Merge remote-tracking branch 'upstream/master' into mix_20240309HEADrahm

21 files changed, 6606 insertions, 0 deletions

diff --git a/src/nvim/vterm/LICENSE b/src/nvim/vterm/LICENSE
new file mode 100644
index 0000000000..0d051634b2
--- /dev/null
+++ b/src/nvim/vterm/LICENSE
@@ -0,0 +1,23 @@
+
+
+The MIT License
+
+Copyright (c) 2008 Paul Evans <leonerd@leonerd.org.uk>
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in
+all copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+THE SOFTWARE.
diff --git a/src/nvim/vterm/README.md b/src/nvim/vterm/README.md
new file mode 100644
index 0000000000..4cc43bfb94
--- /dev/null
+++ b/src/nvim/vterm/README.md
@@ -0,0 +1 @@
+Adopted from [libvterm](https://www.leonerd.org.uk/code/libvterm/)
diff --git a/src/nvim/vterm/encoding.c b/src/nvim/vterm/encoding.c
new file mode 100644
index 0000000000..f9061e8e50
--- /dev/null
+++ b/src/nvim/vterm/encoding.c
@@ -0,0 +1,271 @@
+#include "nvim/vterm/encoding.h"
+#include "nvim/vterm/vterm_internal_defs.h"
+
+#ifdef INCLUDE_GENERATED_DECLARATIONS
+# include "vterm/encoding.c.generated.h"
+#endif
+
+#define UNICODE_INVALID 0xFFFD
+
+#if defined(DEBUG) && DEBUG > 1
+# define DEBUG_PRINT_UTF8
+#endif
+
+struct UTF8DecoderData {
+  // number of bytes remaining in this codepoint
+  int bytes_remaining;
+
+  // number of bytes total in this codepoint once it's finished
+  // (for detecting overlongs)
+  int bytes_total;
+
+  int this_cp;
+};
+
+static void init_utf8(VTermEncoding *enc, void *data_)
+{
+  struct UTF8DecoderData *data = data_;
+
+  data->bytes_remaining = 0;
+  data->bytes_total = 0;
+}
+
+static void decode_utf8(VTermEncoding *enc, void *data_, uint32_t cp[], int *cpi, int cplen,
+                        const char bytes[], size_t *pos, size_t bytelen)
+{
+  struct UTF8DecoderData *data = data_;
+
+#ifdef DEBUG_PRINT_UTF8
+  printf("BEGIN UTF-8\n");
+#endif
+
+  for (; *pos < bytelen && *cpi < cplen; (*pos)++) {
+    uint8_t c = (uint8_t)bytes[*pos];
+
+#ifdef DEBUG_PRINT_UTF8
+    printf(" pos=%zd c=%02x rem=%d\n", *pos, c, data->bytes_remaining);
+#endif
+
+    if (c < 0x20) {  // C0
+      return;
+    } else if (c >= 0x20 && c < 0x7f) {
+      if (data->bytes_remaining) {
+        cp[(*cpi)++] = UNICODE_INVALID;
+      }
+
+      cp[(*cpi)++] = c;
+#ifdef DEBUG_PRINT_UTF8
+      printf(" UTF-8 char: U+%04x\n", c);
+#endif
+      data->bytes_remaining = 0;
+    } else if (c == 0x7f) {  // DEL
+      return;
+    } else if (c >= 0x80 && c < 0xc0) {
+      if (!data->bytes_remaining) {
+        cp[(*cpi)++] = UNICODE_INVALID;
+        continue;
+      }
+
+      data->this_cp <<= 6;
+      data->this_cp |= c & 0x3f;
+      data->bytes_remaining--;
+
+      if (!data->bytes_remaining) {
+#ifdef DEBUG_PRINT_UTF8
+        printf(" UTF-8 raw char U+%04x bytelen=%d ", data->this_cp, data->bytes_total);
+#endif
+        // Check for overlong sequences
+        switch (data->bytes_total) {
+        case 2:
+          if (data->this_cp < 0x0080) {
+            data->this_cp = UNICODE_INVALID;
+          }
+          break;
+        case 3:
+          if (data->this_cp < 0x0800) {
+            data->this_cp = UNICODE_INVALID;
+          }
+          break;
+        case 4:
+          if (data->this_cp < 0x10000) {
+            data->this_cp = UNICODE_INVALID;
+          }
+          break;
+        case 5:
+          if (data->this_cp < 0x200000) {
+            data->this_cp = UNICODE_INVALID;
+          }
+          break;
+        case 6:
+          if (data->this_cp < 0x4000000) {
+            data->this_cp = UNICODE_INVALID;
+          }
+          break;
+        }
+        // Now look for plain invalid ones
+        if ((data->this_cp >= 0xD800 && data->this_cp <= 0xDFFF)
