*
* Generally speaking, a fastbuf consists of a buffer and a set of callbacks.
* All front-end functions operate on the buffer and if the buffer becomes
- * empty or fills up, they ask the corresponding callback to solve the
+ * empty or fills up, they ask the corresponding callback to handle the
* situation. Back-ends then differ just in the definition of the callbacks.
- * The state of the fastbuf is represented by `struct fastbuf`, which
- * is a simple structure describing the state of the buffer and pointers
- * to the callback functions.
- *
- * When we are reading from the fastbuf, the buffer is laid out as follows:
- *
- * +----------------+---------------------------+
- * | read data | free space |
- * +----------------+---------------------------+
- * ^ ^ ^ ^
- * buffer bptr bstop bufend
- *
- * Here `bptr` points to the next character to be read. After the last character is
- * read, `bptr == bstop` and the `refill` callback gets called upon the next read
- * attempt to bring further data. This gives us an easy way how to implement bungetc().
- *
- * When writing, the situation looks like:
- *
- * +--------+--------------+--------------------+
- * | unused | written data | free space |
- * +--------+--------------+--------------------+
- * ^ ^ ^ ^
- * buffer bstop bptr bufend
- *
- * In this case, the `bptr` points at the position where the next character
- * will be written to. When we want to write, but `bptr == bufend`, we call
- * the `spout` hook to flush the data and get an empty buffer.
- *
- * Several dirty tricks can be played:
- *
- * - The `spout`/`refill` hooks can change not only `bptr` and `bstop`, but also
- * the location and size of the buffer; the fb-mem back-end takes advantage of it.
- * - In some cases, the user of the `bdirect` interface can be allowed to modify
- * the data in the buffer to avoid unnecessary copying. If the back-end
- * allows such modifications, it can set `fastbuf->can_overwrite_buffer` accordingly:
- * * 0 if no modification is allowed,
- * * 1 if the user can modify the buffer on the condition that
- * the modifications will be undone before calling the next
- * fastbuf operation
- * * 2 if the user is allowed to overwrite the data in the buffer
- * if bdirect_read_commit_modified() is called afterwards.
- * In this case, the back-end must be prepared for trimming
- * of the buffer which is done by the commit function.
+ *
+ * The state of the fastbuf is represented by a `struct fastbuf`, which
+ * is a simple structure describing the state of the buffer (the pointers
+ * `buffer`, `bufend`), two front-end cursors (`bptr`, `bstop`), position in the file (`pos`)
+ * and pointers to the callback functions.
+ *
+ * The buffer can be in one of the following states:
+ *
+ * 1. Flushed:
+ *
+ * +----------------+---------------------------+
+ * | unused | free space |
+ * +----------------+---------------------------+
+ * ^ ^ ^
+ * buffer <= bptr == bstop (pos) <= bufend
+ *
+ * * If `bptr == bstop`, then there is no cached data and
+ * the fastbuf is ready for any read or write operation.
+ * Position of the back-end's cursor equals the front-end's one.
+ * * The interval `[bstop, bufend]` can be used by front-ends
+ * for writing. If it is empty, the `spout` callback gets called
+ * upon the first write attempt to allocate a new buffer.
+ * * When a front-end needs to read something, it calls the `spout` callback.
+ * * Any of the pointers can be NULL.
+ *
+ * 2. Reading:
+ *
+ * +----------------+---------------------------+
+ * | read data | unused |
+ * +----------------+---------------------------+
+ * ^ ^ ^ ^
+ * buffer <= bptr <= bstop (pos) <= bufend
+ *
+ * * If we try to read something, we get to the reading mode.
+ * * No writing is allowed until a flush operation. But note that @bflush()
+ * will simply set `bptr` to `bstop` and it breaks the position of the front-end's cursor.
+ * * The interval `[buffer, bstop]` contains a block of data read by the back-end.
+ * `bptr` is the front-end's cursor which points to the next character to be read.
+ * After the last character is read, `bptr == bstop` and the `refill` callback
+ * gets called upon the next read attempt to bring further data.
+ * This gives us an easy way how to implement @bungetc().
