = render_text(window, toot, pos, cursor)
> print('')
> window:attron(curses.A_BOLD)
> window:addstr(toot:len())
> window:attroff(curses.A_BOLD)
> end
> window:refresh()
>end
- __teliva_timestamp: original
render_delimiter:
>function render_delimiter(window, s, pos, cursor)
> local newpos = pos
> for i=1,s:len() do
> if newpos == cursor and i ~= 1 then
> if s[i] == '\n' then
> -- newline at cursor = render extra space in reverse video before jumping to new line
> window:attron(curses.A_REVERSE)
> window:addch(' ')
> window:attroff(curses.A_REVERSE)
> window:addstr(s[i])
> else
> -- most characters at cursor = render in reverse video
> window:attron(curses.A_REVERSE)
> window:addstr(s[i])
> window:attroff(curses.A_REVERSE)
> end
> else
> window:addstr(s[i])
> end
> newpos = newpos+1
> end
> return newpos
>end
- __teliva_timestamp: original
render_text:
>-- https://gankra.github.io/blah/text-hates-you
>-- https://lord.io/text-editing-hates-you-too
>
>-- manual tests:
>-- cursor on some character
>-- cursor on (within) '\n\n===\n\n' delimiter (delimiter is hardcoded; things may break if you change it)
>-- cursor at end of each line
>-- render digits
>
>-- positions serve two purposes:
>-- character to index into prose
>-- cursor-printing
>
>-- sequence of stories
>-- focus on rendering a single piece of text, first get that rock-solid
>-- split prose into toots, manage transitions between toots in response to cursor movements
>-- cursor movement: left/right vs up/down
>
>-- what is the ideal representation?
>-- prose + cursor has issues in multi-toot context. when to display cursor?
>function render_text(window, s, pos, cursor)
> local newpos = pos
>--? dbg(window, '--')
> for i=1,s:len() do
>--? dbg(window, tostring(newpos)..' '..tostring(string.byte(s[i])))
> if newpos == cursor then
>--? dbg(window, 'cursor: '..tostring(cursor))
> if s[i] == '\n' then
> -- newline at cursor = render extra space in reverse video before jumping to new line
> window:attron(curses.A_REVERSE)
> window:addch(' ')
> window:attroff(curses.A_REVERSE)
> window:addstr(s[i])
> else
> -- most characters at cursor = render in reverse video
> window:attron(curses.A_REVERSE)
> window:addstr(s[i])
> window:attroff(curses.A_REVERSE)
> end
> else
> window:addstr(s[i])
> end
> newpos = newpos+1
> end
> if newpos == cursor then
> window:attron(curses.A_REVERSE)
> window:addch(' ')
> window:attroff(curses.A_REVERSE)
> end
> return newpos
>end
- __teliva_timestamp: original
update:
>function update(window)
> local key = window:getch()
> local h, w = window:getmaxyx()
> if key == curses.KEY_LEFT then
> if cursor > 1 then
> cursor = cursor-1
> end
> elseif key == curses.KEY_RIGHT then
> if cursor <= #prose then
> cursor = cursor+1
> end
> elseif key == curses.KEY_DOWN then
> cursor = cursor_down(prose, cursor, w)
> elseif key == curses.KEY_UP then
> cursor = cursor_up(prose, cursor, w)
> elseif key == curses.KEY_BACKSPACE or key == 8 or key == 127 then -- ctrl-h, ctrl-?, delete
> if cursor > 1 then
> cursor = cursor-1
> prose = prose:remove(cursor)
> end
> elseif key == 11 then -- ctrl-k
> prose = ''
> cursor = 1
> elseif key == 23 then -- ctrl-w
> local out = io.open('toot', 'w')
> if out ~= nil then
> out:write(prose, '\n')
> out:close()
> end
