1PERLIOL(1) Perl Programmers Reference Guide PERLIOL(1)
2
6 perliol - C API for Perl's implementation of IO in Layers.
7
9 /* Defining a layer ... */
10 #include <perliol.h>
11
13 This document describes the behavior and implementation of the PerlIO
14 abstraction described in perlapio when "USE_PERLIO" is defined (and
15 "USE_SFIO" is not).
16 History and Background
18 The PerlIO abstraction was introduced in perl5.003_02 but languished as
19 just an abstraction until perl5.7.0. However during that time a number
20 of perl extensions switched to using it, so the API is mostly fixed to
21 maintain (source) compatibility.
22 The aim of the implementation is to provide the PerlIO API in a
24 flexible and platform neutral manner. It is also a trial of an "Object
25 Oriented C, with vtables" approach which may be applied to Perl 6.
26 Basic Structure
28 PerlIO is a stack of layers.
29 The low levels of the stack work with the low-level operating system
31 calls (file descriptors in C) getting bytes in and out, the higher
32 layers of the stack buffer, filter, and otherwise manipulate the I/O,
33 and return characters (or bytes) to Perl. Terms above and below are
34 used to refer to the relative positioning of the stack layers.
35 A layer contains a "vtable", the table of I/O operations (at C level a
37 table of function pointers), and status flags. The functions in the
38 vtable implement operations like "open", "read", and "write".
39 When I/O, for example "read", is requested, the request goes from Perl
41 first down the stack using "read" functions of each layer, then at the
42 bottom the input is requested from the operating system services, then
43 the result is returned up the stack, finally being interpreted as Perl
44 data.
45 The requests do not necessarily go always all the way down to the
47 operating system: that's where PerlIO buffering comes into play.
48 When you do an open() and specify extra PerlIO layers to be deployed,
50 the layers you specify are "pushed" on top of the already existing
51 default stack. One way to see it is that "operating system is on the
52 left" and "Perl is on the right".
53 What exact layers are in this default stack depends on a lot of things:
55 your operating system, Perl version, Perl compile time configuration,
56 and Perl runtime configuration. See PerlIO, "PERLIO" in perlrun, and
57 open for more information.
58 binmode() operates similarly to open(): by default the specified layers
60 are pushed on top of the existing stack.
61 However, note that even as the specified layers are "pushed on top" for
63 open() and binmode(), this doesn't mean that the effects are limited to
64 the "top": PerlIO layers can be very 'active' and inspect and affect
65 layers also deeper in the stack. As an example there is a layer called
66 "raw" which repeatedly "pops" layers until it reaches the first layer
67 that has declared itself capable of handling binary data. The "pushed"
68 layers are processed in left-to-right order.
69 sysopen() operates (unsurprisingly) at a lower level in the stack than
71 open(). For example in UNIX or UNIX-like systems sysopen() operates
72 directly at the level of file descriptors: in the terms of PerlIO
73 layers, it uses only the "unix" layer, which is a rather thin wrapper
74 on top of the UNIX file descriptors.
75 Layers vs Disciplines
77 Initial discussion of the ability to modify IO streams behaviour used
78 the term "discipline" for the entities which were added. This came (I
79 believe) from the use of the term in "sfio", which in turn borrowed it
80 from "line disciplines" on Unix terminals. However, this document (and
81 the C code) uses the term "layer".
82 This is, I hope, a natural term given the implementation, and should
84 avoid connotations that are inherent in earlier uses of "discipline"
85 for things which are rather different.
86 Data Structures
88 The basic data structure is a PerlIOl:
89 typedef struct _PerlIO PerlIOl;
91 typedef struct _PerlIO_funcs PerlIO_funcs;
92 typedef PerlIOl *PerlIO;
93 struct _PerlIO
95 {
96 PerlIOl * next; /* Lower layer */
97 PerlIO_funcs * tab; /* Functions for this layer */
98 IV flags; /* Various flags for state */
99 };
100 A "PerlIOl *" is a pointer to the struct, and the application level
102 "PerlIO *" is a pointer to a "PerlIOl *" - i.e. a pointer to a pointer
103 to the struct. This allows the application level "PerlIO *" to remain
104 constant while the actual "PerlIOl *" underneath changes. (Compare
105 perl's "SV *" which remains constant while its "sv_any" field changes
106 as the scalar's type changes.) An IO stream is then in general
107 represented as a pointer to this linked-list of "layers".