+ *
+ * 3. Writing:
+ *
+ * +---------+--------------+-------------------+
+ * | unused | written data | free space |
+ * +---------+--------------+-------------------+
+ * ^ ^ ^ ^
+ * buffer <= bstop (pos) < bptr <= bufend
+ *
+ * * This schema corresponds to the situation after a write attempt.
+ * * No reading is allowed until a flush operation.
+ * * The `bptr` points at the position where the next character
+ * will be written to. When we want to write, but `bptr == bufend`, we call
+ * the `spout` hook to flush the data and get an empty buffer.
+ *
+ *
+ * Rules for back-ends:
+ *
+ * - Front-ends are only allowed to change the value of `bptr`, some flags
+ * and if a fatal error occurs, then also `bstop`.
+ * - `buffer <= bstop <= bufend`.
+ * - `pos` should be the position in the file corresponding of the location of `bstop` in the buffer.
+ * - Failed callbacks (except `close`) should use @bthrow().
+ * - Any callback pointers may be NULL in case the callback is not implemented.
+ * - Callbacks can change not only `bptr` and `bstop`, but also the location and size of the buffer;
+ * the fb-mem back-end takes advantage of it.
+ *
+ * - initialization:
+ * * out: `buffer <= bptr == bstop <= bufend` (flushed)
+ *
+ * - `refill`:
+ * * in: `buffer <= bptr == bstop <= bufend` (reading or flushed)
+ * * out: `buffer <= bptr < bstop <= bufend` (reading)
+ *
+ * - `spout`:
+ * * in: `buffer <= bstop <= bptr <= bufend` (writing or flushed)
+ * * out: `buffer <= bstop <= bufend` (flushed)
+ * * `bptr` is set automatically to `bstop`.
+ * * If the input `bptr` equals ` bstop`, then the resulting `bstop` must be lower than `bufend`.
+ *
+ * - `seek`:
+ * * in: `buffer <= bstop == bptr <= bufend` (flushed)
+ * * out: `buffer <= bstop <= bufend` (flushed)
+ * * `bptr` is set automatically to `bstop`.
+ *
+ * - `close`:
+ * * in: `buffer <= bptr == bstop <= bufend` (flushed)
+ * * `close` must always free all internal structures, even when it throws an exception.
*
***/
ucw_off_t pos; /* Position of bstop in the file */
int (*refill)(struct fastbuf *); /* Get a buffer with new data, returns 0 on EOF */
void (*spout)(struct fastbuf *); /* Write buffer data to the file */
- int (*seek)(struct fastbuf *, ucw_off_t, int);/* Slow path for bseek(), buffer already flushed; returns success */
+ int (*seek)(struct fastbuf *, ucw_off_t, int);/* Slow path for @bseek(), buffer already flushed; returns success */
void (*close)(struct fastbuf *); /* Close the stream */
int (*config)(struct fastbuf *, uns, int); /* Configure the stream */
int can_overwrite_buffer; /* Can the buffer be altered? 0=never, 1=temporarily, 2=permanently */
* Creates a fastbuf from a file descriptor @fd and sets its filename
* to @name (the name is used only in error messages).
* When the fastbuf is closed, the fd is closed as well. You can override
- * this behavior by calling bconfig().
+ * this behavior by calling @bconfig().
*/
struct fastbuf *bopen_fd_name(int fd, struct fb_params *params, const char *name);
static inline struct fastbuf *bopen_fd(int fd, struct fb_params *params) /** Same as above, but with an auto-generated filename. **/
* up any parameters, there is a couple of shortcuts.