+            || data->this_cp == 0xFFFE
+            || data->this_cp == 0xFFFF) {
+          data->this_cp = UNICODE_INVALID;
+        }
+#ifdef DEBUG_PRINT_UTF8
+        printf(" char: U+%04x\n", data->this_cp);
+#endif
+        cp[(*cpi)++] = (uint32_t)data->this_cp;
+      }
+    } else if (c >= 0xc0 && c < 0xe0) {
+      if (data->bytes_remaining) {
+        cp[(*cpi)++] = UNICODE_INVALID;
+      }
+
+      data->this_cp = c & 0x1f;
+      data->bytes_total = 2;
+      data->bytes_remaining = 1;
+    } else if (c >= 0xe0 && c < 0xf0) {
+      if (data->bytes_remaining) {
+        cp[(*cpi)++] = UNICODE_INVALID;
+      }
+
+      data->this_cp = c & 0x0f;
+      data->bytes_total = 3;
+      data->bytes_remaining = 2;
+    } else if (c >= 0xf0 && c < 0xf8) {
+      if (data->bytes_remaining) {
+        cp[(*cpi)++] = UNICODE_INVALID;
+      }
+
+      data->this_cp = c & 0x07;
+      data->bytes_total = 4;
+      data->bytes_remaining = 3;
+    } else if (c >= 0xf8 && c < 0xfc) {
+      if (data->bytes_remaining) {
+        cp[(*cpi)++] = UNICODE_INVALID;
+      }
+
+      data->this_cp = c & 0x03;
+      data->bytes_total = 5;
+      data->bytes_remaining = 4;
+    } else if (c >= 0xfc && c < 0xfe) {
+      if (data->bytes_remaining) {
+        cp[(*cpi)++] = UNICODE_INVALID;
+      }
+
+      data->this_cp = c & 0x01;
+      data->bytes_total = 6;
+      data->bytes_remaining = 5;
+    } else {
+      cp[(*cpi)++] = UNICODE_INVALID;
+    }
+  }
+}
+
+static VTermEncoding encoding_utf8 = {
+  .init = &init_utf8,
+  .decode = &decode_utf8,
+};
+
+static void decode_usascii(VTermEncoding *enc, void *data, uint32_t cp[], int *cpi, int cplen,
+                           const char bytes[], size_t *pos, size_t bytelen)
+{
+  int is_gr = bytes[*pos] & 0x80;
+
+  for (; *pos < bytelen && *cpi < cplen; (*pos)++) {
+    uint8_t c = (uint8_t)(bytes[*pos] ^ is_gr);
+
+    if (c < 0x20 || c == 0x7f || c >= 0x80) {
+      return;
+    }
+
+    cp[(*cpi)++] = c;
+  }
+}
+
+static VTermEncoding encoding_usascii = {
+  .decode = &decode_usascii,
+};
+
+struct StaticTableEncoding {
+  const VTermEncoding enc;
+  const uint32_t chars[128];
+};
+
+static void decode_table(VTermEncoding *enc, void *data, uint32_t cp[], int *cpi, int cplen,
+                         const char bytes[], size_t *pos, size_t bytelen)
+{
+  struct StaticTableEncoding *table = (struct StaticTableEncoding *)enc;
+  int is_gr = bytes[*pos] & 0x80;
+
+  for (; *pos < bytelen && *cpi < cplen; (*pos)++) {
+    uint8_t c = (uint8_t)(bytes[*pos] ^ is_gr);
+
+    if (c < 0x20 || c == 0x7f || c >= 0x80) {
+      return;
+    }
+
+    if (table->chars[c]) {
+      cp[(*cpi)++] = table->chars[c];
+    } else {
+      cp[(*cpi)++] = c;
+    }
+  }
+}
+
+// https://en.wikipedia.org/wiki/DEC_Special_Graphics
+static const struct StaticTableEncoding encoding_DECdrawing = {
+  { .decode = &decode_table },
+  {
+    [0x60] = 0x25C6,  // BLACK DIAMOND
+    [0x61] = 0x2592,  // MEDIUM SHADE (checkerboard)
+    [0x62] = 0x2409,  // SYMBOL FOR HORIZONTAL TAB
+    [0x63] = 0x240C,  // SYMBOL FOR FORM FEED
+    [0x64] = 0x240D,  // SYMBOL FOR CARRIAGE RETURN
+    [0x65] = 0x240A,  // SYMBOL FOR LINE FEED
+    [0x66] = 0x00B0,  // DEGREE SIGN
+    [0x67] = 0x00B1,  // PLUS-MINUS SIGN (plus or minus)
+    [0x68] = 0x2424,  // SYMBOL FOR NEW LINE
+    [0x69] = 0x240B,  // SYMBOL FOR VERTICAL TAB
+    [0x6a] = 0x2518,  // BOX DRAWINGS LIGHT UP AND LEFT (bottom-right corner)
+    [0x6b] = 0x2510,  // BOX DRAWINGS LIGHT DOWN AND LEFT (top-right corner)
+    [0x6c] = 0x250C,  // BOX DRAWINGS LIGHT DOWN AND RIGHT (top-left corner)
+    [0x6d] = 0x2514,  // BOX DRAWINGS LIGHT UP AND RIGHT (bottom-left corner)
