PATCH: support OSC52 paste sequence

[Oliver Kiddle <opk@xxxxxxx>]

We already have support for copying to the system clipboard via the
OSC52 sequence. Some terminals also allow the system clipboard or
primary selection to be retrieved (with \e]52;p;?\e\\). You may find
that that's disabled by default for security/privacy reasons but can be
enabled. Mostly, Shift-Insert is more convenient but I sometimes miss it
when using vi-style editing.

This patch adds support for this with the "* and "+ vim registers. Much
of the patch involves moving some of the existing copy code to the new
termquery.c source file.

Oliver

diff --git a/Doc/Zsh/zle.yo b/Doc/Zsh/zle.yo
index 31eb3f3ba..316d232c3 100644
--- a/Doc/Zsh/zle.yo
+++ b/Doc/Zsh/zle.yo
@@ -2483,9 +2483,8 @@ appended to the buffer instead of overwriting it. When using the tt("_)
 buffer, nothing happens. This can be useful for deleting text without
 affecting any buffers.
 
-Updating the system clipboard relies on specific support from the terminal.
-Reading it is not possible so a paste command with tt("*) or tt("+) will do
-nothing.
+Accessing and updating the system clipboard relies on specific support from
+the terminal.
 
 If no buffer is specified for a cut or change command, tt("1) is used, and
 the contents of tt("1) to tt("8) are each shifted along one buffer;
diff --git a/Src/Zle/termquery.c b/Src/Zle/termquery.c
index ccd6eaef0..1232a96bf 100644
--- a/Src/Zle/termquery.c
+++ b/Src/Zle/termquery.c
@@ -79,6 +79,7 @@ typedef const unsigned char seqstate_t;
 #define TINFO "\xc3"
 #define XTID  "\xc4"
 #define XTVER "\xc5"
+#define CLIP  "\xc6"
 
 /* Deterministic finite state automata for parsing terminal response sequences.
  * - alternatives decided by the first character (no back-tracking)
@@ -122,6 +123,12 @@ typedef const unsigned char seqstate_t;
 	    OR MATCH("u", KITTY ) \
 	    OR MATCH("c", DA) )))
 
+#define OSC52_STATES \
+    "\033]52;" EITHER("p" OR "c") ";" RECORD \
+    REPEAT( \
+	CAPTURE EITHER( "\033" MATCH("\\", CLIP) OR MATCH("\007", CLIP) ) \
+    OR WILDCARD )
+
 static char *EXTVAR  = ".term.extensions";
 static char *IDVAR   = ".term.id";
 static char *VERVAR  = ".term.version";
@@ -191,7 +198,7 @@ find_matching(seqstate_t* pos, int direction)
 static void
 probe_terminal(const char *tquery, seqstate_t *states,
 	void (*handle_seq) (int seq, int *numbers, int len,
-	    char *capture, int clen))
+	    char *capture, int clen, void *output), void *output)
 {
     size_t blen = 256, nlen = 16;
     char *buf = zhalloc(blen);
@@ -201,11 +208,12 @@ probe_terminal(const char *tquery, seqstate_t *states,
     int *num = numbers;
     int finish = 0, number = 0;
     int ch;
-    struct ttyinfo ti;
+    struct ttyinfo ti, torig;
 
     seqstate_t *curstate = states;
 
     gettyinfo(&ti);
+    memcpy(&torig, &ti, sizeof(torig));
 #ifdef HAS_TIO
     ti.tio.c_lflag &= (~ECHO & ~ICANON & ~ISIG);
     ti.tio.c_iflag &= ~ICRNL;
@@ -389,7 +397,7 @@ probe_terminal(const char *tquery, seqstate_t *states,
 	} else {
 	    if (sequence && !(finish = sequence == SEQ))
 		handle_seq(sequence & ~SEQ, numbers, num - numbers,
-			buf + record, capture - record);
+			buf + record, capture - record, output);
 
 	    if ((sequence || (action & 1)) &&
 		(current = start) && /* drop input from sequence */
@@ -409,7 +417,7 @@ probe_terminal(const char *tquery, seqstate_t *states,
     if (current > buf)
 	ungetbytes(buf, current - buf);
 
-    settyinfo(&shttyinfo);
+    settyinfo(&torig);
 }
 
 static void
@@ -440,7 +448,8 @@ static const char *queries[] =
 	{ TQ_BGCOLOR, TQ_FGCOLOR, TQ_KITTYKB, TQ_RGB, TQ_XTVERSION, TQ_DA };
 
 static void
-handle_query(int sequence, int *numbers, int len, char *capture, int clen)
+handle_query(int sequence, int *numbers, int len, char *capture, int clen,
+	UNUSED(void *output))
 {
     char **feat;
 
