* Copyright (c) 1988 Mark Nudleman
* Copyright (c) 1988 Regents of the University of California.
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* 1. Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* 2. Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
* 3. All advertising materials mentioning features or use of this software
* must display the following acknowledgement:
* This product includes software developed by the University of
* California, Berkeley and its contributors.
* 4. Neither the name of the University nor the names of its contributors
* may be used to endorse or promote products derived from this software
* without specific prior written permission.
* THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
* ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
* ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
* FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
* DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
* OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
* HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
* LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
* OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
static char sccsid
[] = "@(#)ch.c 5.11 (Berkeley) 6/21/92";
* Low level character input from the input file.
* We use these special purpose routines which optimize moving
* both forward and backward from the current read pointer.
int file
= -1; /* File descriptor of the input file */
* Pool of buffers holding the most recently used blocks of the input file.
* The buffer pool is kept as a doubly-linked circular list, in order from
* most- to least-recently used. The circular list is anchored by buf_anchor.
#define END_OF_CHAIN ((struct buf *)&buf_anchor)
#define buf_head buf_anchor.next
#define buf_tail buf_anchor.prev
} buf_anchor
= { END_OF_CHAIN
, END_OF_CHAIN
};
extern int ispipe
, cbufs
, sigs
;
* Current position in file.
* Stored as a block number and an offset into the block.
/* Length of file, needed if input is a pipe. */
/* Number of bytes read, if input is standard input (a pipe). */
static off_t last_piped_pos
;
* Get the character pointed to by the read pointer. ch_get() is a macro
* which is more efficient to call than fch_get (the function), in the usual
* case that the block desired is at the head of the chain.
((buf_head->block == ch_block && \
ch_offset < buf_head->datasize) ? \
buf_head->data[ch_offset] : fch_get())
/* look for a buffer holding the desired block. */
for (bp
= buf_head
; bp
!= END_OF_CHAIN
; bp
= bp
->next
)
if (bp
->block
== ch_block
) {
if (ch_offset
>= bp
->datasize
)
* Need more data in this buffer.
* On a pipe, we don't sort the buffers LRU
* because this can cause gaps in the buffers.
* For example, suppose we've got 12 1K buffers,
* and a 15K input stream. If we read the first 12K
* sequentially, then jump to line 1, then jump to
* the end, the buffers have blocks 0,4,5,6,..,14.
* If we then jump to line 1 again and try to
* read sequentially, we're out of luck when we
* get to block 1 (we'd get the "pipe error" below).
* To avoid this, we only sort buffers on a pipe
* when we actually READ the data, not when we
* find it already buffered.
return(bp
->data
[ch_offset
]);
* Block is not in a buffer. Take the least recently used buffer
* and read the desired block into it. If the LRU buffer has data
* in it, and input is a pipe, then try to allocate a new buffer first.
if (ispipe
&& buf_tail
->block
!= (long)(-1))
pos
= (ch_block
* BUFSIZ
) + bp
->datasize
;
* The data requested should be immediately after
* the last data read from the pipe.
if (pos
!= last_piped_pos
) {
(void)lseek(file
, pos
, L_SET
);
* If we read less than a full block, we just return the
* partial block and pick up the rest next time.
n
= iread(file
, &bp
->data
[bp
->datasize
], BUFSIZ
- bp
->datasize
);
p
= &bp
->data
[bp
->datasize
];
* Set an EOI marker in the buffered data itself. Then ensure the
* data is "clean": there are no extra EOI chars in the data and
* that the "meta" bit (the 0200 bit) is reset in each char;
* also translate \r\n sequences to \n if -u flag not set.
bp
->data
[bp
->datasize
++] = EOI
;
for (p
= &bp
->data
[bp
->datasize
]; --n
>= 0;) {
for (t
= p
; --n
>= 0; ++p
) {
if (ch
== '\r' && n
&& (p
[1] & 0177) == '\n') {
*t
++ = (ch
== EOI
) ? 0200 : ch
;
* Move the buffer to the head of the buffer chain.
* This orders the buffer chain, most- to least-recently used.
bp
->next
->prev
= bp
->prev
;
bp
->prev
->next
= bp
->next
;
if (ch_offset
>= bp
->datasize
)
* After all that, we still don't have enough data.
return(bp
->data
[ch_offset
]);
* Determine if a specific block is currently in one of the buffers.
for (bp
= buf_head
; bp
!= END_OF_CHAIN
; bp
= bp
->next
)
* Seek to a specified position in the file.
* Return 0 if successful, non-zero if can't seek there.
new_block
= pos
/ BUFSIZ
;
if (!ispipe
|| pos
== last_piped_pos
|| buffered(new_block
)) {
ch_offset
= pos
% BUFSIZ
;
* Seek to the end of the file.
return(ch_seek(ch_length()));
* Do it the slow way: read till end of data.
while (ch_forw_get() != EOI
)
* Seek to the beginning of the file, or as close to it as we can get.
* We may not be able to seek there if input is a pipe and the
* beginning of the pipe is no longer buffered.
register struct buf
*bp
, *firstbp
;
* Try a plain ch_seek first.
if (ch_seek((off_t
)0) == 0)
* Can't get to position 0.
* Look thru the buffers for the one closest to position 0.
while ((bp
= bp
->next
) != END_OF_CHAIN
)
if (bp
->block
< firstbp
->block
)
ch_block
= firstbp
->block
;
* Return the length of the file, if known.
return((off_t
)(lseek(file
, (off_t
)0, L_XTND
)));
* Return the current position in the file.
return(ch_block
* BUFSIZ
+ ch_offset
);
* Get the current char and post-increment the read pointer.
if (c
!= EOI
&& ++ch_offset
>= BUFSIZ
) {
* Pre-decrement the read pointer and get the new current char.
if (ch_block
<= 0 || (ispipe
&& !buffered(ch_block
-1))) {
* Caller wants us to have a total of at least want_nbufs buffers.
* keep==1 means keep the data in the current buffers;
* otherwise discard the old data.
ch_init(want_nbufs
, keep
)
if (nbufs
< want_nbufs
&& ch_addbuf(want_nbufs
- nbufs
)) {
* Cannot allocate enough buffers.
* If we don't have ANY, then quit.
* Otherwise, just report the error and return.
(void)sprintf(message
, "cannot allocate %d buffers",
* We don't want to keep the old data,
* so initialize all the buffers now.
for (bp
= buf_head
; bp
!= END_OF_CHAIN
; bp
= bp
->next
)
last_piped_pos
= (off_t
)0;
ch_fsize
= NULL_POSITION
;
* Allocate some new buffers.
* The buffers are added to the tail of the buffer chain.
register struct buf
*newbufs
;
* We don't have enough buffers.
* Allocate some new ones.
newbufs
= (struct buf
*)calloc((u_int
)nnew
, sizeof(struct buf
));
* Initialize the new buffers and link them together.
* Link them all onto the tail of the buffer list.
for (bp
= &newbufs
[0]; bp
< &newbufs
[nnew
]; bp
++) {
newbufs
[nnew
-1].next
= END_OF_CHAIN
;
newbufs
[0].prev
= buf_tail
;
buf_tail
->next
= &newbufs
[0];
buf_tail
= &newbufs
[nnew
-1];