+    [0x6e] = 0x253C,  // BOX DRAWINGS LIGHT VERTICAL AND HORIZONTAL (crossing lines)
+    [0x6f] = 0x23BA,  // HORIZONTAL SCAN LINE-1
+    [0x70] = 0x23BB,  // HORIZONTAL SCAN LINE-3
+    [0x71] = 0x2500,  // BOX DRAWINGS LIGHT HORIZONTAL
+    [0x72] = 0x23BC,  // HORIZONTAL SCAN LINE-7
+    [0x73] = 0x23BD,  // HORIZONTAL SCAN LINE-9
+    [0x74] = 0x251C,  // BOX DRAWINGS LIGHT VERTICAL AND RIGHT
+    [0x75] = 0x2524,  // BOX DRAWINGS LIGHT VERTICAL AND LEFT
+    [0x76] = 0x2534,  // BOX DRAWINGS LIGHT UP AND HORIZONTAL
+    [0x77] = 0x252C,  // BOX DRAWINGS LIGHT DOWN AND HORIZONTAL
+    [0x78] = 0x2502,  // BOX DRAWINGS LIGHT VERTICAL
+    [0x79] = 0x2A7D,  // LESS-THAN OR SLANTED EQUAL-TO
+    [0x7a] = 0x2A7E,  // GREATER-THAN OR SLANTED EQUAL-TO
+    [0x7b] = 0x03C0,  // GREEK SMALL LETTER PI
+    [0x7c] = 0x2260,  // NOT EQUAL TO
+    [0x7d] = 0x00A3,  // POUND SIGN
+    [0x7e] = 0x00B7,  // MIDDLE DOT
+  }
+};
+
+static struct {
+  VTermEncodingType type;
+  char designation;
+  VTermEncoding *enc;
+}
+encodings[] = {
+  { ENC_UTF8,      'u', &encoding_utf8 },
+  { ENC_SINGLE_94, '0', (VTermEncoding *)&encoding_DECdrawing },
+  { ENC_SINGLE_94, 'B', &encoding_usascii },
+  { 0 },
+};
+
+VTermEncoding *vterm_lookup_encoding(VTermEncodingType type, char designation)
+{
+  for (int i = 0; encodings[i].designation; i++) {
+    if (encodings[i].type == type && encodings[i].designation == designation) {
+      return encodings[i].enc;
+    }
+  }
+  return NULL;
+}
diff --git a/src/nvim/vterm/encoding.h b/src/nvim/vterm/encoding.h
new file mode 100644
index 0000000000..204b6d90c9
--- /dev/null
+++ b/src/nvim/vterm/encoding.h
@@ -0,0 +1,10 @@
+#pragma once
+
+#include <stddef.h>
+
+#include "nvim/vterm/vterm_defs.h"
+#include "nvim/vterm/vterm_internal_defs.h"
+
+#ifdef INCLUDE_GENERATED_DECLARATIONS
+# include "vterm/encoding.h.generated.h"
+#endif
diff --git a/src/nvim/vterm/keyboard.c b/src/nvim/vterm/keyboard.c
new file mode 100644
index 0000000000..dd088ac40e
--- /dev/null
+++ b/src/nvim/vterm/keyboard.c
@@ -0,0 +1,318 @@
+#include <stdio.h>
+
+#include "nvim/ascii_defs.h"
+#include "nvim/tui/termkey/termkey.h"
+#include "nvim/vterm/keyboard.h"
+#include "nvim/vterm/vterm.h"
+#include "nvim/vterm/vterm_internal_defs.h"
+
+#ifdef INCLUDE_GENERATED_DECLARATIONS
+# include "vterm/keyboard.c.generated.h"
+#endif
+
+static VTermKeyEncodingFlags vterm_state_get_key_encoding_flags(const VTermState *state)
+{
+  int screen = state->mode.alt_screen ? BUFIDX_ALTSCREEN : BUFIDX_PRIMARY;
+  const struct VTermKeyEncodingStack *stack = &state->key_encoding_stacks[screen];
+  assert(stack->size > 0);
+  return stack->items[stack->size - 1];
+}
+
+void vterm_keyboard_unichar(VTerm *vt, uint32_t c, VTermModifier mod)
+{
+  bool passthru = false;
+  if (c == ' ') {
+    // Space is passed through only when there are no modifiers (including shift)
+    passthru = mod == VTERM_MOD_NONE;
+  } else {
+    // Otherwise pass through when there are no modifiers (ignoring shift)
+    passthru = (mod & (unsigned)~VTERM_MOD_SHIFT) == 0;
+  }
+
+  if (passthru) {
+    char str[6];
+    int seqlen = fill_utf8((int)c, str);
+    vterm_push_output_bytes(vt, str, (size_t)seqlen);
+    return;
+  }
+
+  VTermKeyEncodingFlags flags = vterm_state_get_key_encoding_flags(vt->state);
+  if (flags.disambiguate) {
+    // Always use unshifted codepoint
+    if (c >= 'A' && c <= 'Z') {
+      c += 'a' - 'A';
+      mod |= VTERM_MOD_SHIFT;
+    }
+
+    vterm_push_output_sprintf_ctrl(vt, C1_CSI, "%d;%du", c, mod + 1);
+    return;
+  }
+
+  if (mod & VTERM_MOD_CTRL) {
+    // Handle special cases. These are taken from kitty, but seem mostly
+    // consistent across terminals.