108 It should be noted that because of the double indirection in a "PerlIO
110 *", a "&(perlio->next)" "is" a "PerlIO *", and so to some degree at
111 least one layer can use the "standard" API on the next layer down.
112 A "layer" is composed of two parts:
114 1. The functions and attributes of the "layer class".
116 2. The per-instance data for a particular handle.
118 Functions and Attributes
120 The functions and attributes are accessed via the "tab" (for table)
121 member of "PerlIOl". The functions (methods of the layer "class") are
122 fixed, and are defined by the "PerlIO_funcs" type. They are broadly the
123 same as the public "PerlIO_xxxxx" functions:
124 struct _PerlIO_funcs
126 {
127 Size_t fsize;
128 char * name;
129 Size_t size;
130 IV kind;
131 IV (*Pushed)(pTHX_ PerlIO *f,const char *mode,SV *arg, PerlIO_funcs *tab);
132 IV (*Popped)(pTHX_ PerlIO *f);
133 PerlIO * (*Open)(pTHX_ PerlIO_funcs *tab,
134 PerlIO_list_t *layers, IV n,
135 const char *mode,
136 int fd, int imode, int perm,
137 PerlIO *old,
138 int narg, SV **args);
139 IV (*Binmode)(pTHX_ PerlIO *f);
140 SV * (*Getarg)(pTHX_ PerlIO *f, CLONE_PARAMS *param, int flags)
141 IV (*Fileno)(pTHX_ PerlIO *f);
142 PerlIO * (*Dup)(pTHX_ PerlIO *f, PerlIO *o, CLONE_PARAMS *param, int flags)
143 /* Unix-like functions - cf sfio line disciplines */
144 SSize_t (*Read)(pTHX_ PerlIO *f, void *vbuf, Size_t count);
145 SSize_t (*Unread)(pTHX_ PerlIO *f, const void *vbuf, Size_t count);
146 SSize_t (*Write)(pTHX_ PerlIO *f, const void *vbuf, Size_t count);
147 IV (*Seek)(pTHX_ PerlIO *f, Off_t offset, int whence);
148 Off_t (*Tell)(pTHX_ PerlIO *f);
149 IV (*Close)(pTHX_ PerlIO *f);
150 /* Stdio-like buffered IO functions */
151 IV (*Flush)(pTHX_ PerlIO *f);
152 IV (*Fill)(pTHX_ PerlIO *f);
153 IV (*Eof)(pTHX_ PerlIO *f);
154 IV (*Error)(pTHX_ PerlIO *f);
155 void (*Clearerr)(pTHX_ PerlIO *f);
156 void (*Setlinebuf)(pTHX_ PerlIO *f);
157 /* Perl's snooping functions */
158 STDCHAR * (*Get_base)(pTHX_ PerlIO *f);
159 Size_t (*Get_bufsiz)(pTHX_ PerlIO *f);
160 STDCHAR * (*Get_ptr)(pTHX_ PerlIO *f);
161 SSize_t (*Get_cnt)(pTHX_ PerlIO *f);
162 void (*Set_ptrcnt)(pTHX_ PerlIO *f,STDCHAR *ptr,SSize_t cnt);
163 };
164 The first few members of the struct give a function table size for
166 compatibility check "name" for the layer, the size to "malloc" for the
167 per-instance data, and some flags which are attributes of the class as
168 whole (such as whether it is a buffering layer), then follow the
169 functions which fall into four basic groups:
170 1. Opening and setup functions
172 2. Basic IO operations
174 3. Stdio class buffering options.
176 4. Functions to support Perl's traditional "fast" access to the
178 buffer.
179 A layer does not have to implement all the functions, but the whole
181 table has to be present. Unimplemented slots can be NULL (which will
182 result in an error when called) or can be filled in with stubs to
183 "inherit" behaviour from a "base class". This "inheritance" is fixed
184 for all instances of the layer, but as the layer chooses which stubs to
185 populate the table, limited "multiple inheritance" is possible.
