Merge remote-tracking branch 'origin/front' into front

lib/ktrans/README

@@ -0,0 +1,12 @@
+#
+# The following are a set of tools to obtain and process dictionaries from the SKK project in order to use them with ktrans(1).
+#
+# grabskkdicts pulls the skk kana-kanji conversion dictionaries from the skk-dev/dict repo.
+# skk2ktrans takes an skk dictionary and converts it into a kanji jisho suitable to be used with ktrans(1).
+#
+# You can fetch and convert all the dictionaries by running this file.
+#
+
+grabskkdicts
+for(d in skkdicts/SKK-JISYO.*)
+	<$d skk2ktrans >$d.jisho

lib/ktrans/grabskkdicts

@@ -0,0 +1,2 @@
+#!/bin/rc
+git/clone https://github.com/skk-dev/dict skkdicts

lib/ktrans/skk2ktrans

@@ -0,0 +1,2 @@
+#!/bin/rc
+tcs -sf jis | awk '$1 !~ /;;/ {gsub("(^\/|\/$)", "", $2); gsub(" ", "	"); gsub("\/", " ", $2);} {print}'

lib/theo

@@ -956,3 +956,7 @@ I really don't understand the approach being taken here.
 I want to understand what the goal here is.
 It won't happen.
 That is either a funny joke, or it shows you don't understand Unix.
+The smallest diffs are the best.
+I don't understand what the proposal is.
+The man page says "don't do that".
+That seems disgustingly verbose to me.

rc/bin/inst/bootplan9

@@ -22,6 +22,7 @@ if(~ $#p9part 0){
 	exit ''
 }
 
+if(test -f /386/mbr){
 	echo
 	echo 'If you use the Windows NT/2000/XP master boot record'
 	echo 'or a master boot record from a Unix clone (e.g., LILO or'
@@ -33,6 +34,7 @@ switch($rd) {
 	case yes
 		disk/mbr -m /386/mbr /dev/$disk/data
 	}
+}
 
 prompt 'Mark the Plan 9 partition active' yes no
 switch($rd) {

rc/bin/inst/bootsetup

@@ -30,16 +30,26 @@ case go
 	9fat=$rd
 	export 9fat
 
+	if(~ $#bootfile 1){
 		bootfile=`{basename $bootfile}
+	}
+	if(! ~ $#bootfile 1 || ! test -f /$cputype/$bootfile) {
+		bootfile=()
+	}
 
 	if(! test -f /tmp/plan9.ini) {
 		@{
+			if(~ $#bootfile 1)
 				echo 'bootfile='^$bootfile
+			if(~ $#fs 1)
 				echo 'bootargs=local!'^$fs' '$"fsflags
 			if(~ $#nvram 1)
 				echo 'nvram='^$nvram
+			if(~ $#mouseport 1)
 				echo 'mouseport='^$mouseport
+			if(~ $#monitor 1)
 				echo 'monitor='^$monitor
+			if(~ $#vgasize 1)
 				echo 'vgasize='^$vgasize
 			cd '#ec'
 			for(i in *){
@@ -72,32 +82,50 @@ case go
 	# always make backup of old bootsector
 	logprog dd -bs 512 -count 1 -if $9fat -of /tmp/pbs.bak
 
+	if (! test -f /386/pbs) {
 		if(~ $need9fatformat yes){
 			log Initializing Plan 9 FAT partition.
-		logprog disk/format -r 2 -d -b /n/newfs/386/pbs -l PLAN9 $9fat
+			logprog disk/format -r 2 -d -l PLAN9 $9fat
+		}
+	}
+	if not {
+		if(~ $need9fatformat yes){
+			log Initializing Plan 9 FAT partition.
+			logprog disk/format -r 2 -d -b /386/pbs -l PLAN9 $9fat
 		}
 		if not {
 			log Updating bootsector.
-		logprog disk/format -b /n/newfs/386/pbs $9fat
+			logprog disk/format -b /386/pbs $9fat
+		}
 	}
 
 	logprog mount -c /srv/dos /n/9fat $9fat
 
-	logprog rm -f /n/9fat/^(9bootfat plan9.ini $bootfile)
+	if(test -f /386/9bootfat) {
+		logprog rm -f /n/9fat/9bootfat
+		logprog cp /386/9bootfat /n/9fat/9bootfat
 
-	logprog cp /n/newfs/386/9bootfat /n/9fat/9bootfat
 		# make file continous on disk
 		logprog chmod +al /n/9fat/9bootfat
+	}
 
 	# copy config
+	logprog rm -f /n/9fat/plan9.ini
 	logprog cp /tmp/plan9.ini /n/9fat/plan9.ini
 
 	# copy kernel
-	logprog cp /n/newfs/$cputype/$bootfile /n/9fat/
+	if(~ $#bootfile 1){
+		logprog rm -f /n/9fat/$bootfile
+		logprog cp /$cputype/$bootfile /n/9fat/$bootfile
+	}
 
 	# copy efi bootloader
+	for(i in /386/*.efi){
+		if(test -f $i) {
 			logprog mkdir -p /n/9fat/efi/boot
-	logprog cp /386/^(bootia32.efi bootx64.efi) /n/9fat/efi/boot
+			logprog cp $i /n/9fat/efi/boot/
+		}
+	}
 