> elseif key == 10 or (key >= 32 and key < 127) then
> prose = prose:insert(string.char(key), cursor-1)
> cursor = cursor+1
> end
>end
- __teliva_timestamp: original
cursor_down:
>function cursor_down(s, old_idx, width)
> local max = string.len(s)
> local i = 1
> -- compute oldcol, the screen column of old_idx
> local oldcol = 0
> local col = 0
> while true do
> if i > max then
> -- abnormal old_idx
> return old_idx
> end
> if i == old_idx then
> oldcol = col
> break
> end
> if s[i] == '\n' then
> col = 0
> else
> col = col+1
> end
> i = i+1
> end
> -- skip rest of line
> while true do
> if i > max then
> -- current line is at bottom
> if col >= width then
> return i
> end
> return old_idx
> end
> if s[i] == '\n' then
> break
> end
> if i - old_idx >= width then
> return i
> end
> col = col+1
> i = i+1
> end
> -- compute index at same column on next line
> -- i is at a newline
> i = i+1
> col = 0
> while true do
> if i > max then
> -- next line is at bottom and is too short; position at end of it
> return i
> end
> if s[i] == '\n' then
> -- next line is too short; position at end of it
> return i
> end
> if col == oldcol then
> return i
> end
> col = col+1
> i = i+1
> end
>end
>
>function test_cursor_down()
> -- lines that don't wrap
> check_eq(cursor_down('abc\ndef', 1, 5), 5, 'cursor_down: non-bottom line first char')
> check_eq(cursor_down('abc\ndef', 2, 5), 6, 'cursor_down: non-bottom line mid char')
> check_eq(cursor_down('abc\ndef', 3, 5), 7, 'cursor_down: non-bottom line final char')
> check_eq(cursor_down('abc\ndef', 4, 5), 8, 'cursor_down: non-bottom line end')
> check_eq(cursor_down('abc\ndef', 5, 5), 5, 'cursor_down: bottom line first char')
> check_eq(cursor_down('abc\ndef', 6, 5), 6, 'cursor_down: bottom line mid char')
> check_eq(cursor_down('abc\ndef', 7, 5), 7, 'cursor_down: bottom line final char')
> check_eq(cursor_down('abc\n\ndef', 2, 5), 5, 'cursor_down: to shorter line')
>
> -- within a single wrapping line
> -- |abcde| <-- wrap, no newline
> -- |fgh |
> check_eq(cursor_down('abcdefgh', 1, 5), 6, 'cursor_down from wrapping line: first char')
> check_eq(cursor_down('abcdefgh', 2, 5), 7, 'cursor_down from wrapping line: mid char')
> check_eq(cursor_down('abcdefgh', 5, 5), 9, 'cursor_down from wrapping line: to shorter line')
>
> -- within a single very long wrapping line
> -- |abcde| <-- wrap, no newline
> -- |fghij| <-- wrap, no newline
> -- |klm |
> check_eq(cursor_down('abcdefghijklm', 1, 5), 6, 'cursor_down within wrapping line: first char')
> check_eq(cursor_down('abcdefghijklm', 2, 5), 7, 'cursor_down within wrapping line: mid char')
> check_eq(cursor_down('abcdefghijklm', 5, 5), 10, 'cursor_down within wrapping line: final char')
>end
- __teliva_timestamp: original
cursor_up:
>function cursor_up(s, old_idx, width)
> local max = string.len(s)
> local i = 1
> -- compute oldcol, the screen column of old_idx
> local oldcol = 0
> local col = 0
> local newline_before_current_line = 0
> while true do
> if i > max or i == old_idx then
> oldcol = col
> break
> end
> if s[i] == '\n' then
> col = 0
> newline_before_current_line = i
> else
> col = col+1
> if col == width then
> col = 0
> end
> end
> i = i+1
> end
> -- find previous newline
> i = i-col-1
> if old_idx - newline_before_current_line > width then
> -- we're in a wrapped line