+    switch (c) {
+    case '2':
+    case ' ':
+      // Ctrl+2 is NUL to match Ctrl+@ (which is Shift+2 on US keyboards)
+      // Ctrl+Space is also NUL for some reason
+      c = 0x00;
+      break;
+    case '3':
+    case '4':
+    case '5':
+    case '6':
+    case '7':
+      // Ctrl+3 through Ctrl+7 are sequential starting from 0x1b. Importantly,
+      // this means that Ctrl+6 emits 0x1e (the same as Ctrl+^ on US keyboards)
+      c = 0x1b + c - '3';
+      break;
+    case '8':
+      // Ctrl+8 is DEL
+      c = 0x7f;
+      break;
+    case '/':
+      // Ctrl+/ is equivalent to Ctrl+_ for historic reasons
+      c = 0x1f;
+      break;
+    default:
+      if (c >= '@' && c <= 0x7f) {
+        c &= 0x1f;
+      }
+      break;
+    }
+  }
+
+  vterm_push_output_sprintf(vt, "%s%c", mod & VTERM_MOD_ALT ? ESC_S : "", c);
+}
+
+typedef struct {
+  enum {
+    KEYCODE_NONE,
+    KEYCODE_LITERAL,
+    KEYCODE_TAB,
+    KEYCODE_ENTER,
+    KEYCODE_SS3,
+    KEYCODE_CSI,
+    KEYCODE_CSI_CURSOR,
+    KEYCODE_CSINUM,
+    KEYCODE_KEYPAD,
+  } type;
+  int literal;
+  int csinum;
+} keycodes_s;
+
+static keycodes_s keycodes[] = {
+  { KEYCODE_NONE, NUL, 0 },  // NONE
+
+  { KEYCODE_ENTER,   '\r',   0 },  // ENTER
+  { KEYCODE_TAB,     '\t',   0 },  // TAB
+  { KEYCODE_LITERAL, '\x7f', 0 },  // BACKSPACE == ASCII DEL
+  { KEYCODE_LITERAL, '\x1b', 0 },  // ESCAPE
+
+  { KEYCODE_CSI_CURSOR, 'A', 0 },  // UP
+  { KEYCODE_CSI_CURSOR, 'B', 0 },  // DOWN
+  { KEYCODE_CSI_CURSOR, 'D', 0 },  // LEFT
+  { KEYCODE_CSI_CURSOR, 'C', 0 },  // RIGHT
+
+  { KEYCODE_CSINUM, '~', 2     },  // INS
+  { KEYCODE_CSINUM, '~', 3     },  // DEL
+  { KEYCODE_CSI_CURSOR, 'H', 0 },  // HOME
+  { KEYCODE_CSI_CURSOR, 'F', 0 },  // END
+  { KEYCODE_CSINUM, '~', 5     },  // PAGEUP
+  { KEYCODE_CSINUM, '~', 6     },  // PAGEDOWN
+};
+
+static keycodes_s keycodes_fn[] = {
+  { KEYCODE_NONE,   NUL,  0 },    // F0 - shouldn't happen
+  { KEYCODE_SS3,    'P',  0 },    // F1
+  { KEYCODE_SS3,    'Q',  0 },    // F2
+  { KEYCODE_SS3,    'R',  0 },    // F3
+  { KEYCODE_SS3,    'S',  0 },    // F4
+  { KEYCODE_CSINUM, '~',  15 },   // F5
+  { KEYCODE_CSINUM, '~',  17 },   // F6
+  { KEYCODE_CSINUM, '~',  18 },   // F7
+  { KEYCODE_CSINUM, '~',  19 },   // F8
+  { KEYCODE_CSINUM, '~',  20 },   // F9
+  { KEYCODE_CSINUM, '~',  21 },   // F10
+  { KEYCODE_CSINUM, '~',  23 },   // F11
+  { KEYCODE_CSINUM, '~',  24 },   // F12
+};
+
+static keycodes_s keycodes_kp[] = {
+  { KEYCODE_KEYPAD, '0',  'p' },  // KP_0
+  { KEYCODE_KEYPAD, '1',  'q' },  // KP_1
+  { KEYCODE_KEYPAD, '2',  'r' },  // KP_2
+  { KEYCODE_KEYPAD, '3',  's' },  // KP_3
+  { KEYCODE_KEYPAD, '4',  't' },  // KP_4
+  { KEYCODE_KEYPAD, '5',  'u' },  // KP_5
+  { KEYCODE_KEYPAD, '6',  'v' },  // KP_6
+  { KEYCODE_KEYPAD, '7',  'w' },  // KP_7
+  { KEYCODE_KEYPAD, '8',  'x' },  // KP_8
+  { KEYCODE_KEYPAD, '9',  'y' },  // KP_9
+  { KEYCODE_KEYPAD, '*',  'j' },  // KP_MULT
+  { KEYCODE_KEYPAD, '+',  'k' },  // KP_PLUS
+  { KEYCODE_KEYPAD, ',',  'l' },  // KP_COMMA
+  { KEYCODE_KEYPAD, '-',  'm' },  // KP_MINUS
+  { KEYCODE_KEYPAD, '.',  'n' },  // KP_PERIOD
+  { KEYCODE_KEYPAD, '/',  'o' },  // KP_DIVIDE
+  { KEYCODE_KEYPAD, '\n', 'M' },  // KP_ENTER