 	# copy in backups
 	if(test -f /tmp/oldplan9.ini)

sys/games/lib/fortunes

@@ -5260,3 +5260,5 @@ MY NAME IS BEING USED IN A PHISHING ATTACK.  DO NOT RESPOND TO MAIL OFFERING MON
 Xterm.js
 GPU.js
 Joys of Porting a Kernel to WebAssembly
+Never change anything. You can only make it worse. -- Ted Unangst
+Use a client written in Go? -- Skip Tavakkolian

sys/man/3/i2c

@@ -0,0 +1,98 @@
+.TH I2C 3
+.SH NAME
+i2c \- I²C Bus Interface
+.SH SYNOPSIS
+.nf
+.B bind -a #J /dev
+.PP
+.nf
+.BI /dev/ bus /i2c. n .ctl
+.BI /dev/ bus /i2c. n .data
+.SH DESCRIPTION
+The I²C (Inter-Integrated Circuit) bus is a serial bus standard
+invented by Philips.
+It is used to connect chips over short-distances like on a motherboard
+or between boards.
+The bus requires only a single data signal 
+.IR SDA ,
+a clock signal
+.I SDC
+and ground.
+Devices on the bus have 7-bit or 10-bit addresses, which is
+determined by the device (see the datasheet).
+Every device monitors the bus and waits for the master to initiate
+a transaction.
+Each transaction starts with the master sending the target device address
+with a direction bit and the device acknowledges each byte
+received if its address matched.
+Depending on the device, more bytes must be transmitted after the
+.IR device-address .
+For example, encoding a register number at a memory offset.
+We refer to these bytes as the
+.IR sub-address .
+After the address phase, the data phase begins where
+zero or more data bytes are either transmitted or received
+depending on the transaction direction.
+.PP
+This driver supports multiple buses (controllers) to be registered
+by platform specific code in the kernel.
+Each bus is represented as a directory with a platform specific name.
+.PP
+Because all received bytes are acknowledged by target devices,
+it is possible for the master to enumerate all existing addresses
+on a bus, which is done when accessing the
+.I bus
+directory.
+.PP
+Platform specific code can also register individual devices,
+supplying a meaningful device configuration.
+It can also be necessary to manually register devices if
+the device has been disabled by an enable signal or
+otherwise inhibited during enumeration.
+.PP
+Each device appears as a control (
+.BI i2c. n .ctl
+)
+and a data file (
+.BI i2c. n .data
+) in the
+.I bus
+directory
+where
+.I n
+is the device-address in hexadecimal.
+.PP
+The data file is used to initiate transactions to the device.
+The initial
+.I sub-address
+bytes sent are derived from the file position according to the device
+configuration, and the length of the data phase is controlled by
+the requested read or write size.
+The requested size might be truncated by the driver if
+the request exceeds the logical file-size.
+.PP
+The control-file provides access to the device configuration.
+When read,
+it reproduces this configuration in text form.
+The following text lines are accepted when written to the control-file:
+.TF "subaddress n"
+.TP
+.BI "subaddress " n
+Sets the length
+.IR n ,
+in bytes, of the sub-address for the device
+that is derived from the file-position of the data-file.
+This can be set to zero if no sub-address bytes should
+be generated by the driver.
+By default, the sub-address length is 1.
+.TP
+.BI "size " n
+Changes the logical size of the data-file to
+.IR n .
+Making sure the file-position is not exceeding this size.
+.SH SOURCE
+.B /sys/src/9/port/devi2c.c
+.SH HISTORY
+The file-interface has been inspired by infernos devi2c,
+but with the addition of supporting multiple buses
+and device enumeration.

sys/man/3/ip

@@ -682,33 +682,43 @@ Otherwise a free port number starting at 5000 is chosen.
 The connect fails if the combination of local and remote address/port pairs
 are already assigned to another port.
 .TP
-.BI announce\  X
+.BI announce\  [ip-address ! ]port
-.I X
+Set the local IP address and port number and accept calls there.  If
-is a decimal port number or
+ip-address is left out, accept calls on any address.  If port is 0, a
-.LR * .
+port is automatically choosen that is not yet announced.
-Set the local port
+If the address is
-number to
-.I X
-and accept calls to
-.IR X .
-If
-.I X
-is
 .LR * ,
-accept
+accept calls on any address.
-calls for any port that no process has explicitly announced.
+If port is
-The local IP address cannot be set.
+.LR * ,
+accept calls on any port.
+If port is
+.LR * ,
+and the address is left out, accept calls on any address and port.
 .B Announce
-fails if the connection is already announced or connected.
+fails if the connection is already announced.
 .TP
-.BI bind\  X
+.BI bind\  [ip-address ! ]port
-.I X
+Set the local IP address and port number like for a server connection
-is a decimal port number or
+similar to the
-.LR * .
+.B announce
-Set the local port number to
+command.
-.IR X .
+If ip-address is left out, an address is automatically selected.  If
-This exists to support emulation
+port is 0, a port is automatically choosen that is not yet announced.
-of BSD sockets by the APE libraries (see
+This command has no actual effect, beyond remembering the parameters
+and possibly selecting an unused port.
+The commands
+.B announce
+and
+.B connect
+reset both the local address and the port according to their own
+parameters.
+This command also does
+.B not
+reserve the IP address and port, another connection can use them, even
+while they are registered in a connection by this command.
+This exists to support emulation of BSD sockets by the APE libraries
+(see
 .IR pcc (1))
 and is not otherwise used.
 .\" this is gone

sys/src/9/bcm/pi

@@ -13,6 +13,7 @@ dev
 	tls
 	cap
 	fs
+	bridge	log
 	ip	arp chandial ip ipv6 ipaux iproute netlog nullmedium pktmedium inferno
 	draw	screen swcursor
 	mouse	mouse
@@ -22,10 +23,11 @@ dev
 	usb
 
 link
-	loopbackmedium
-	ethermedium
 	archbcm
 	usbdwc
+	ethermedium
+	loopbackmedium
+	netdevmedium
 
 ip
 	tcp

sys/src/9/bcm/pi2

@@ -13,6 +13,7 @@ dev
 	tls
 	cap
 	fs
+	bridge	log
 	ip	arp chandial ip ipv6 ipaux iproute netlog nullmedium pktmedium inferno
 	draw	screen swcursor
 	mouse	mouse
@@ -22,10 +23,11 @@ dev
 	usb
 
 link
-	loopbackmedium
-	ethermedium
 	archbcm2
 	usbdwc
+	ethermedium
+	loopbackmedium
+	netdevmedium
 
 ip
 	tcp

sys/src/9/bcm64/pi3

@@ -13,6 +13,7 @@ dev
 	tls
 	cap
 	fs
+	bridge	log
 	ip	arp chandial ip ipv6 ipaux iproute netlog nullmedium pktmedium inferno
 	draw	screen swcursor
 	mouse	mouse
@@ -23,10 +24,11 @@ dev
 	dtracy
 
