Re: recent terminfo changes

[christos%astron.com@localhost (Christos Zoulas)]

In article <20200324144303.E147610756%quasar.astron.com@localhost>,
Christos Zoulas <christos%quasar.astron.com@localhost> wrote:
>Hello,
>
>The recent terminfo ABI change to widen numeric constants from 16 bits to
>32 bits was done in the following way:
>
>1. Create a new terminfo2 database and rely on the fact that old
>   programs will use the old filename and new ones will use the new
>   filename.
>2. Add compatibility code so that the new library can read the old
>   terminfo database.
>
>This approach has issues. Aside from the visible effect to have a
>new file, it causes us to have to carry an old terminfo file around
>to keep old binaries working because we don't have a way to build
>the old terminfo file anymore.
>
>With a little more effort all this can be avoided:
>
>We can modify the writing code to only write entries in the new
>format for the terminal descriptions that require it. We can store
>those descriptions under a different name <name>@v2 for example,
>while storing a "compatible" description with the original name
>with the values that don't fit in 16 bits clamped. The code can
>try to find <name>@v2 first and if that fails, attempt to find the
>original name. This way we can store all the data in the same file
>(no need to create terminfo2 anymore), and have both old programs
>keep working, and new programs be able to use the wide fields.
>

Below is a patch that does most of the work described above. It
creates type 1 and type 3 records in the same file as needed. Old
code will not be able to read the type 3 records, which is just a
handful of them... I suggest that we apply that now, and keep using
/usr/share/misc/terminfo.cdb as before.

Index: compile.c
===================================================================
RCS file: /cvsroot/src/lib/libterminfo/compile.c,v
retrieving revision 1.14
diff -u -u -r1.14 compile.c
--- compile.c	13 Mar 2020 15:19:25 -0000	1.14
+++ compile.c	26 Mar 2020 23:48:02 -0000
@@ -51,6 +51,9 @@
 #include <term_private.h>
 #include <term.h>
 
+#define _ti_numsize(tic) \
+    ((tic)->rtype == TERMINFO_RTYPE_O1 ? sizeof(uint16_t) : sizeof(uint32_t))
+
 static void __printflike(2, 3)
 dowarn(int flags, const char *fmt, ...)
 {
@@ -87,7 +90,7 @@
 }
 
 char *
-_ti_find_cap(TBUF *tbuf, char type, short ind)
+_ti_find_cap(TIC *tic, TBUF *tbuf, char type, short ind)
 {
 	size_t n;
 	uint16_t num;
@@ -106,7 +109,7 @@
 			cap++;
 			break;
 		case 'n':
-			cap += sizeof(uint32_t);
+			cap += _ti_numsize(tic);
 			break;
 		case 's':
 			num = le16dec(cap);
@@ -121,7 +124,7 @@
 }
 
 char *
-_ti_find_extra(TBUF *tbuf, const char *code)
+_ti_find_extra(TIC *tic, TBUF *tbuf, const char *code)
 {
 	size_t n;
 	uint16_t num;
@@ -142,7 +145,7 @@
 			cap++;
 			break;
 		case 'n':
-			cap += sizeof(uint32_t);
+			cap += _ti_numsize(tic);
 			break;
 		case 's':
 			num = le16dec(cap);
@@ -156,6 +159,16 @@
 	return NULL;
 }
 
+static void
+_ti_encode_num(TIC *tic, TBUF *rbuf, int num) {
+	if (_ti_numsize(tic) == sizeof(uint16_t)) {
+		le16enc(rbuf->buf + rbuf->bufpos, (uint16_t)num);
+	} else {
+		le32enc(rbuf->buf + rbuf->bufpos, (uint32_t)num);
+	}
+	rbuf->bufpos += _ti_numsize(tic);
+}
+
 size_t
 _ti_store_extra(TIC *tic, int wrn, char *id, char type, char flag, int num,
     char *str, size_t strl, int flags)
@@ -165,7 +178,7 @@
 	_DIAGASSERT(tic != NULL);
 
 	if (strcmp(id, "use") != 0) {
-		if (_ti_find_extra(&tic->extras, id) != NULL)
+		if (_ti_find_extra(tic, &tic->extras, id) != NULL)
 			return 0;
 		if (!(flags & TIC_EXTRA)) {
 			if (wrn != 0)
@@ -182,7 +195,7 @@
 	}
 