+  { KEYCODE_KEYPAD, '=',  'X' },  // KP_EQUAL
+};
+
+static keycodes_s keycodes_kp_csiu[] = {
+  { KEYCODE_KEYPAD, 57399, 'p' },  // KP_0
+  { KEYCODE_KEYPAD, 57400, 'q' },  // KP_1
+  { KEYCODE_KEYPAD, 57401, 'r' },  // KP_2
+  { KEYCODE_KEYPAD, 57402, 's' },  // KP_3
+  { KEYCODE_KEYPAD, 57403, 't' },  // KP_4
+  { KEYCODE_KEYPAD, 57404, 'u' },  // KP_5
+  { KEYCODE_KEYPAD, 57405, 'v' },  // KP_6
+  { KEYCODE_KEYPAD, 57406, 'w' },  // KP_7
+  { KEYCODE_KEYPAD, 57407, 'x' },  // KP_8
+  { KEYCODE_KEYPAD, 57408, 'y' },  // KP_9
+  { KEYCODE_KEYPAD, 57411, 'j' },  // KP_MULT
+  { KEYCODE_KEYPAD, 57413, 'k' },  // KP_PLUS
+  { KEYCODE_KEYPAD, 57416, 'l' },  // KP_COMMA
+  { KEYCODE_KEYPAD, 57412, 'm' },  // KP_MINUS
+  { KEYCODE_KEYPAD, 57409, 'n' },  // KP_PERIOD
+  { KEYCODE_KEYPAD, 57410, 'o' },  // KP_DIVIDE
+  { KEYCODE_KEYPAD, 57414, 'M' },  // KP_ENTER
+  { KEYCODE_KEYPAD, 57415, 'X' },  // KP_EQUAL
+};
+
+void vterm_keyboard_key(VTerm *vt, VTermKey key, VTermModifier mod)
+{
+  if (key == VTERM_KEY_NONE) {
+    return;
+  }
+
+  VTermKeyEncodingFlags flags = vterm_state_get_key_encoding_flags(vt->state);
+
+  keycodes_s k;
+  if (key < VTERM_KEY_FUNCTION_0) {
+    if (key >= sizeof(keycodes)/sizeof(keycodes[0])) {
+      return;
+    }
+    k = keycodes[key];
+  } else if (key >= VTERM_KEY_FUNCTION_0 && key <= VTERM_KEY_FUNCTION_MAX) {
+    if ((key - VTERM_KEY_FUNCTION_0) >= sizeof(keycodes_fn)/sizeof(keycodes_fn[0])) {
+      return;
+    }
+    k = keycodes_fn[key - VTERM_KEY_FUNCTION_0];
+  } else if (key >= VTERM_KEY_KP_0) {
+    if ((key - VTERM_KEY_KP_0) >= sizeof(keycodes_kp)/sizeof(keycodes_kp[0])) {
+      return;
+    }
+
+    if (flags.disambiguate) {
+      k = keycodes_kp_csiu[key - VTERM_KEY_KP_0];
+    } else {
+      k = keycodes_kp[key - VTERM_KEY_KP_0];
+    }
+  }
+
+  switch (k.type) {
+  case KEYCODE_NONE:
+    break;
+
+  case KEYCODE_TAB:
+    // Shift-Tab is CSI Z but plain Tab is 0x09
+    if (flags.disambiguate) {
+      goto case_LITERAL;
+    } else if (mod == VTERM_MOD_SHIFT) {
+      vterm_push_output_sprintf_ctrl(vt, C1_CSI, "Z");
+    } else if (mod & VTERM_MOD_SHIFT) {
+      vterm_push_output_sprintf_ctrl(vt, C1_CSI, "1;%dZ", mod + 1);
+    } else {
+      goto case_LITERAL;
+    }
+    break;
+
+  case KEYCODE_ENTER:
+    // Enter is CRLF in newline mode, but just LF in linefeed
+    if (vt->state->mode.newline) {
+      vterm_push_output_sprintf(vt, "\r\n");
+    } else {
+      goto case_LITERAL;
+    }
+    break;
+
+  case KEYCODE_LITERAL:
+                        case_LITERAL:
+    if (flags.disambiguate) {
+      switch (key) {
+      case VTERM_KEY_TAB:
+      case VTERM_KEY_ENTER:
+      case VTERM_KEY_BACKSPACE:
+        // If there are no mods then leave these as-is
+        flags.disambiguate = mod != VTERM_MOD_NONE;
+        break;
+      default:
+        break;
+      }
+    }
+
+    if (flags.disambiguate) {
+      vterm_push_output_sprintf_ctrl(vt, C1_CSI, "%d;%du", k.literal, mod + 1);
+    } else {
+      vterm_push_output_sprintf(vt, mod & VTERM_MOD_ALT ? ESC_S "%c" : "%c", k.literal);
+    }
+    break;
+
+  case KEYCODE_SS3:
+                    case_SS3:
+    if (mod == 0) {