> return old_idx - width
> end
> -- scan back to start of previous line
> if s[i] == '\n' then
> i = i-1
> end
> while true do
> if i < 1 then
> -- current line is at top
> break
> end
> if s[i] == '\n' then
> break
> end
> i = i-1
> end
> -- i is at a newline
> i = i+1
> -- skip whole screen lines within previous line
> while newline_before_current_line - i > width do
> i = i + width
> end
> -- compute index at same column on previous screen line
> col = 0
> while true do
> if i > max then
> -- next line is at bottom and is too short; position at end of it
> return i
> end
> if s[i] == '\n' then
> -- next line is too short; position at end of it
> return i
> end
> if col == oldcol then
> return i
> end
> col = col+1
> i = i+1
> end
>end
>
>function test_cursor_up()
> -- lines that don't wrap
> check_eq(cursor_up('abc\ndef', 1, 5), 1, 'cursor_up: top line first char')
> check_eq(cursor_up('abc\ndef', 2, 5), 2, 'cursor_up: top line mid char')
> check_eq(cursor_up('abc\ndef', 3, 5), 3, 'cursor_up: top line final char')
> check_eq(cursor_up('abc\ndef', 4, 5), 4, 'cursor_up: top line end')
> check_eq(cursor_up('abc\ndef', 5, 5), 1, 'cursor_up: non-top line first char')
> check_eq(cursor_up('abc\ndef', 6, 5), 2, 'cursor_up: non-top line mid char')
> check_eq(cursor_up('abc\ndef', 7, 5), 3, 'cursor_up: non-top line final char')
> check_eq(cursor_up('abc\ndef\n', 8, 5), 4, 'cursor_up: non-top line end')
> check_eq(cursor_up('ab\ndef\n', 7, 5), 3, 'cursor_up: to shorter line')
>
> -- within a single wrapping line
> -- |abcde| <-- wrap, no newline
> -- |fgh |
> check_eq(cursor_up('abcdefgh', 6, 5), 1, 'cursor_up from wrapping line: first char')
> check_eq(cursor_up('abcdefgh', 7, 5), 2, 'cursor_up from wrapping line: mid char')
> check_eq(cursor_up('abcdefgh', 8, 5), 3, 'cursor_up from wrapping line: final char')
> check_eq(cursor_up('abcdefgh', 9, 5), 4, 'cursor_up from wrapping line: wrapped line end')
>
> -- within a single very long wrapping line
> -- |abcde| <-- wrap, no newline
> -- |fghij| <-- wrap, no newline
> -- |klm |
> check_eq(cursor_up('abcdefghijklm', 11, 5), 6, 'cursor_up within wrapping line: first char')
> check_eq(cursor_up('abcdefghijklm', 12, 5), 7, 'cursor_up within wrapping line: mid char')
> check_eq(cursor_up('abcdefghijklm', 13, 5), 8, 'cursor_up within wrapping line: final char')
> check_eq(cursor_up('abcdefghijklm', 14, 5), 9, 'cursor_up within wrapping line: wrapped line end')
>
> -- from below to (the bottom of) a wrapping line
> -- |abcde| <-- wrap, no newline
> -- |fg |
> -- |hij |
> check_eq(cursor_up('abcdefg\nhij', 9, 5), 6, 'cursor_up to wrapping line: first char')
> check_eq(cursor_up('abcdefg\nhij', 10, 5), 7, 'cursor_up to wrapping line: mid char')
> check_eq(cursor_up('abcdefg\nhij', 11, 5), 8, 'cursor_up to wrapping line: final char')
> check_eq(cursor_up('abcdefg\nhij', 12, 5), 8, 'cursor_up to wrapping line: to shorter line')
>end
- __teliva_timestamp:
>Thu Feb 17 19:52:30 2022
doc:blurb:
>A tiny editor (no scrolling) for composing a series of toots or tweets. Always shows character counts for current state of prose.
>
>Typing '===' on its own lines, surrounded by empty lines, partitions prose and gives all segments independent character counts. Good for threads (tweetstorms).
