Commit Graph

617 Commits

Author SHA1 Message Date
4f20379219 take care of a few tests after refactoring/making a different design decision 2025-09-23 11:04:47 +01:00
961ce63023 begin refactoring to let cursor move to very last char of file (including unix line ending) 2025-09-23 10:50:10 +01:00
8931d4d071 fix minor compilation error introduced in previous commit 2025-09-23 09:29:02 +01:00
ec462a4df7 remove special case from 'k' motion: the exact semantics we want for trailing newlines at the end of a file are currently a bit unclear and require exploration 2025-09-23 09:27:13 +01:00
21e12addf3 in text builders, check when on newline if this is the second last character in the file, and draw it on a newline if so. This is to align with the design decision that we want the second-last character to be selectable 2025-09-23 09:16:02 +01:00
505521f5e2 minor cosmetic change 2025-09-23 09:04:18 +01:00
c9365076ad handle special case with 'k' motion: when file ends with \n\n and cursor is on first \n: we should be able to move upwards from here too 2025-09-23 08:54:19 +01:00
79d232df7a amend very minor bug: when looping through DFA and hitting a 'final state', we let the next chr afterwards handle the start state, but we want the current char to handle the start state instead, as this matches vim's behaviour 2025-09-23 07:36:51 +01:00
0b0d5268fc handle edge case when deleting by DFA: when we are deleting to the end of the file and if cursor is on first character after newline, then we also want to delete preceding newline too. Or else we will end up with two newlines at the end of the file, which is not what we want. 2025-09-23 07:18:58 +01:00
ec10f3e025 tiny amount of formatting 2025-09-23 07:04:41 +01:00
9174e4923c don't deleteByDfa when buffer's length is 1, because that would mean that we would be deleting last newline in file 2025-09-23 07:04:20 +01:00
02ada5765b improve 'MakeNormalDelete.deleteByDfa' function by refraining from deleting the unix-style line ending, if DFA takes us past the ending 2025-09-23 06:09:03 +01:00
cb0dec93a0 simplify implementation of 'MakeNormalDelete.deleteLine 2025-09-22 13:02:08 +01:00
e3148e6ac1 when deleting using a DFA motion, make sure we have a linebreak at the end so that we still have unix-style line endings as an invariant 2025-09-22 12:41:33 +01:00
351f7e9bf9 fix edge case with 'd$' motion: if deleting to end of file, we need to append two newlines 2025-09-22 12:28:08 +01:00
6f968ee8e5 in 'dk' motion, make sure we move the cursor to the start of the previous line in all cases 2025-09-22 09:46:43 +01:00
e6ea96f995 fix bug in 'k' motion: when we move from any other line to line 0, always unconditionally decrement column by 1 2025-09-22 07:02:20 +01:00
3128647a5f reimplement 'k' motion so that newlines which are immediately preceded by non-newline chars are skipped 2025-09-22 06:49:00 +01:00
5315aff600 reimplement 'j' motion so that newline is skipped only if immediately preceded by char 2025-09-22 05:32:03 +01:00
cdc23945a5 reimplement vi's 'h' movement, so that we only skip newline characters that immediately follow a non-newline 2025-09-21 21:36:59 +01:00
d63201ada7 reimplement vi's 'l' motion so that we only skip a newline when it is preceded by a non-newline character 2025-09-21 17:49:30 +01:00
639f71009c done reimplementing 'MakeNormalDelete.deleteLineBack', taking care of cases where we use this motion while on the first line or on the last 2025-09-21 16:40:05 +01:00
3ea44d9725 progress reimplementing 'NormalDelete.deleteLineBack' function 2025-09-21 15:45:14 +01:00
6af2b58b0f make 'dl' motion (including count) call the same function as the 'x' motion, because they have identical behaviour. 2025-09-21 00:22:24 +01:00
2758b864bc fix bug found by unit test: when deleting leftwards, we should never delete beyond the end of the line 2025-09-20 18:04:59 +01:00
c29afbdb7c fix minor bug regarding 'NormalMove.moveCursorDown': after a <count-j> that moved the cursor to the last line in the buffer, the cursor would be at the last column too, even if that wasn't the previous column. To address this issue, we take out code that calculates the lineIdx from the 'NormalMove.finishMoveCursorUpDown' function, and have both 'NormalMove.moveCursorUp' and 'NormalMove.moveCursorDown' calculate their own lineIdx in different ways and pass it to the 'NormalMove.finishMoveCursorUpAndDown' function. 2025-09-20 02:58:51 +01:00