+      vterm_push_output_sprintf_ctrl(vt, C1_SS3, "%c", k.literal);
+    } else {
+      goto case_CSI;
+    }
+    break;
+
+  case KEYCODE_CSI:
+                    case_CSI:
+    if (mod == 0) {
+      vterm_push_output_sprintf_ctrl(vt, C1_CSI, "%c", k.literal);
+    } else {
+      vterm_push_output_sprintf_ctrl(vt, C1_CSI, "1;%d%c", mod + 1, k.literal);
+    }
+    break;
+
+  case KEYCODE_CSINUM:
+    if (mod == 0) {
+      vterm_push_output_sprintf_ctrl(vt, C1_CSI, "%d%c", k.csinum, k.literal);
+    } else {
+      vterm_push_output_sprintf_ctrl(vt, C1_CSI, "%d;%d%c", k.csinum, mod + 1, k.literal);
+    }
+    break;
+
+  case KEYCODE_CSI_CURSOR:
+    if (vt->state->mode.cursor) {
+      goto case_SS3;
+    } else {
+      goto case_CSI;
+    }
+
+  case KEYCODE_KEYPAD:
+    if (vt->state->mode.keypad) {
+      k.literal = k.csinum;
+      goto case_SS3;
+    } else {
+      goto case_LITERAL;
+    }
+  }
+}
+
+void vterm_keyboard_start_paste(VTerm *vt)
+{
+  if (vt->state->mode.bracketpaste) {
+    vterm_push_output_sprintf_ctrl(vt, C1_CSI, "200~");
+  }
+}
+
+void vterm_keyboard_end_paste(VTerm *vt)
+{
+  if (vt->state->mode.bracketpaste) {
+    vterm_push_output_sprintf_ctrl(vt, C1_CSI, "201~");
+  }
+}
diff --git a/src/nvim/vterm/keyboard.h b/src/nvim/vterm/keyboard.h
new file mode 100644
index 0000000000..af5f4a3ed1
--- /dev/null
+++ b/src/nvim/vterm/keyboard.h
@@ -0,0 +1,10 @@
+#pragma once
+
+#include <stddef.h>
+
+#include "nvim/vterm/vterm_defs.h"
+#include "nvim/vterm/vterm_keycodes_defs.h"
+
+#ifdef INCLUDE_GENERATED_DECLARATIONS
+# include "vterm/keyboard.h.generated.h"
+#endif
diff --git a/src/nvim/vterm/mouse.c b/src/nvim/vterm/mouse.c
new file mode 100644
index 0000000000..2f3b1d9e2f
--- /dev/null
+++ b/src/nvim/vterm/mouse.c
@@ -0,0 +1,124 @@
+#include "nvim/math.h"
+#include "nvim/tui/termkey/termkey.h"
+#include "nvim/vterm/mouse.h"
+#include "nvim/vterm/vterm.h"
+#include "nvim/vterm/vterm_internal_defs.h"
+
+#ifdef INCLUDE_GENERATED_DECLARATIONS
+# include "vterm/mouse.c.generated.h"
+#endif
+
+static void output_mouse(VTermState *state, int code, int pressed, int modifiers, int col, int row)
+{
+  modifiers <<= 2;
+
+  switch (state->mouse_protocol) {
+  case MOUSE_X10:
+    if (col + 0x21 > 0xff) {
+      col = 0xff - 0x21;
+    }
+    if (row + 0x21 > 0xff) {
+      row = 0xff - 0x21;
+    }
+
+    if (!pressed) {
+      code = 3;
+    }
+
+    if (code & 0x80) {
+      break;
+    }
+    vterm_push_output_sprintf_ctrl(state->vt, C1_CSI, "M%c%c%c",
+                                   (code | modifiers) + 0x20, col + 0x21, row + 0x21);
+    break;
+
+  case MOUSE_UTF8: {
+    char utf8[18];
+    size_t len = 0;
+
+    if (!pressed) {
+      code = 3;
+    }
+
+    len += (size_t)fill_utf8((code | modifiers) + 0x20, utf8 + len);
+    len += (size_t)fill_utf8(col + 0x21, utf8 + len);
+    len += (size_t)fill_utf8(row + 0x21, utf8 + len);
+    utf8[len] = 0;
+
+    vterm_push_output_sprintf_ctrl(state->vt, C1_CSI, "M%s", utf8);
+  }
+  break;
+
+  case MOUSE_SGR:
+    vterm_push_output_sprintf_ctrl(state->vt, C1_CSI, "<%d;%d;%d%c",
+                                   code | modifiers, col + 1, row + 1, pressed ? 'M' : 'm');
+    break;
+
+  case MOUSE_RXVT:
+    if (!pressed) {