 link
-	loopbackmedium
-	ethermedium
 	archbcm3
 	usbdwc
+	ethermedium
+	loopbackmedium 
+	netdevmedium
 
 ip
 	tcp

sys/src/9/bcm64/pi4

@@ -14,6 +14,7 @@ dev
 	cap
 	fs
 	ether	netif
+	bridge	log
 	ip	arp chandial ip ipv6 ipaux iproute netlog nullmedium pktmedium inferno
 	draw	screen swcursor
 	mouse	mouse
@@ -31,6 +32,7 @@ link
 	ethergenet	ethermii
 	ethermedium
 	loopbackmedium
+	netdevmedium
 
 ip
 	tcp

sys/src/9/imx8/reform

@@ -13,6 +13,7 @@ dev
 	cap
 	fs
 	ether	netif
+	bridge	log
 	ip	arp chandial ip ipv6 ipaux iproute netlog nullmedium pktmedium inferno
 	draw	screen swcursor
 	mouse	screen swcursor
@@ -26,8 +27,10 @@ link
 	etherimx	ethermii
 	ethermedium
 	loopbackmedium
+	netdevmedium
 	i2cimx		devi2c
 	pciimx		pci
+
 ip
 	tcp
 	udp

sys/src/9/pc/pc

@@ -20,7 +20,7 @@ dev
 	dtracy
 
 	ether		netif
-	bridge		netif log
+	bridge		log
 	ip		arp chandial ip ipv6 ipaux iproute netlog nullmedium pktmedium inferno
 
 	draw		screen vga vgax vgasoft
@@ -81,9 +81,9 @@ link
 	ethervirtio	pci
 	ethervirtio10	pci
 	ethermedium
-	pcmciamodem
-	netdevmedium
 	loopbackmedium
+	netdevmedium
+	pcmciamodem
 	usbuhci		pci
 	usbohci		pci
 	usbehci		usbehcipc

sys/src/9/pc64/pc64

@@ -19,8 +19,8 @@ dev
 	fs
 
 	ether		netif
-	bridge		netif log
+	bridge		log
-	ip		arp chandial ip ipv6 ipaux iproute netlog ethermedium nullmedium pktmedium inferno
+	ip		arp chandial ip ipv6 ipaux iproute netlog nullmedium pktmedium inferno
 
 	draw		screen vga vgax vgasoft
 	mouse		mouse
@@ -79,9 +79,9 @@ link
 	ethervirtio	pci
 	ethervirtio10	pci
 	ethermedium
-#	pcmciamodem
-	netdevmedium
 	loopbackmedium
+	netdevmedium
+#	pcmciamodem
 	usbuhci		pci
 	usbohci		pci
 	usbehci		pci usbehcipc

sys/src/9/zynq/zynq

@@ -12,6 +12,7 @@ dev
 	pipe
 	dup
 	ether	netif
+	bridge	log
 	ip	arp chandial ip ipv6 ipaux iproute netlog nullmedium pktmedium inferno
 	tls
 	cap

sys/src/ape/lib/bsd/_sock_ingetaddr.c

@@ -39,6 +39,30 @@ _sock_inport(struct sockaddr *a)
 	return 0;
 }
 
+int
+_sock_inisany(int af, void *addr)
+{
+	int alen;
+	void *any;
+	/* an IPv4 address that is auto-initialized to all zeros */
+	static struct in_addr inaddr_any;
+
+	switch(af){
+	case AF_INET:
+		alen = sizeof inaddr_any.s_addr;
+		any = &inaddr_any;
+		break;
+	case AF_INET6:
+		alen = sizeof in6addr_any;
+		any = &in6addr_any;
+		break;
+	default:
+		return 0;
+	}
+
+	return 0 == memcmp(addr, any, alen);
+}
+
 int
 _sock_inaddr(int af, char *ip, char *port, void *a, int *alen)
 {
@@ -97,3 +121,23 @@ _sock_ingetaddr(Rock *r, void *a, int *alen, char *file)
 		close(fd);
 	}
 }
+
+char *
+_sock_inaddr2string(Rock *r, char *dest, int dlen)
+{
+	int af = r->domain;
+	void *addr = _sock_inip(&r->addr);
+	int port = _sock_inport(&r->addr);
+	char *d = dest;
+	char *dend = dest+dlen;
+
+	if(!_sock_inisany(af, addr)){
+		inet_ntop(af, addr, d, dlen-1);
+		d = memchr(d, 0, dlen-1);
+		*(d++) = '!';
+	}
+
+	snprintf(d, dend-d, "%d", port);
+
+	return dest;
+}

sys/src/ape/lib/bsd/bind.c

@@ -24,7 +24,7 @@
 int
 bind(int fd, void *a, int alen)
 {
-	int n, len, cfd, port;
+	int n, len, cfd;
 	struct sockaddr *sa;
 	Rock *r;
 	char msg[128];
@@ -55,20 +55,25 @@ bind(int fd, void *a, int alen)
 		errno = EBADF;
 		return -1;
 	}
-	port = _sock_inport(&r->addr);
+
-	if(port > 0)
+	strcpy(msg, "bind ");
-		snprintf(msg, sizeof msg, "bind %d", port);
+	_sock_inaddr2string(r, msg + 5, sizeof msg - 5);
-	else
+
-		strcpy(msg, "bind *");
 	n = write(cfd, msg, strlen(msg));
 	if(n < 0){
-		errno = EOPNOTSUPP;	/* Improve error reporting!!! */
+		_syserrno();
+		if(errno == EPLAN9)
+			errno = EOPNOTSUPP;
 		close(cfd);
 		return -1;
 	}
+
 	close(cfd);
-	if(port <= 0)
+
+	if(_sock_inport(&r->addr) == 0)
 		_sock_ingetaddr(r, &r->addr, 0, "local");
 
 	return 0;
 }
+
+

sys/src/ape/lib/bsd/listen.c

@@ -121,7 +121,7 @@ listen(fd, backlog)
 	int backlog;
 {
 	Rock *r;
-	int n, cfd, port;
+	int n, cfd;
 	char msg[128];
 	struct sockaddr_un *lunix;
 