- __teliva_timestamp:
>Fri Mar 11 09:45:27 2022
first_toot:
>first_toot = 1
- __teliva_timestamp:
>Fri Mar 11 11:47:34 2022
update:
>function update(window)
> local key = window:getch()
> local h, w = window:getmaxyx()
> if key == curses.KEY_LEFT then
> if cursor > 1 then
> cursor = cursor-1
> end
> elseif key == curses.KEY_RIGHT then
> if cursor <= #prose then
> cursor = cursor+1
> end
> elseif key == curses.KEY_DOWN then
> cursor = cursor_down(prose, cursor, w)
> elseif key == curses.KEY_UP then
> cursor = cursor_up(prose, cursor, w)
> elseif key == curses.KEY_BACKSPACE or key == 8 or key == 127 then -- ctrl-h, ctrl-?, delete
> if cursor > 1 then
> cursor = cursor-1
> prose = prose:remove(cursor)
> end
> elseif key == 6 then -- ctrl-f
> first_toot = first_toot+1
> elseif key == 2 then -- ctrl-b
> if first_toot > 1 then
> first_toot = first_toot-1
> end
> elseif key == 11 then -- ctrl-k
> prose = ''
> cursor = 1
> elseif key == 23 then -- ctrl-w
> local out = io.open('toot', 'w')
> if out ~= nil then
> out:write(prose, '\n')
> out:close()
> end
> elseif key == 10 or (key >= 32 and key < 127) then
> prose = prose:insert(string.char(key), cursor-1)
> cursor = cursor+1
> end
>end
- __teliva_timestamp:
>Fri Mar 11 11:48:43 2022
menu:
>-- To show app-specific hotkeys in the menu bar, add hotkey/command
>-- arrays of strings to the menu array.
>menu = {
> {'^w', 'write to "toot"'},
> {'^f|^b', 'scroll'},
> {'^k', 'clear'},
>}
- __teliva_timestamp:
>Sat Mar 12 08:48:44 2022
render:
>function render(window)
> window:clear()
> debugy = 5
> local toots = split(prose, '\n\n===\n\n')
> pos = 1
> for i, toot in ipairs(toots) do
>--? dbg(window, "render: "..i.." pos "..pos.." cursor "..cursor)
> if i > 1 then
> pos = render_delimiter(window, '\n\n===\n\n', pos, cursor)
>--? dbg(window, "delim: "..pos.." cursor "..cursor)
> end
> if i <= first_toot then
> window:clear()
> end
> pos = render_text(window, toot, pos, cursor)
> print('')
>--? dbg(window, "text: "..pos.." cursor "..cursor)
> window:attron(curses.A_BOLD)
> window:addstr(toot:len())
> window:attroff(curses.A_BOLD)
> end
> window:refresh()
>end
- __teliva_timestamp:
>Sat Mar 12 08:57:41 2022
doc:blurb:
>A tiny editor (no scrolling) for composing a series of toots or tweets.
>Always shows character counts for current state of prose.
>
>Typing '===' on its own lines, surrounded by empty lines, partitions prose and gives all segments independent character counts. Good for threads (tweetstorms).
- __teliva_timestamp:
>Sat Mar 12 08:59:52 2022
__teliva_note:
>hacky scrolling support
>
>Since I started out rendering a toot at a time and tracking the position
>as I rendered each toot, the easiest way to build this was to scroll a
>toot at a time, always render each toot and just decide when to stop
>clearing the screen. This way I don't mess with the position computation
>logic which is carefully synced between render and cursor_up/cursor_down.
>
>But there may be a more elegant approach if I was building the current state
>from scratch.
doc:blurb:
>A tiny editor for composing a short series of toots or tweets. Always shows character counts for current state of prose.
>
>Typing '===' on its own lines, surrounded by empty lines, partitions prose and gives all segments independent character counts. Good for threads (tweetstorms).
>
>Scrolling support is rudimentary. Keys to scroll are independent of cursor movement, so cursor can move off the screen and confusingly 'get lost'.