+      code = 3;
+    }
+
+    vterm_push_output_sprintf_ctrl(state->vt, C1_CSI, "%d;%d;%dM",
+                                   code | modifiers, col + 1, row + 1);
+    break;
+  }
+}
+
+void vterm_mouse_move(VTerm *vt, int row, int col, VTermModifier mod)
+{
+  VTermState *state = vt->state;
+
+  if (col == state->mouse_col && row == state->mouse_row) {
+    return;
+  }
+
+  state->mouse_col = col;
+  state->mouse_row = row;
+
+  if ((state->mouse_flags & MOUSE_WANT_DRAG && state->mouse_buttons)
+      || (state->mouse_flags & MOUSE_WANT_MOVE)) {
+    if (state->mouse_buttons) {
+      int button = xctz((uint64_t)state->mouse_buttons) + 1;
+      if (button < 4) {
+        output_mouse(state, button - 1 + 0x20, 1, (int)mod, col, row);
+      } else if (button >= 8 && button < 12) {
+        output_mouse(state, button - 8 + 0x80 + 0x20, 1, (int)mod, col, row);
+      }
+    } else {
+      output_mouse(state, 3 + 0x20, 1, (int)mod, col, row);
+    }
+  }
+}
+
+void vterm_mouse_button(VTerm *vt, int button, bool pressed, VTermModifier mod)
+{
+  VTermState *state = vt->state;
+
+  int old_buttons = state->mouse_buttons;
+
+  if ((button > 0 && button <= 3) || (button >= 8 && button <= 11)) {
+    if (pressed) {
+      state->mouse_buttons |= (1 << (button - 1));
+    } else {
+      state->mouse_buttons &= ~(1 << (button - 1));
+    }
+  }
+
+  // Most of the time we don't get button releases from 4/5/6/7
+  if (state->mouse_buttons == old_buttons && (button < 4 || button > 7)) {
+    return;
+  }
+
+  if (!state->mouse_flags) {
+    return;
+  }
+
+  if (button < 4) {
+    output_mouse(state, button - 1, pressed, (int)mod, state->mouse_col, state->mouse_row);
+  } else if (button < 8) {
+    output_mouse(state, button - 4 + 0x40, pressed, (int)mod, state->mouse_col, state->mouse_row);
+  } else if (button < 12) {
+    output_mouse(state, button - 8 + 0x80, pressed, (int)mod, state->mouse_col, state->mouse_row);
+  }
+}
diff --git a/src/nvim/vterm/mouse.h b/src/nvim/vterm/mouse.h
new file mode 100644
index 0000000000..477f4028a2
--- /dev/null
+++ b/src/nvim/vterm/mouse.h
@@ -0,0 +1,10 @@
+#pragma once
+
+#include <stddef.h>
+
+#include "nvim/vterm/vterm_defs.h"
+#include "nvim/vterm/vterm_keycodes_defs.h"
+
+#ifdef INCLUDE_GENERATED_DECLARATIONS
+# include "vterm/mouse.h.generated.h"
+#endif
diff --git a/src/nvim/vterm/parser.c b/src/nvim/vterm/parser.c
new file mode 100644
index 0000000000..79d348f2c1
--- /dev/null
+++ b/src/nvim/vterm/parser.c
@@ -0,0 +1,411 @@
+#include <assert.h>
+#include <stdio.h>
+#include <string.h>
+
+#include "nvim/vterm/parser.h"
+#include "nvim/vterm/vterm.h"
+#include "nvim/vterm/vterm_internal_defs.h"
+
+#ifdef INCLUDE_GENERATED_DECLARATIONS
+# include "vterm/parser.c.generated.h"
+#endif
+
+#undef DEBUG_PARSER
+
+static bool is_intermed(uint8_t c)
+{
+  return c >= 0x20 && c <= 0x2f;
+}
+
+static void do_control(VTerm *vt, uint8_t control)
+{
+  if (vt->parser.callbacks && vt->parser.callbacks->control) {
+    if ((*vt->parser.callbacks->control)(control, vt->parser.cbdata)) {
+      return;
+    }
+  }
+
+  DEBUG_LOG("libvterm: Unhandled control 0x%02x\n", control);
+}
+
+static void do_csi(VTerm *vt, char command)
+{
+#ifdef DEBUG_PARSER