186 Per-instance Data
188 The per-instance data are held in memory beyond the basic PerlIOl
189 struct, by making a PerlIOl the first member of the layer's struct
190 thus:
191 typedef struct
193 {
194 struct _PerlIO base; /* Base "class" info */
195 STDCHAR * buf; /* Start of buffer */
196 STDCHAR * end; /* End of valid part of buffer */
197 STDCHAR * ptr; /* Current position in buffer */
198 Off_t posn; /* Offset of buf into the file */
199 Size_t bufsiz; /* Real size of buffer */
200 IV oneword; /* Emergency buffer */
201 } PerlIOBuf;
202 In this way (as for perl's scalars) a pointer to a PerlIOBuf can be
204 treated as a pointer to a PerlIOl.
205 Layers in action.
207 table perlio unix
208 | |
209 +-----------+ +----------+ +--------+
210 PerlIO ->| |--->| next |--->| NULL |
211 +-----------+ +----------+ +--------+
212 | | | buffer | | fd |
213 +-----------+ | | +--------+
214 | | +----------+
215 The above attempts to show how the layer scheme works in a simple case.
217 The application's "PerlIO *" points to an entry in the table(s)
218 representing open (allocated) handles. For example the first three
219 slots in the table correspond to "stdin","stdout" and "stderr". The
220 table in turn points to the current "top" layer for the handle - in
221 this case an instance of the generic buffering layer "perlio". That
222 layer in turn points to the next layer down - in this case the low-
223 level "unix" layer.
224 The above is roughly equivalent to a "stdio" buffered stream, but with
226 much more flexibility:
227 · If Unix level "read"/"write"/"lseek" is not appropriate for (say)
229 sockets then the "unix" layer can be replaced (at open time or even
230 dynamically) with a "socket" layer.
231 · Different handles can have different buffering schemes. The "top"
233 layer could be the "mmap" layer if reading disk files was quicker
234 using "mmap" than "read". An "unbuffered" stream can be implemented
235 simply by not having a buffer layer.
236 · Extra layers can be inserted to process the data as it flows
238 through. This was the driving need for including the scheme in
239 perl 5.7.0+ - we needed a mechanism to allow data to be translated
240 between perl's internal encoding (conceptually at least Unicode as
241 UTF-8), and the "native" format used by the system. This is
242 provided by the ":encoding(xxxx)" layer which typically sits above
243 the buffering layer.
244 · A layer can be added that does "\n" to CRLF translation. This layer
246 can be used on any platform, not just those that normally do such
247 things.
248 Per-instance flag bits
250 The generic flag bits are a hybrid of "O_XXXXX" style flags deduced
251 from the mode string passed to "PerlIO_open()", and state bits for
252 typical buffer layers.
253 PERLIO_F_EOF
255 End of file.
256 PERLIO_F_CANWRITE
258 Writes are permitted, i.e. opened as "w" or "r+" or "a", etc.
259 PERLIO_F_CANREAD
261 Reads are permitted i.e. opened "r" or "w+" (or even "a+" - ick).
262 PERLIO_F_ERROR
264 An error has occurred (for "PerlIO_error()").
265 PERLIO_F_TRUNCATE
267 Truncate file suggested by open mode.
268 PERLIO_F_APPEND
270 All writes should be appends.
271 PERLIO_F_CRLF
273 Layer is performing Win32-like "\n" mapped to CR,LF for output and
274 CR,LF mapped to "\n" for input. Normally the provided "crlf" layer
275 is the only layer that need bother about this. "PerlIO_binmode()"
276 will mess with this flag rather than add/remove layers if the
277 "PERLIO_K_CANCRLF" bit is set for the layers class.
278 PERLIO_F_UTF8
280 Data written to this layer should be UTF-8 encoded; data provided
281 by this layer should be considered UTF-8 encoded. Can be set on any
282 layer by ":utf8" dummy layer. Also set on ":encoding" layer.
283 PERLIO_F_UNBUF
285 Layer is unbuffered - i.e. write to next layer down should occur
286 for each write to this layer.
287 PERLIO_F_WRBUF
289 The buffer for this layer currently holds data written to it but
290 not sent to next layer.
291 PERLIO_F_RDBUF
293 The buffer for this layer currently holds unconsumed data read from
294 layer below.
295 PERLIO_F_LINEBUF
297 Layer is line buffered. Write data should be passed to next layer
298 down whenever a "\n" is seen. Any data beyond the "\n" should then
299 be processed.