@@ -499,12 +508,85 @@ query_terminal(void) {
 		((cterm = getsparam("COLORTERM")) &&
 		    (!strcmp(cterm, "truecolor") ||
 			!strcmp(cterm, "24bit")))))
-	    handle_query(3, 0, 0, 0, 0);
+	    handle_query(3, NULL, 0, NULL, 0, NULL);
 	else
 	    struncpy(&tqend, (char *) queries[i], /* collate escape sequences */
 		sizeof(tquery) - (tqend - tquery));
     }
 
     if (tqend != tquery) /* unless nothing left after filtering */
-	probe_terminal(tquery, states, &handle_query);
+	probe_terminal(tquery, states, &handle_query, NULL);
+}
+
+static char*
+base64_encode(const char *src, size_t len) {
+    static const char* base64_table =
+	"ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789+/";
+
+    const unsigned char *end = (unsigned char *)src + len;
+    const unsigned char *in = (unsigned char *)src;
+    char *ret = zhalloc(1 + 4 * ((len + 2) / 3)); /* 4 bytes out for 3 in */
+    char *cur = ret;
+
+    for (; end - in > 0; in += 3, cur += 4) {
+	unsigned int n = *in << 16;
+	cur[3] = end - in > 2 ? base64_table[(n |= in[2]) & 0x3f] : '=';
+	cur[2] = end - in > 1 ? base64_table[((n |= in[1]<<8) >> 6) & 0x3f] : '=';
+	cur[1] = base64_table[(n >> 12) & 0x3f];
+	cur[0] = base64_table[n >> 18];
+    }
+    *cur = '\0';
+
+    return ret;
+}
+
+static char*
+base64_decode(const char *src, size_t len)
+{
+    int i = 0;
+    unsigned int n;
+    char *buf = hcalloc((3 * len) / 4 + 1);
+    char *b = buf;
+    char c;
+
+    while (len && (c = src[i]) != '=') {
+	n = isdigit(c) ? c - '0' + 52 :
+	    islower(c) ? c - 'a' + 26 :
+	    isupper(c) ? c - 'A' :
+	    (c == '+') ? 62 :
+	    (c == '/') ? 63 : 0;
+	if (i % 4)
+	    *b++ |= n >> (2 * (3 - (i % 4)));
+	if (++i >= len)
+	    break;
+	*b = n << (2 * (i % 4));
+    }
+    return buf;
+}
+
+static void
+handle_paste(UNUSED(int sequence), UNUSED(int *numbers), UNUSED(int len),
+	char *capture, int clen, void *output)
+{
+    *(char**) output = base64_decode(capture, clen);
+}
+
+/**/
+char *
+system_clipget(char clip)
+{
+    static seqstate_t osc52[] = OSC52_STATES;
+    char seq[] = "\033]52;.;?\033\\";
+    char *contents;
+    seq[5] = clip;
+    probe_terminal(seq, osc52, &handle_paste, &contents);
+    return contents;
+}
+
+/**/
+void
+system_clipput(char clip, char *content, size_t clen)
+{
+    char *encoded = base64_encode(content, clen);
+    fprintf(shout, "\033]52;%c;%s\a", clip, encoded);
 }
diff --git a/Src/Zle/zle_misc.c b/Src/Zle/zle_misc.c
index e17a08d53..a2d07ee03 100644
--- a/Src/Zle/zle_misc.c
+++ b/Src/Zle/zle_misc.c
@@ -554,7 +554,8 @@ yank(UNUSED(char **args))
 }
 
 /* position: 0 is before, 1 after, 2 split the line */
-static void pastebuf(Cutbuffer buf, int mult, int position)
+static void
+pastebuf(Cutbuffer buf, int mult, int position)
 {
     int cc;
     if (buf->flags & CUTBUFFER_LINE) {
@@ -613,7 +614,20 @@ viputbefore(UNUSED(char **args))
 	return 1;
     if (zmod.flags & MOD_NULL)
 	return 0;
-    if (zmod.flags & MOD_VIBUF)
+    if (zmod.flags & MOD_OSSEL) {
+	struct cutbuffer kbuf;
+	int x;
+	char *pbuf = system_clipget(zmod.flags & MOD_CLIP ? 'c' : 'p');
+	if (!pbuf || !*pbuf)
+	    return 1;
+	kbuf.buf = stringaszleline(pbuf, 0, &x, NULL, NULL);
+	kbuf.len = x;
+	kbuf.flags = 0;
+	kct = -1;
+	yankcs = zlecs;
+	pastebuf(&kbuf, n, 0);
+	return 0;
+    } else if (zmod.flags & MOD_VIBUF)
 	kctbuf = &vibuf[zmod.vibuf];
     else
 	kctbuf = &cutbuf;
@@ -636,7 +650,20 @@ viputafter(UNUSED(char **args))
 	return 1;
     if (zmod.flags & MOD_NULL)
 	return 0;
-    if (zmod.flags & MOD_VIBUF)
+    if (zmod.flags & MOD_OSSEL) {
+	struct cutbuffer kbuf;
+	int x;
+	char *pbuf = system_clipget(zmod.flags & MOD_CLIP ? 'c' : 'p');
+	if (!pbuf || !*pbuf)
+	    return 1;
+	kbuf.buf = stringaszleline(pbuf, 0, &x, NULL, NULL);
+	kbuf.len = x;
+	kbuf.flags = 0;
+	kct = -1;
+	yankcs = zlecs;
+	pastebuf(&kbuf, n, 1);
+	return 0;
+    } else if (zmod.flags & MOD_VIBUF)
 	kctbuf = &vibuf[zmod.vibuf];
     else
 	kctbuf = &cutbuf;
@@ -661,15 +688,25 @@ putreplaceselection(UNUSED(char **args))
     startvichange(-1);
     if (n < 0 || zmod.flags & MOD_NULL)
 	return 1;
-    putbuf = (zmod.flags & MOD_VIBUF) ? &vibuf[zmod.vibuf] : &cutbuf;
-    if (!putbuf->buf)
-	return 1;
-    memcpy(&prevbuf, putbuf, sizeof(prevbuf));
+    if (zmod.flags & MOD_OSSEL) {
+        int x;
+        char *pbuf = system_clipget(zmod.flags & MOD_CLIP ? 'c' : 'p');
+        if (!pbuf || !*pbuf)
+            return 1;
+        prevbuf.buf = stringaszleline(pbuf, 0, &x, NULL, NULL);
+        prevbuf.len = x;
+        prevbuf.flags = 0;
+    } else {
+	putbuf = (zmod.flags & MOD_VIBUF) ? &vibuf[zmod.vibuf] : &cutbuf;
+	if (!putbuf->buf)
+	    return 1;
+	memcpy(&prevbuf, putbuf, sizeof(prevbuf));
 