090e7ccea2 pass 'maxLength' value to TextScroll function so that we can prevent scrolling past the bottom of the file (unless we want to od so by centtering the screen) 2025-09-19 23:17:28 +01:00
d3fa4d08bf extract reusable function containing common code previously used in both 'NormalMove.moveCursorUp' and 'NormalMove.moveCursorDown' functions, and make them call the reusable function instead 2025-09-19 22:40:29 +01:00
859860b19f reimplement vi's 'j' motion, and remove the original function for this motion which was in cursor.sml 2025-09-19 22:33:37 +01:00
2c88341c37 remove code which became dead after reimplementing vi's 'k' motion 2025-09-19 05:27:10 +01:00
9ec8891ce5 fix bugs in new function for moving cursor upwards, taking care to calculate column different if on first line (where there is no preceding newline) vs any line other than the first 2025-09-19 05:23:20 +01:00
4c1a76ae67 begin reimplementing motion to move cursor up (vi's 'k' command), although it is in progress and buggy 2025-09-19 04:59:40 +01:00
e4e3961d41 in NormalMove.moveToLine function, use new LineGap function, 'LineGap.lineNumberToIdx', to get the index a line is at, and delete the Cursor.getLineStartIdx' function which we no longer need, as that functionality is in brolib-sml itself. 2025-09-19 04:45:29 +01:00
b6244d4962 amend motion to move to a line: if the requested line has a linebreak as the next character after it, stay at the current position, or else add 1 to the linebreak's idx so that we go to the first character of the line 2025-09-19 04:03:54 +01:00
d871ac451f restore temp.txt to previous state, as it was temporarily being edited for testing 2025-09-19 03:04:10 +01:00
ab2f2cc752 simplify 'vi-hl-dfa's DFA, removing one state 2025-09-19 02:40:14 +01:00
ad7d103f61 rename 'ViLDfa' structure to 'ViHlDfa' since this DFA implements both vi's 'h' motion and vi's 'l' motion 2025-09-17 11:02:11 +01:00
dbf05ec8dc refactor Cursor.viH to use DFA 2025-09-17 10:45:37 +01:00
9ddb5f68d7 fix bugs in vi-l-dfa (we want to stop looping when the counter is ~1, because the starting character will likely be final/a not-newline, and we want to loop at least once; also, we want to special case 'twoNewlineState' as a final case which causes us to go backwards by 1 instead of treating 'oneNewlineState' as a final/special case). We also modify other code to use the new vi-l implementation in the program. 2025-09-17 03:52:31 +01:00
5174a5b0a4 adjustments to vi-l-dfa: we should accept a single-newline as a final state, but not a double-newline 2025-09-17 02:35:32 +01:00
161caee62a create DFA for Vi's 'l' motion 2025-09-17 02:31:38 +01:00
c94aa72015 create 'MakeIfCharFolderNext' functor, which mirrors the 'MakeIfCharFolderPrev' functor that we already have 2025-09-17 02:20:36 +01:00
97936b83a0 in TextBuilder functions, when we encounter a newline, we want any cursor here to be drawn after a linebreak (after the posX and posY values go to the next line) and not before 2025-09-16 22:29:19 +01:00
1adb74df2c pull in new brolib-sml changes (which just changes one particular function's name to be clearer), and modify code to use the correct name 2025-09-16 22:20:55 +01:00
c4fedc6434 fix 'Cursor.getLineStartIdx' function, which should look at the left nodes if there is no relevant line in this node 2025-09-16 22:09:15 +01:00
96521f358f when looping through a DFA in a functor, make sure we return the absolute index to the calling function and not a modified version of the absolute index which has already been clipped 2025-09-16 21:35:48 +01:00
12478f8ac2 remove references to 'Cursor.clipIdx' and delete that function once it is no longer used 2025-09-16 16:08:01 +01:00
19d2235686 refactor 'MakeNormalDelete.removeLineBreaks' function to remove function call to 'Cursor.clipIdx', instead relying on textLength field in buffer 2025-09-16 15:53:41 +01:00
2c5449dc0a rewrite 'MakeMove' functor in fcore/move.sml to use textLength field instead of Cursor.clipIdx function 2025-09-16 07:59:26 +01:00
148b108062 when using 'G' command to move to end fo file, use the 'textLength' field to calculate the new bufferIdx, which is more sensible than using the LineGap's idx for the same purpose 2025-09-16 07:52:47 +01:00