@@ -139,20 +139,13 @@ listen(fd, backlog)
 			errno = EBADF;
 			return -1;
 		}
-		port = _sock_inport(&r->addr);
+		strcpy(msg, "announce ");
-		if(port >= 0) {
+		_sock_inaddr2string(r, msg + 9, sizeof msg - 9);
-			if(write(cfd, "bind 0", 6) < 0) {
-				errno = EGREG;
-				close(cfd);
-				return -1;
-			}
-			snprintf(msg, sizeof msg, "announce %d", port);
-		}
-		else
-			strcpy(msg, "announce *");
 		n = write(cfd, msg, strlen(msg));
 		if(n < 0){
-			errno = EOPNOTSUPP;	/* Improve error reporting!!! */
+			_syserrno();
+			if(errno == EPLAN9)
+				errno = EOPNOTSUPP;
 			close(cfd);
 			return -1;
 		}

sys/src/ape/lib/bsd/priv.h

@@ -43,5 +43,7 @@ extern int	_sock_data(int, char*, int, int, int, Rock**);
 extern int	_sock_ipattr(char*);
 extern void*	_sock_inip(struct sockaddr*);
 extern int	_sock_inport(struct sockaddr*);
+extern int	_sock_inisany(int af, void *addr);
 extern int	_sock_inaddr(int, char*, char*, void*, int*);
 extern void	_sock_ingetaddr(Rock*, void*, int*, char*);
+extern char*	_sock_inaddr2string(Rock *r, char *dest, int dlen);

sys/src/cmd/ktrans/hash.c

@@ -0,0 +1,224 @@
+#include <u.h>
+#include <libc.h>
+#include "hash.h"
+
+typedef struct Hnode Hnode;
+struct Hnode {
+	int filled;
+	int next;
+	void *key;
+};
+
+enum{
+	Tagsize = sizeof(Hnode),
+};
+
+uvlong
+shash(char *s)
+{
+	uvlong hash;
+
+	hash = 7;
+	for(; *s; s++)
+		hash = hash*31  + *s;
+	return hash;
+}
+
+Hmap*
+hmapalloc(int nbuckets, int size)
+{
+	void *store;
+	Hmap *h;
+	int nsz;
+
+	nsz = Tagsize + size;
+	store = mallocz(sizeof(*h) + (nbuckets * nsz), 1);
+	if(store == nil)
+		return nil;
+
+	h = store;
+	h->nbs = nbuckets;
+	h->nsz = nsz;
+	h->len = h->cap = nbuckets;
+
+	h->nodes = store;
+	h->nodes += sizeof(*h);
+	return store;
+}
+
+int
+hmaprepl(Hmap **store, char *key, void *new, void *old, int freekeys)
+{
+	Hnode *n;
+	uchar *v;
+	uchar *oldv;
+	Hmap *h;
+	int next;
+	vlong diff;
+
+	h = *store;
+	oldv = nil;
+	v = h->nodes + (shash(key)%h->nbs) * h->nsz;
+	for(;;){
+		n = (Hnode*)v;
+		next = n->next;
+
+		if(n->filled == 0)
+			goto replace;
+		if(strcmp(n->key, key) == 0){
+			if(freekeys)
+				free(n->key);
+			oldv = v + Tagsize;
+			goto replace;
+		}
+		if(next == 0)
+			break;
+		v = h->nodes + next*h->nsz;
+	}
+
+	if(h->cap == h->len){
+		/* figure out way back from a relocation */
+		diff = v - h->nodes;
+
+		h->cap *= 2;
+		*store = realloc(*store, sizeof(*h) + h->cap*h->nsz);
+		h = *store;
+		h->nodes = (uchar*)*store + sizeof(*h);
+		memset(h->nodes + h->len*h->nsz, 0, h->nsz);
+
+		v = h->nodes + diff;
+		n = (Hnode*)v;
+	}
+	n->next = h->len;
+	h->len++;
+	assert(h->len <= h->cap);
+	v = h->nodes + n->next*h->nsz;
+	n = (Hnode*)v;
+
+replace:
+	memmove(v + Tagsize, new, h->nsz - Tagsize);
+	n->filled++;
+	n->key = key;
+	n->next = next;
+	if(old != nil && oldv != nil){
+		memmove(old, oldv, h->nsz - Tagsize);
+		return 1;
+	}
+	return 0;
+}
+
+int
+hmapupd(Hmap **h, char *key, void *new)
+{
+	char *prev;
+
+	prev = hmapkey(*h, key);
+	if(prev == nil)
+		prev = key;
+
+	return hmaprepl(h, prev, new, nil, 0);
+}
+
+void*
+_hmapget(Hmap *h, char *key)
+{
+	Hnode *n;
+	uchar *v;
+
+	v = h->nodes + (shash(key)%h->nbs)*h->nsz;
+	for(;;){
+		n = (Hnode*)v;
+		if(n->filled != 0 && strcmp(n->key, key) == 0)
+			return v;
+		if(n->next == 0)
+			break;
+		v = h->nodes + n->next*h->nsz;
+	}
+	return nil;
+}
+
+int
+hmapget(Hmap *h, char *key, void *dst)
+{
+	uchar *v;
+
+	v = _hmapget(h, key);
+	if(v == nil)
+		return -1;
+	if(dst != nil)
+		memmove(dst, v + Tagsize, h->nsz - Tagsize);
+	return 0;
+}
+
+int
+hmapdel(Hmap *h, char *key, void *dst, int freekey)
+{
+	uchar *v;
+	Hnode *n;
+
+	v = _hmapget(h, key);
+	if(v == nil)
+		return -1;
+
+	n = (Hnode*)v;
+	n->filled = 0;
+	if(freekey)
+		free(n->key);
+	if(dst != nil)
+		memmove(dst, v + Tagsize, h->nsz - Tagsize);
+	return 0;
+}
+
+char*
+hmapkey(Hmap *h, char *key)
+{
+	uchar *v;
+	Hnode *n;
+
+	v = _hmapget(h, key);
+	if(v == nil)
+		return nil;
+
+	n = (Hnode*)v;
+	return n->key;
+}
+
+Hmap*
+hmaprehash(Hmap *old, int buckets)
+{
+	int i;
+	uchar *v;
+	Hnode *n;
+	Hmap *new;
+
+	if(buckets == 0)
+		buckets = old->len;
+
+	new = hmapalloc(buckets, old->nsz - Tagsize);
+	for(i=0 ; i < old->len; i++){
+		v = old->nodes + i*old->nsz;
+		n = (Hnode*)v;
+		hmaprepl(&new, n->key, v + Tagsize, nil, 0);
+	}
+	free(old);
+	return new;
+}
+
+void
+hmapreset(Hmap *h, int freekeys)
+{
+	Hnode *n;
+	uchar *v;
+	int i;
+
+	for(i=0; i < h->len; i++){
+		v = h->nodes + i*h->nsz;
+		n = (Hnode*)v;
+		if(n->filled == 0)
+			continue;
+		if(freekeys)
+			free(n->key);
+		n->filled = 0;
+	}
+	h->len = 0;
+}