 	if (!_ti_grow_tbuf(&tic->extras,
-		l + strl + sizeof(uint16_t) + sizeof(uint32_t) + 1))
+		l + strl + sizeof(uint16_t) + _ti_numsize(tic) + 1))
 		return 0;
 	le16enc(tic->extras.buf + tic->extras.bufpos, (uint16_t)l);
 	tic->extras.bufpos += sizeof(uint16_t);
@@ -194,8 +207,7 @@
 		tic->extras.buf[tic->extras.bufpos++] = flag;
 		break;
 	case 'n':
-		le32enc(tic->extras.buf + tic->extras.bufpos, (uint32_t)num);
-		tic->extras.bufpos += sizeof(uint32_t);
+		_ti_encode_num(tic, &tic->extras, num);
 		break;
 	case 's':
 		le16enc(tic->extras.buf + tic->extras.bufpos, (uint16_t)strl);
@@ -239,7 +251,7 @@
 		return -1;
 
 	cap = *buf;
-	*cap++ = TERMINFO_RTYPE; /* Record type 3 */
+	*cap++ = tic->rtype;
 	le16enc(cap, (uint16_t)len);
 	cap += sizeof(uint16_t);
 	memcpy(cap, tic->name, len);
@@ -451,6 +463,17 @@
 	return token;
 }
 
+static int
+_ti_find_rtype(const char *cap)
+{
+	for (const char *ptr = cap; (ptr = strchr(ptr, '#')) != NULL;) {
+		if (strtol(++ptr, NULL, 0) > 0xffff) {
+			return TERMINFO_RTYPE;
+		}
+	}
+	return TERMINFO_RTYPE_O1;
+}
+
 TIC *
 _ti_compile(char *cap, int flags)
 {
@@ -467,7 +490,7 @@
 
 	name = _ti_get_token(&cap, ',');
 	if (name == NULL) {
-		dowarn(flags, "no seperator found: %s", cap);
+		dowarn(flags, "no separator found: %s", cap);
 		return NULL;
 	}
 	desc = strrchr(name, '|');
@@ -494,6 +517,7 @@
 	if (tic == NULL)
 		return NULL;
 
+	tic->rtype = _ti_find_rtype(cap);
 	buf.buf = NULL;
 	buf.buflen = 0;
 
@@ -533,7 +557,7 @@
 			/* Don't use the string if we already have it */
 			ind = (short)_ti_strindex(token);
 			if (ind != -1 &&
-			    _ti_find_cap(&tic->strs, 's', ind) != NULL)
+			    _ti_find_cap(tic, &tic->strs, 's', ind) != NULL)
 				continue;
 
 			/* Encode the string to our scratch buffer */
@@ -579,7 +603,7 @@
 			/* Don't use the number if we already have it */
 			ind = (short)_ti_numindex(token);
 			if (ind != -1 &&
-			    _ti_find_cap(&tic->nums, 'n', ind) != NULL)
+			    _ti_find_cap(tic, &tic->nums, 'n', ind) != NULL)
 				continue;
 
 			cnum = strtol(p, &e, 0);
@@ -600,14 +624,12 @@
 				    num, NULL, 0, flags);
 			else {
 				if (_ti_grow_tbuf(&tic->nums,
-				    sizeof(uint16_t) + sizeof(uint32_t))==NULL)
+				    sizeof(uint16_t) + _ti_numsize(tic))==NULL)
 					goto error;
 				le16enc(tic->nums.buf + tic->nums.bufpos,
 				    (uint16_t)ind);
 				tic->nums.bufpos += sizeof(uint16_t);
-				le32enc(tic->nums.buf + tic->nums.bufpos,
-				    (uint32_t)num);
-				tic->nums.bufpos += sizeof(uint32_t);
+				_ti_encode_num(tic, &tic->nums, num);
 				tic->nums.entries++;
 			}
 			continue;
@@ -622,7 +644,8 @@
 		ind = (short)_ti_flagindex(token);
 		if (ind == -1 && flag == CANCELLED_BOOLEAN) {
 			if ((ind = (short)_ti_numindex(token)) != -1) {
-				if (_ti_find_cap(&tic->nums, 'n', ind) != NULL)
+				if (_ti_find_cap(tic, &tic->nums, 'n', ind)
+				    != NULL)
 					continue;
 				if (_ti_grow_tbuf(&tic->nums,
 					sizeof(uint16_t) * 2) == NULL)
@@ -630,13 +653,13 @@
 				le16enc(tic->nums.buf + tic->nums.bufpos,
 				    (uint16_t)ind);
 				tic->nums.bufpos += sizeof(uint16_t);
-				le32enc(tic->nums.buf + tic->nums.bufpos,
-				    (uint32_t)CANCELLED_NUMERIC);
-				tic->nums.bufpos += sizeof(uint32_t);
+				_ti_encode_num(tic, &tic->nums,
+				    CANCELLED_NUMERIC); 
 				tic->nums.entries++;
 				continue;
 			} else if ((ind = (short)_ti_strindex(token)) != -1) {
-				if (_ti_find_cap(&tic->strs, 's', ind) != NULL)
+				if (_ti_find_cap(tic, &tic->strs, 's', ind)
+				    != NULL)
 					continue;
 				if (_ti_grow_tbuf(&tic->strs,
 				    (sizeof(uint16_t) * 2) + 1) == NULL)
@@ -652,7 +675,7 @@
 		if (ind == -1)
 			_ti_store_extra(tic, 1, token, 'f', flag, 0, NULL, 0,
 			    flags);
-		else if (_ti_find_cap(&tic->flags, 'f', ind) == NULL) {
+		else if (_ti_find_cap(tic, &tic->flags, 'f', ind) == NULL) {
 			if (_ti_grow_tbuf(&tic->flags, sizeof(uint16_t) + 1)
 			    == NULL)
 				goto error;
Index: term.c
===================================================================
RCS file: /cvsroot/src/lib/libterminfo/term.c,v
retrieving revision 1.30
diff -u -u -r1.30 term.c
--- term.c	13 Mar 2020 15:19:25 -0000	1.30
+++ term.c	26 Mar 2020 23:48:02 -0000
@@ -74,6 +74,21 @@
 }
 