***/
-struct fastbuf *bopen(const char *name, uns mode, uns buflen); /** Equivalent to bopen_file() with `FB_STD` back-end. **/
-struct fastbuf *bopen_try(const char *name, uns mode, uns buflen); /** Equivalent to bopen_file_try() with `FB_STD` back-end. **/
-struct fastbuf *bopen_tmp(uns buflen); /** Equivalent to bopen_tmp_file() with `FB_STD` back-end. **/
-struct fastbuf *bfdopen(int fd, uns buflen); /** Equivalent to bopen_fd() with `FB_STD` back-end. **/
-struct fastbuf *bfdopen_shared(int fd, uns buflen); /** Like bfdopen(), but it does not close the @fd on bclose(). **/
+struct fastbuf *bopen(const char *name, uns mode, uns buflen); /** Equivalent to @bopen_file() with `FB_STD` back-end. **/
+struct fastbuf *bopen_try(const char *name, uns mode, uns buflen); /** Equivalent to @bopen_file_try() with `FB_STD` back-end. **/
+struct fastbuf *bopen_tmp(uns buflen); /** Equivalent to @bopen_tmp_file() with `FB_STD` back-end. **/
+struct fastbuf *bfdopen(int fd, uns buflen); /** Equivalent to @bopen_fd() with `FB_STD` back-end. **/
+struct fastbuf *bfdopen_shared(int fd, uns buflen); /** Like @bfdopen(), but it does not close the @fd on @bclose(). **/
/***
* === Temporary files [[fbtemp]]
*
- * Usually, bopen_tmp_file() is the best way how to come to a temporary file.
+ * Usually, @bopen_tmp_file() is the best way how to come to a temporary file.
* However, in some specific cases you can need more, so there is also a set
* of more general functions.
***/
/**
* Sometimes, a file is created as temporary and then moved to a stable
- * location. This function takes a fastbuf created by bopen_tmp_file()
- * or bopen_tmp(), marks it as permanent, closes it and renames it to
+ * location. This function takes a fastbuf created by @bopen_tmp_file()
+ * or @bopen_tmp(), marks it as permanent, closes it and renames it to
* @name.
*
* Please note that it assumes that the temporary file and the @name
* in memory (as a linked list of memory blocks, so address space
* fragmentation is avoided).
*
- * First, you use fbmem_create() to create the stream and the fastbuf
- * used for writing to it. Then you can call fbmem_clone_read() to get
+ * First, you use @fbmem_create() to create the stream and the fastbuf
+ * used for writing to it. Then you can call @fbmem_clone_read() to get
* an arbitrary number of fastbuf for reading from the stream.
***/
* The fastbuf structure is allocated by the caller and pointed to by @f.
* An attempt to write behind the end of the buffer dies.
*
- * Data are written directly into the buffer, so it is not necessary to call bflush()
+ * Data are written directly into the buffer, so it is not necessary to call @bflush()
* at any moment.
*
* It is not possible to close this fastbuf.
**/
void fbpool_init(struct fbpool *fb); /** Initialize a new mempool fastbuf. **/
/**
- * Start a new continuous block and prepare for writing (see mp_start()).
+ * Start a new continuous block and prepare for writing (see <<mempool:mp_start()>>).
* Provide the memory pool you want to use for this block as @mp.
**/
void fbpool_start(struct fbpool *fb, struct mempool *mp, uns init_size);
/**
- * Close the block and return the address of its start (see mp_end()).
- * The length can be determined by calling mp_size(mp, ptr).
+ * Close the block and return the address of its start (see <<mempool:mp_end()>>).
+ * The length can be determined by calling <<mempool:mp_size(mp, ptr)>>.
**/
void *fbpool_end(struct fbpool *fb);
* If the file has fixed record length, just set @record_len to it.
* Otherwise set @record_len to the expected maximum record length
* with a negative sign (you need not fit in this length, but as long
- * as you do, the fastbuf is more efficient) and call fbatomic_commit()
+ * as you do, the fastbuf is more efficient) and call @fbatomic_commit()
* after each record.
*
* You can specify @record_len, if it is known (for optimisations).
bputc_slow(f, c);
}
-static inline uns bavailr(struct fastbuf *f)
+static inline uns bavailr(struct fastbuf *f) /** Return the length of the cached data to be read. Do not use directly. **/
{
return f->bstop - f->bptr;
}
-static inline uns bavailw(struct fastbuf *f)
+static inline uns bavailw(struct fastbuf *f) /** Return the length of the buffer available for writing. Do not use directly. **/
{
return f->bufend - f->bptr;
}
* Dies if the line is longer than @l.