sys/src/cmd/ktrans/hash.h

@@ -0,0 +1,23 @@
+typedef union Hkey Hkey;
+union Hkey {
+	void *p;
+	int v;
+};
+
+typedef struct Hmap Hmap;
+struct Hmap {
+	int nbs;
+	int nsz;
+
+	int len;
+	int cap;
+	uchar *nodes;
+};
+
+Hmap*	hmapalloc(int nbuckets, int size);
+int	hmapget(Hmap *h, char *key, void *dst);
+int	hmaprepl(Hmap **h, char *key, void *new, void *old, int freekeys);
+int	hmapupd(Hmap **h, char *key, void *new);
+int	hmapdel(Hmap *h, char *key, void *dst, int freekey);
+char*	hmapkey(Hmap *h, char *key);
+void	hmapreset(Hmap *h, int freekeys);

sys/src/cmd/ktrans/jisho.c

@@ -1,211 +0,0 @@
-/* 
- *  open jisho file, and set the size of this jisho etc
- *
- *          Kenji Okamoto   August 4, 2000
- *		Osaka Prefecture Univ.
- *            okamoto@granite.cias.osakafu-u.ac.jp
- */
-
-#include <u.h>
-#include <libc.h>
-#include <bio.h>
-#include "jisho.h"
-
-Dictionary *openQDIC(char *);
-void freeQDIC(Dictionary*);
-KouhoList *getKouhoHash(Dictionary*, char *);
-KouhoList *getKouhoFile(DicList*, char *);
-void selectKouho(KouhoList **, KouhoList*);
-int hashVal(char *);
-void addHash(Hash **, DicList*);
-
-
-
-/*
- * Open QuickDIC (hashed personal dictionary)
- *   open skk styled ktrans dictionary file, and make its hash table 
- *    based on individual header kana strings
- *
- *                                        KouhoList
- *                                       |---------|
- *                  Hash         |---->kouho---->kouhotop
- *               |-------|       |
- *    dic---->dhash---->dicindex---->kanahead
- *     |--------|         |--------|
- *     Dictionary           DicList
- *
- */
-Dictionary *
-openQDIC(char *dicname)
-{
-	Biobuf *f;
-	void *Bbuf;
-	Dictionary *dic;
-	DicList *dicitem;			/* for a future extension */
-	char buf[1024], *startstr, *endstr;
-	int i;
-
-	SET(dicitem);		 /* yes, I know I'm wrong, but... */
-
-	dic = (Dictionary*)malloc(sizeof(Dictionary));
-  	      /* make room for pointer array (size=HASHSIZE) of hash table */
-	for(i=0; i< HASHSIZE; i++) dic->dhash[i] = 0;
-	dic->dlist = 0;		/* for a future extension (more than one dics ^_^ */
-
-	if ((f = Bopen(dicname, OREAD)) == 0)
-	    return dic;
-
-    /* make hash table by the dic's header word */
-
-	while(Bbuf = Brdline(f, '\n')) {
-	   strncpy(buf, (char *)Bbuf, Blinelen(f));
-
-	   if (buf[0] == ';')    /* comment line */
-		continue;
-	   else {
-    	  	/* get header word from jisho */
-	  	startstr = buf;
-	  	if(!(endstr = utfutf(startstr, "\t"))) break;
-	  	*endstr = '\0';
-			/* dicitem includes each header word from the jisho */
-
-		dicitem = (DicList*)malloc(sizeof(DicList)+(endstr-startstr+1));
-		dicitem->nextitem = 0;		/* for a future extension */
-		strcpy(dicitem->kanahead, startstr);
-
-		dicitem->kouho = getKouhoFile(dicitem, endstr);    /* read kouho from jisho */
-		addHash(dic->dhash, dicitem);
-	   }
-	   continue;
-	}
-	dic->dlist = dicitem;
-	Bterm(f);
-	return dic;
-}
-
-/* 
- * free dynamically allocated memory
- */
-void
-freeQDIC(Dictionary *dic)
-{
-   Hash *hash1, *hash2;
-   DicList *dlist, *dlist2;
-   int l;
-
-	for (dlist = dic->dlist;
-		dlist != 0;
-		dlist2 = dlist, dlist = dlist->nextitem, free((void *)dlist2));
-	for (l = 0; l < HASHSIZE; l++) {
-		for (hash1 = dic->dhash[l]; hash1; hash1 = hash2) {
-			if (hash1->next !=0) {
-				hash2 = hash1->next;
-				free((void *)hash1);
-			}else
-				break;
-		}
-	}
-	free((void *)dic);
-}
-
-int
-hashVal(char *s)
-{
-	uint h;
-
-	h = 0x811c9dc5;
-	while(*s != 0)
-		h = (h^(uchar)*s++) * 0x1000193;
-	return h % HASHSIZE;
-}
-
-void
-addHash(Hash **hash, DicList *ditem)
-{
-	Hash *h;
-	int v;
-
-	v = hashVal(ditem->kanahead);
-	h = (Hash*)malloc(sizeof(Hash));
-	h->dicindex = ditem;
-	h->length = strlen(ditem->kanahead);
-	h->next = hash[v];
-	hash[v] = h;
-}
-
-/* 
- * read Kouho list from the jisho file defined by Biobuf descriptor f
- * 
- *  revised for Plan 9 by K.Okamoto
- */
-KouhoList *
-getKouhoFile(DicList *dicitem, char * endstr)
-{
-	char *kouhostart, *kouhoend;
-	KouhoList *kouhoitem, *currntkouhoitem=0, *prevkouhoitem;
-
-	prevkouhoitem = 0;
-	kouhostart = endstr + 1;
-	while((kouhoend = utfutf(kouhostart, " ")) || 
-			(kouhoend = utfutf(kouhostart, "\n"))) {
-	   *kouhoend = '\0';
-
-	   kouhoitem = (KouhoList*)malloc(sizeof(KouhoList)+(kouhoend-kouhostart+1));
-	   kouhoitem->nextkouho = 0;
-	   kouhoitem->prevkouho = prevkouhoitem;
-	   kouhoitem->dicitem = dicitem;
-	   strcpy(kouhoitem->kouhotop, kouhostart);
-	   if (prevkouhoitem)
-		prevkouhoitem->nextkouho = kouhoitem;
-	   else
-		currntkouhoitem = kouhoitem;
-	   prevkouhoitem = kouhoitem;
-	   kouhostart = kouhoend + 1;
-	}
-	return currntkouhoitem;
-}
-
-/*
- * get matched kouho from the hash table of header word of the dict
- * if found, returns pointer to the first candidate in the hash table.
- * if not found, returns 0.
- * 
- * from getCand() in skklib.c by Akinori Ito et al.,(aito@ei5sun.yz.yamagata-u.ac.jp)
- */
-KouhoList *
-getKouhoHash(Dictionary *dic, char *s)
-{
-	int l, v;
-	Hash *h;
-
-	l = strlen(s);
-	v = hashVal(s);
-	for (h = dic->dhash[v]; h != 0; h = h->next) {
-		if (h->length != l ||
-		    strcmp(h->dicindex->kanahead, s)) continue;
-		return h->dicindex->kouho;      /* return matched kouho */
-	}
-	return 0;
-}
-
-/* 
- * from skklib.c by Akinori Ito et al.,(aito@ei5sun.yz.yamagata-u.ac.jp)
- * just modified to read easier for current purpose
- */
-void
-selectKouho(KouhoList **first, KouhoList *current)
-{
-	/* take off currentkouho from the kouholist table */
-	if (current->prevkouho) {
-		current->prevkouho->nextkouho = current->nextkouho;
-		if (current->nextkouho)
-			current->nextkouho->prevkouho = current->prevkouho;
-		current->prevkouho = 0;
-	}
-	/* take place of firstkouho by currentkouho  */
-	if (*first != current) {
-		(*first)->prevkouho = current;
-		current->nextkouho = *first;
-		*first = current;
-	}
-}