 static int
+_ti_decode_num(int rtype, const char **cap)
+{
+	int rv;
+
+	if (rtype == TERMINFO_RTYPE_O1) {
+		rv = (int)le16dec(*cap);
+		*cap += sizeof(uint16_t);
+	} else {
+		rv = (int)le32dec(*cap);
+		*cap += sizeof(uint32_t);
+	}
+	return rv;
+}
+
+static int
 _ti_readterm(TERMINAL *term, const char *cap, size_t caplen, int flags)
 {
 	char rtype;
@@ -150,13 +165,7 @@
 		for (; num != 0; num--) {
 			ind = le16dec(cap);
 			cap += sizeof(uint16_t);
-			if (rtype == TERMINFO_RTYPE_O1) {
-				term->nums[ind] = (int)le16dec(cap);
-				cap += sizeof(uint16_t);
-			} else {
-				term->nums[ind] = (int)le32dec(cap);
-				cap += sizeof(uint32_t);
-			}
+			term->nums[ind] = _ti_decode_num(rtype, &cap);
 			if (flags == 0 && !VALID_NUMERIC(term->nums[ind]))
 				term->nums[ind] = ABSENT_NUMERIC;
 		}
@@ -213,13 +222,7 @@
 				break;
 			case 'n':
 				ud->flag = ABSENT_BOOLEAN;
-				if (rtype == TERMINFO_RTYPE_O1) {
-					ud->num = (int)le16dec(cap);
-					cap += sizeof(uint16_t);
-				} else {
-					ud->num = (int)le32dec(cap);
-					cap += sizeof(uint32_t);
-				}
+				ud->num = _ti_decode_num(rtype, &cap);
 				if (flags == 0 &&
 				    !VALID_NUMERIC(ud->num))
 					ud->num = ABSENT_NUMERIC;
Index: term_private.h
===================================================================
RCS file: /cvsroot/src/lib/libterminfo/term_private.h,v
retrieving revision 1.12
diff -u -u -r1.12 term_private.h
--- term_private.h	13 Mar 2020 15:19:25 -0000	1.12
+++ term_private.h	26 Mar 2020 23:48:02 -0000
@@ -154,6 +154,7 @@
 	char *name;
 	char *alias;
 	char *desc;
+	int  rtype;
 	TBUF flags;
 	TBUF nums;
 	TBUF strs;
@@ -162,8 +163,8 @@
 
 char *_ti_grow_tbuf(TBUF *, size_t);
 char *_ti_get_token(char **, char);
-char *_ti_find_cap(TBUF *, char,  short);
-char *_ti_find_extra(TBUF *, const char *);
+char *_ti_find_cap(TIC *, TBUF *, char,  short);
+char *_ti_find_extra(TIC *, TBUF *, const char *);
 size_t _ti_store_extra(TIC *, int, char *, char, char, int,
     char *, size_t, int);
 TIC *_ti_compile(char *, int);