-    /* if "9 was specified, prevent killregion from freeing it */
-    if (zmod.vibuf == 35) {
-	putbuf->buf = 0;
-	clear = 1;
+	/* if "9 was specified, prevent killregion from freeing it */
+	if (zmod.vibuf == 35) {
+	    putbuf->buf = 0;
+	    clear = 1;
+	}
     }
 
     zmod.flags = 0; /* flags apply to paste not kill */
diff --git a/Src/Zle/zle_utils.c b/Src/Zle/zle_utils.c
index e2b86e863..6e2456b1f 100644
--- a/Src/Zle/zle_utils.c
+++ b/Src/Zle/zle_utils.c
@@ -937,28 +937,6 @@ cut(int i, int ct, int flags)
   cuttext(zleline + i, ct, flags);
 }
 
-static char*
-base64_encode(const char *src, size_t len) {
-    static const char* base64_table =
-        "ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789+/";
-
-    const unsigned char *end = (unsigned char *)src + len;
-    const unsigned char *in = (unsigned char *)src;
-    char *ret = zhalloc(1 + 4 * ((len + 2) / 3)); /* 4 bytes out for 3 in */
-    char *cur = ret;
-
-    for (; end - in > 0; in += 3, cur += 4) {
-        unsigned int n = *in << 16;
-        cur[3] = end - in > 2 ? base64_table[(n |= in[2]) & 0x3f] : '=';
-        cur[2] = end - in > 1 ? base64_table[((n |= in[1]<<8) >> 6) & 0x3f] : '=';
-        cur[1] = base64_table[(n >> 12) & 0x3f];
-        cur[0] = base64_table[n >> 18];
-    }
-    *cur = '\0';
-
-    return ret;
-}
-
 /*
  * As cut, but explicitly supply the text together with its length.
  */
@@ -975,10 +953,7 @@ cuttext(ZLE_STRING_T line, int ct, int flags)
 	int cutll;
 	char *mbcut = zlelineasstring(line, ct, 0, &cutll, NULL, 1);
 	unmetafy(mbcut, &cutll);
-	mbcut = base64_encode(mbcut, cutll);
-
-	fprintf(shout, "\033]52;%c;%s\a", zmod.flags & MOD_CLIP ? 'c' : 'p',
-		mbcut);
+	system_clipput(zmod.flags & MOD_CLIP ? 'c' : 'p', mbcut, cutll);
     } else if (zmod.flags & MOD_VIBUF) {
 	struct cutbuffer *b = &vibuf[zmod.vibuf];
 
diff --git a/Test/X06termquery.ztst b/Test/X06termquery.ztst
index 52f39ddf1..cce4e626f 100644
--- a/Test/X06termquery.ztst
+++ b/Test/X06termquery.ztst
@@ -84,6 +84,20 @@
   termresp ''
 0:no response - timeout
 
+# Following three vi-put tests also cover 0, 1 and 2 `=' padding
+# characters in the base64 decoding.
+  termresp $'\e[?0;cbindkey -v\necho \e"*p\e]52;p;YWZ0ZXI=\aa\n'
+0:paste after from clipboard
+>after
+
+  termresp $'\e[?0;cbindkey -v\necho X\e"*P\e]52;p;YmVmb3Jl\aa\n'
+0:paste before from clipboard
+>beforeX
+
+  termresp $'\e[?0;cbindkey -v\necho X\ev"*p\e]52;p;cmVwbGFjZQ==\aa\n'
+0:paste over from clipboard
+>replace
+
 %clean
 
   zmodload -ui zsh/zpty