sys/src/cmd/ktrans/jisho.h

@@ -1,41 +0,0 @@
-/*
- *          Kenji Okamoto   August 4, 2000
- *		Osaka Prefecture Univ.
- *            okamoto@granite.cias.osakafu-u.ac.jp
- */
-
-#define HASHSIZE 257
-
-/*
- * Structure for Dictionary's header word (in Hiragana)
- */
-typedef	struct DicList DicList;
-struct DicList {
-	struct KouhoList *kouho;
-	struct DicList *nextitem; /* for a future extension */
-	char kanahead[1];
-};
-
-/*
- * Structure for Kouho of each index word in the dictionary
- */
-typedef	struct KouhoList KouhoList;
-struct KouhoList {
-	struct KouhoList *nextkouho;
-	struct KouhoList *prevkouho;
-	struct DicList *dicitem;
-	char kouhotop[1]; /* top of the kouhos */
-} ;
-
-typedef	struct Hash Hash;
-struct Hash {
-	DicList *dicindex; /* pointer to a KouhoList and kanahead etc */
-	short length;
-	struct Hash *next;
-};
-
-typedef	struct Dictionary Dictionary;
-struct Dictionary {
-	DicList *dlist; /* for a future extension, having more than one dictionaries */
-	Hash *dhash[HASHSIZE];
-};

sys/src/cmd/ktrans/main.c

@@ -6,16 +6,26 @@
  *                   okamoto@granite.cias.osakafu-u.ac.jp
  */
 
+/*
+ * A glossary on some of the Japanese vocabulary used:
+ * kana: syllabic letting, either hiragana(ひらがな) or katakana(カタカナ)
+ * kanji(漢字): borrowed characters, 楽 in 楽しい
+ * Okurigana(送り仮名): kana tail to kanji, しい in 楽しい
+ * Joshi(助詞): particle, は in 私は
+ * Jisho(辞書): dictionary
+ * kouho(候補): candidate
+ */
+
 #include <u.h>
 #include <libc.h>
 #include <bio.h>
+#include "hash.h"
 #include "ktrans.h"
-#include "jisho.h"
 