**/
char *bgets(struct fastbuf *f, char *b, uns l);
-char *bgets0(struct fastbuf *f, char *b, uns l); /** The same as bgets(), but for 0-terminated strings. **/
+char *bgets0(struct fastbuf *f, char *b, uns l); /** The same as @bgets(), but for 0-terminated strings. **/
/**
* Returns either length of read string (excluding the terminator) or -1 if it is too long.
* In such cases exactly @l bytes are read.
void bbcopy_slow(struct fastbuf *f, struct fastbuf *t, uns l);
/**
* Copy @l bytes of data from fastbuf @f to fastbuf @t.
+ * `UINT_MAX` (`~0U`) means all data, even if more than `UINT_MAX` bytes remain.
**/
static inline void bbcopy(struct fastbuf *f, struct fastbuf *t, uns l)
{
}
/*** === Direct I/O on buffers ***/
-// TODO Documentation -- what do they do?
-static inline uns
-bdirect_read_prepare(struct fastbuf *f, byte **buf)
+/**
+ * Begin direct reading from fastbuf's internal buffer to avoid unnecessary copying.
+ * The function returns a buffer @buf together with its length in bytes (zero means EOF)
+ * with cached data to be read.
+ *
+ * Some back-ends allow the user to modify the data in the returned buffer to avoid unnecessary.
+ * If the back-end allows such modifications, it can set `f->can_overwrite_buffer` accordingly:
+ *
+ * - 0 if no modification is allowed,
+ * - 1 if the user can modify the buffer on the condition that
+ * the modifications will be undone before calling the next
+ * fastbuf operation
+ * - 2 if the user is allowed to overwrite the data in the buffer
+ * if @bdirect_read_commit_modified() is called afterwards.
+ * In this case, the back-end must be prepared for trimming
+ * of the buffer which is done by the commit function.
+ *
+ * The reading must be ended by @bdirect_read_commit() or @bdirect_read_commit_modified(),
+ * unless the user did not read or modify anything.
+ **/
+static inline uns bdirect_read_prepare(struct fastbuf *f, byte **buf)
{
if (f->bptr == f->bstop && !f->refill(f))
{
return bavailr(f);
}
-static inline void
-bdirect_read_commit(struct fastbuf *f, byte *pos)
+/**
+ * End direct reading started by @bdirect_read_prepare() and move the cursor at @pos.
+ * Data in the returned buffer must be same as after @bdirect_read_prepare() and
+ * @pos must point somewhere inside the buffer.
+ **/
+static inline void bdirect_read_commit(struct fastbuf *f, byte *pos)
{
f->bptr = pos;
}
-static inline void
-bdirect_read_commit_modified(struct fastbuf *f, byte *pos)
+/**
+ * Similar to @bdirect_read_commit(), but accepts also modified data before @pos.
+ * Note that such modifications are supported only if `f->can_overwrite_buffer == 2`.
+ **/
+static inline void bdirect_read_commit_modified(struct fastbuf *f, byte *pos)
{
f->bptr = pos;
f->buffer = pos; /* Avoid seeking backwards in the buffer */
}
-static inline uns
-bdirect_write_prepare(struct fastbuf *f, byte **buf)
+/**
+ * Start direct writing to fastbuf's internal buffer to avoid copy overhead.
+ * The function returns the length of the buffer in @buf (at least one byte)
+ * where we can write to. The operation must be ended by @bdirect_write_commit(),
+ * unless nothing is written.
+ **/
+static inline uns bdirect_write_prepare(struct fastbuf *f, byte **buf)
{
if (f->bptr == f->bufend)
f->spout(f);
return bavailw(f);
}
-static inline void
-bdirect_write_commit(struct fastbuf *f, byte *pos)
+/**
+ * Commit the data written to the buffer returned by @bdirect_write_prepare().
+ * The length is specified by @pos which must point just after the written data.
+ * Also moves the cursor to @pos.
+ **/
+static inline void bdirect_write_commit(struct fastbuf *f, byte *pos)
{
f->bptr = pos;
}
projects / libucw.git / blobdiff