+  printf("Parsed CSI args as:\n", arglen, args);
+  printf(" leader: %s\n", vt->parser.v.csi.leader);
+  for (int argi = 0; argi < vt->parser.v.csi.argi; argi++) {
+    printf(" %lu", CSI_ARG(vt->parser.v.csi.args[argi]));
+    if (!CSI_ARG_HAS_MORE(vt->parser.v.csi.args[argi])) {
+      printf("\n");
+    }
+    printf(" intermed: %s\n", vt->parser.intermed);
+  }
+#endif
+
+  if (vt->parser.callbacks && vt->parser.callbacks->csi) {
+    if ((*vt->parser.callbacks->csi)(vt->parser.v.csi.leaderlen ? vt->parser.v.csi.leader : NULL,
+                                     vt->parser.v.csi.args,
+                                     vt->parser.v.csi.argi,
+                                     vt->parser.intermedlen ? vt->parser.intermed : NULL,
+                                     command,
+                                     vt->parser.cbdata)) {
+      return;
+    }
+  }
+
+  DEBUG_LOG("libvterm: Unhandled CSI %c\n", command);
+}
+
+static void do_escape(VTerm *vt, char command)
+{
+  char seq[INTERMED_MAX + 1];
+
+  size_t len = (size_t)vt->parser.intermedlen;
+  strncpy(seq, vt->parser.intermed, len);  // NOLINT(runtime/printf)
+  seq[len++] = command;
+  seq[len] = 0;
+
+  if (vt->parser.callbacks && vt->parser.callbacks->escape) {
+    if ((*vt->parser.callbacks->escape)(seq, len, vt->parser.cbdata)) {
+      return;
+    }
+  }
+
+  DEBUG_LOG("libvterm: Unhandled escape ESC 0x%02x\n", command);
+}
+
+static void string_fragment(VTerm *vt, const char *str, size_t len, bool final)
+{
+  VTermStringFragment frag = {
+    .str = str,
+    .len = len,
+    .initial = vt->parser.string_initial,
+    .final = final,
+  };
+
+  switch (vt->parser.state) {
+  case OSC:
+    if (vt->parser.callbacks && vt->parser.callbacks->osc) {
+      (*vt->parser.callbacks->osc)(vt->parser.v.osc.command, frag, vt->parser.cbdata);
+    }
+    break;
+
+  case DCS_VTERM:
+    if (vt->parser.callbacks && vt->parser.callbacks->dcs) {
+      (*vt->parser.callbacks->dcs)(vt->parser.v.dcs.command, (size_t)vt->parser.v.dcs.commandlen,
+                                   frag,
+                                   vt->parser.cbdata);
+    }
+    break;
+
+  case APC:
+    if (vt->parser.callbacks && vt->parser.callbacks->apc) {
+      (*vt->parser.callbacks->apc)(frag, vt->parser.cbdata);
+    }
+    break;
+
+  case PM:
+    if (vt->parser.callbacks && vt->parser.callbacks->pm) {
+      (*vt->parser.callbacks->pm)(frag, vt->parser.cbdata);
+    }
+    break;
+
+  case SOS:
+    if (vt->parser.callbacks && vt->parser.callbacks->sos) {
+      (*vt->parser.callbacks->sos)(frag, vt->parser.cbdata);
+    }
+    break;
+
+  case NORMAL:
+  case CSI_LEADER:
+  case CSI_ARGS:
+  case CSI_INTERMED:
+  case OSC_COMMAND:
+  case DCS_COMMAND:
+    break;
+  }
+
+  vt->parser.string_initial = false;
+}
+
+size_t vterm_input_write(VTerm *vt, const char *bytes, size_t len)
+{
+  size_t pos = 0;
+  const char *string_start;
+
+  switch (vt->parser.state) {
+  case NORMAL:
+  case CSI_LEADER:
+  case CSI_ARGS:
+  case CSI_INTERMED:
+  case OSC_COMMAND:
+  case DCS_COMMAND:
+    string_start = NULL;
+    break;
+  case OSC:
+  case DCS_VTERM:
+  case APC:
+  case PM:
+  case SOS:
+    string_start = bytes;
+    break;