 #define	LSIZE	256
 
 Rune	lbuf[LSIZE];		/* hiragana buffer for key input written by send() */
-Map	*table = hira;		/* default language conversion table */
+Hmap	*table;
 uchar	okurigana[LSIZE];	/* buffer for okurigana */
 char	okuri = 0;		/* buffer/flag for capital input char */
 int	in, out;
@@ -23,16 +33,10 @@ int	llen, olen, joshi = 0;
 int	natural = 1;		/* not Japanese but English mode */
 
 int	changelang(int);
-int	dotrans(Dictionary*);
+int	dotrans(Hmap*);
 int	nrune(char *);
 void	send(uchar *, int);
-Map	*match(uchar *p, int *nc, Map *table);
+Hmap*	opendict(Hmap *, char *);
-
-extern Dictionary *openQDIC(char *);
-extern KouhoList *getKouhoHash(Dictionary*, char *);
-extern KouhoList *getKouhoFile(DicList*, char *);
-extern void freeQDIC(Dictionary*);
-extern void selectKouho(KouhoList **, KouhoList*);
 
 void
 kbdopen(void)
@@ -89,6 +93,49 @@ kbdopen(void)
 	exits(nil);
 }
 
+Map signalmore = {
+	"_", nil, 1,
+};
+
+Hmap*
+initmap(Map *m, int n)
+{
+	int i, j;
+	char buf[16];
+	char *s;
+	Map prev;
+	Hmap *h;
+
+	h = hmapalloc(n, sizeof(Map));
+	for(i = 0; i < n; i++){
+		if(m[i].roma == nil || m[i].roma[0] == '\0')
+			continue;
+
+		//We mark all partial strings so we know when
+		//we have partial match when ingesting.
+		j = 2;
+		for(s = m[i].roma; *s && j <= sizeof buf; s++){
+			snprint(buf, j, "%s", m[i].roma);
+			prev = m[i];
+			if(hmapget(h, buf, &prev) == 0){
+				if(prev.leadstomore == 1 && s[1] == '\0'){
+					//confict; partial & valid input
+					prev = m[i];
+					prev.leadstomore = 1;
+				}
+			}
+
+			if(s[1] == '\0'){
+				hmaprepl(&h, strdup(buf), &prev, nil, 1);
+			} else {
+				hmaprepl(&h, strdup(buf), &signalmore, nil, 1);
+			}
+			j++;
+		}
+	}
+	return h;
+}
+
 void
 usage(void)
 {
@@ -101,11 +148,11 @@ main(int argc, char *argv[])
 {
 
 	uchar *bp, *ep, buf[128];
-	Map *mp;
+	Map lkup, last;
-	int nchar, wantmore;
+	int wantmore;
 	int n, c;
-	char *dictname;
+	char *jishoname, *zidianname;
-	Dictionary *jisho;
+	Hmap *jisho, *zidian;
 
 	ARGBEGIN{
 	default: usage();
@@ -113,9 +160,20 @@ main(int argc, char *argv[])
 	if(argc != 0)
 		usage();
 
-	if((dictname = getenv("jisho")) == nil)
+	if((jishoname = getenv("jisho")) == nil)
-		dictname = "/lib/kanji.jisho";
+		jishoname = "/lib/kanji.jisho";
-	jisho = openQDIC(dictname);
+	jisho = opendict(nil, jishoname);
+
+	if((zidianname = getenv("zidian")) == nil)
+		zidianname = "/lib/hanzi.zidian";
+	zidian = opendict(nil, zidianname);
+
+	hira 	= table = initmap(mhira, nelem(mhira));
+	kata 	= initmap(mkata, nelem(mkata));
+	greek 	= initmap(mgreek, nelem(mgreek));
+	cyril 	= initmap(mcyril, nelem(mcyril));
+	hangul 	= initmap(mhangul, nelem(mhangul));
+	last = (Map){nil, nil, -1};
 
 	kbdopen();
 	if(fork())
@@ -147,8 +205,8 @@ main(int argc, char *argv[])
 			wantmore = 0;
 
 			if (*bp=='') { /* ^x read ktrans-jisho once more */
-				freeQDIC(jisho);
+				jisho = opendict(jisho, jishoname);
-				jisho = openQDIC(dictname);
+				zidian = opendict(zidian, zidianname);
 				llen = 0;
 				olen = okuri = joshi = 0;
 				wantmore=0;
@@ -156,6 +214,9 @@ main(int argc, char *argv[])
 				continue;
 			}
 			if (*bp=='') { /* ^\ (start translation command) */
+				if (table == hanzi)
+					c = dotrans(zidian);
+				else
 					c = dotrans(jisho);
 				if (c)
 					*bp = c; /* pointer to translated rune */
@@ -167,11 +228,13 @@ main(int argc, char *argv[])
 				bp++;
 				llen = 0;
 				olen = okuri = joshi = 0;
+				last.kana = nil;
 				continue;
 			}
 			if (changelang(*bp)) { /* change language mode OK */
 				bp++;
 				olen = okuri = joshi = 0;
+				last.kana = nil;
 				continue;
 			}
 			if (natural || *bp<=' ' || *bp>='{') { /* English mode but not ascii */
@@ -179,6 +242,7 @@ main(int argc, char *argv[])
 				int rlen = chartorune(&r, (char *)bp);
 				send(bp, rlen); /* write bp to /dev/cons */
 				bp += rlen;
+				last.kana = nil;
 				continue;
 			}
 			if (table == hira && (*bp >= 'A' && *bp <= 'Z') && (*(bp+1) < 'A'
@@ -192,25 +256,31 @@ main(int argc, char *argv[])
 				joshi = 1;
 				olen = 0;
 			}
-			mp = match(bp, &nchar, table);
+			if(hmapget(table, (char*)bp, &lkup) < 0){
-			if (mp == 0) {
+				if(last.kana != nil){
-				if (nchar>0) { /* match, longer possible */
+					send((uchar*)last.kana, strlen(last.kana));
-					wantmore++;
+					bp += strlen(last.roma);
+				} else
+					send(bp++, 1);
+				last.kana = nil;
 				break;
 			}
-				send(bp++, 1); /* alphabet in kana mode */
+			/* concatinations; only advance a single character */
-			} else {
+			if(lkup.kana != nil && strstr("ッっ", lkup.kana))
-				send((uchar*)mp->kana, strlen(mp->kana));
+				lkup.roma = "_";
-				bp += nchar;
+			/* partial match */
-			}
+			if(lkup.kana == nil || lkup.leadstomore == 1){
-		}
+				if(lkup.kana != nil)
-	}
+					last = lkup;
-}
 
-int
+				wantmore = 1;
-min(int a, int b)
+				break;
-{
+			}
-	return a<b? a: b;
+			last.kana = nil;
+			send((uchar*)lkup.kana, strlen(lkup.kana));
+			bp += strlen(lkup.roma);
+		}
+	}
 }
 
 /*
@@ -232,7 +302,9 @@ send(uchar *p, int n)
 		llen -= 64;
 	}
 
-	if (table!=hira || natural)
+	if(table != hira && table != hanzi)
+		return;
+	if(natural && table != hanzi)
 		return;
 
 	ep = p+n;
@@ -253,49 +325,13 @@ send(uchar *p, int n)
 	   }
 }
 
-/*
- * Romaji to Hiragana/Katakana conversion
- *     romaji shoud be input as small letter
- *     returns the matched address in table, hira, kata, etc.
- *     nc: number of character (return value)
- */
-Map *
-match(uchar *p, int *nc, Map *table)
-{
-	register Map *longp = 0, *kp;
-	static char last;
-	int longest = 0;
-
-	*nc = -1;
-	for (kp=table; kp->roma; kp++) {
-		if (*p == *kp->roma) {
-			int lr = strlen(kp->roma);
-			int len = min(lr, strlen((char *)p));
-			if (strncmp(kp->roma, (char *)p, len)==0) {
-				if (len<lr) {
-					*nc = 1;
-					return 0;
-				}
-				if (len>longest) {
-					longest = len;
-					longp = kp;
-				}
-			}
-		}
-	}
-	if (longp) {
-		last = longp->roma[longest-1];
-		*nc = longp->advance;
-	}
-	return longp;
-}
-
 int
 changelang(int c)
 {
 	switch(c){
 	case '': /* ^t (English mode) */
 		natural = 1;
+		table = hira;
 		llen = 0;
 		return 1;
 		break;
@@ -334,23 +370,80 @@ changelang(int c)
 		llen = 0;
 		return 1;
 		break;
+
+	case '': /* ^c (Chinese mode) */
+		natural = 1;
+		table = hanzi;
+		llen = 0;
+		return 1;
+		break;
 	}
 	return 0;
 }
 
+Hmap*
+opendict(Hmap *h, char *name)
+{
+	Biobuf *b;
+	char *p;
+	char *dot, *rest;
+	char *kouho[16];
+	int i;
+
+	b = Bopen(name, OREAD);
+	if(b == nil)
+		return nil;
+
+	if(h == nil)
+		h = hmapalloc(8192, sizeof(kouho));
+	else
+		hmapreset(h, 1);
+	while(p = Brdstr(b, '\n', 1)){
+		if(p[0] == '\0' || p[0] == ';'){
+		Err:
+			free(p);
+			continue;
+		}
+		dot = utfrune(p, '\t');
+		if(dot == nil)
+			goto Err;
+
+		*dot = '\0';
+		rest = dot+1;
+		if(*rest == '\0')
+			goto Err;
+
+		memset(kouho, 0, sizeof kouho);
+		i = 0;
+		while(i < nelem(kouho)-1 && (dot = utfrune(rest, ' '))){
+			*dot = '\0';
+			kouho[i++] = rest;
+			rest = dot+1;
+		}
+		if(i < nelem(kouho)-1)
+			kouho[i] = rest;
+
+		/* key is the base pointer; overwrites clean up for us */
+		hmaprepl(&h, p, kouho, nil, 1);
+	}
+	Bterm(b);
+	return h;
+}
+
 /*
  * write translated kanji runes to stdout and return last character
  * if it's not ctl-\. if the last is ctl-\, proceed with
  * translation of the next kouho
  */
 int
-dotrans(Dictionary *dic)
+dotrans(Hmap *dic)
 {
 	Rune *res, r[1];
 	char v[1024], *p, tbuf[64], hirabuf[64];
 	int j, lastlen, nokouho = 0;
 	char ch;
-	KouhoList *fstkouho, *currentkouho;
+	int i;
+	char *kouho[16];
 
 	if (llen==0)
 		return 0; /* don't use kanji transform function */
@@ -375,15 +468,13 @@ dotrans(Dictionary *dic)
 	if (okuri && joshi != 1) /* verb mode */
 		hirabuf[strlen(hirabuf) - 1] = '\0';
 
-	if(!(fstkouho = getKouhoHash(dic, v))) { /* not found */
+	if(hmapget(dic, v, kouho) < 0){
 		llen = olen = okuri = joshi = 0;
 		okurigana[0] = 0;
 		return 0;
 	}
-
+	for(i = 0; i < nelem(kouho) && kouho[i] != nil; i++) {
-	currentkouho = fstkouho;
+		p = kouho[i];
-	for(;;) {
-		p = currentkouho->kouhotop; /* p to the head of kanji kouho array */
 		lastlen = nrune(tbuf); /* number of rune chars */
 
 		if (okuri && joshi != 1) /* verb mode */
@@ -407,10 +498,9 @@ dotrans(Dictionary *dic)
 			exits(nil);
 
 		if (ch == '') { /* if next input is ^\, once again */
-			if(currentkouho->nextkouho != 0) { /* have next kouho */
+			if(i+1 < nelem(kouho) && kouho[i+1] != nil) { /* have next kouho */
 				nokouho = 0;
 				strcpy(tbuf, p);
-				currentkouho = currentkouho->nextkouho;
 
 				if (okuri && joshi != 1) /* verb mode */
 					for (j=0; j<nrune(tbuf); j++)
@@ -442,8 +532,12 @@ dotrans(Dictionary *dic)
 				break;
 			}
 		} else {
-			if(!nokouho) /* learn the previous use of the kouho */
+			if(!nokouho && i != 0){  /* learn the previous use of the kouho */
-				selectKouho(&(fstkouho->dicitem->kouho), currentkouho);
+				p = kouho[0];
+				kouho[0] = kouho[i];
+				kouho[i] = p;
+				hmapupd(&dic, v, kouho);
+			}
 
 			olen = okuri = joshi = 0;
 			okurigana[0] = 0;

sys/src/cmd/ktrans/mkfile

@@ -2,9 +2,9 @@
 
 BIN=/$objtype/bin
 TARG=ktrans
-HFILES=jisho.h ktrans.h
+HFILES=ktrans.h
 OFILES=\
+	hash.$O\
 	main.$O\
-	jisho.$O
 
 </sys/src/cmd/mkone