300 PERLIO_F_TEMP
302 File has been "unlink()"ed, or should be deleted on "close()".
303 PERLIO_F_OPEN
305 Handle is open.
306 PERLIO_F_FASTGETS
308 This instance of this layer supports the "fast "gets"" interface.
309 Normally set based on "PERLIO_K_FASTGETS" for the class and by the
310 existence of the function(s) in the table. However a class that
311 normally provides that interface may need to avoid it on a
312 particular instance. The "pending" layer needs to do this when it
313 is pushed above a layer which does not support the interface.
314 (Perl's "sv_gets()" does not expect the streams fast "gets"
315 behaviour to change during one "get".)
316 Methods in Detail
318 fsize
319 Size_t fsize;
320 Size of the function table. This is compared against the value
322 PerlIO code "knows" as a compatibility check. Future versions may
323 be able to tolerate layers compiled against an old version of the
324 headers.
325 name
327 char * name;
328 The name of the layer whose open() method Perl should invoke on
330 open(). For example if the layer is called APR, you will call:
331 open $fh, ">:APR", ...
333 and Perl knows that it has to invoke the PerlIOAPR_open() method
335 implemented by the APR layer.
336 size
338 Size_t size;
339 The size of the per-instance data structure, e.g.:
341 sizeof(PerlIOAPR)
343 If this field is zero then "PerlIO_pushed" does not malloc anything
345 and assumes layer's Pushed function will do any required layer
346 stack manipulation - used to avoid malloc/free overhead for dummy
347 layers. If the field is non-zero it must be at least the size of
348 "PerlIOl", "PerlIO_pushed" will allocate memory for the layer's
349 data structures and link new layer onto the stream's stack. (If the
350 layer's Pushed method returns an error indication the layer is
351 popped again.)
352 kind
354 IV kind;
355 · PERLIO_K_BUFFERED
357 The layer is buffered.
359 · PERLIO_K_RAW
361 The layer is acceptable to have in a binmode(FH) stack - i.e.
363 it does not (or will configure itself not to) transform bytes
364 passing through it.
365 · PERLIO_K_CANCRLF
367 Layer can translate between "\n" and CRLF line ends.
369 · PERLIO_K_FASTGETS
371 Layer allows buffer snooping.
373 · PERLIO_K_MULTIARG
375 Used when the layer's open() accepts more arguments than usual.
377 The extra arguments should come not before the "MODE" argument.
378 When this flag is used it's up to the layer to validate the
379 args.
380 Pushed
382 IV (*Pushed)(pTHX_ PerlIO *f,const char *mode, SV *arg);
383 The only absolutely mandatory method. Called when the layer is
385 pushed onto the stack. The "mode" argument may be NULL if this
386 occurs post-open. The "arg" will be non-"NULL" if an argument
387 string was passed. In most cases this should call
388 "PerlIOBase_pushed()" to convert "mode" into the appropriate
389 "PERLIO_F_XXXXX" flags in addition to any actions the layer itself
390 takes. If a layer is not expecting an argument it need neither
391 save the one passed to it, nor provide "Getarg()" (it could perhaps
392 "Perl_warn" that the argument was un-expected).
393 Returns 0 on success. On failure returns -1 and should set errno.
395 Popped
397 IV (*Popped)(pTHX_ PerlIO *f);
398 Called when the layer is popped from the stack. A layer will
400 normally be popped after "Close()" is called. But a layer can be
401 popped without being closed if the program is dynamically managing
402 layers on the stream. In such cases "Popped()" should free any
403 resources (buffers, translation tables, ...) not held directly in
404 the layer's struct. It should also "Unread()" any unconsumed data
405 that has been read and buffered from the layer below back to that
406 layer, so that it can be re-provided to what ever is now above.
407 Returns 0 on success and failure. If "Popped()" returns true then
409 perlio.c assumes that either the layer has popped itself, or the
410 layer is super special and needs to be retained for other reasons.
411 In most cases it should return false.
412 Open
414 PerlIO * (*Open)(...);
415 The "Open()" method has lots of arguments because it combines the
417 functions of perl's "open", "PerlIO_open", perl's "sysopen",
418 "PerlIO_fdopen" and "PerlIO_reopen". The full prototype is as
419 follows:
420 PerlIO * (*Open)(pTHX_ PerlIO_funcs *tab,
422 PerlIO_list_t *layers, IV n,
423 const char *mode,
424 int fd, int imode, int perm,
425 PerlIO *old,
426 int narg, SV **args);
427 Open should (perhaps indirectly) call "PerlIO_allocate()" to
429 allocate a slot in the table and associate it with the layers
430 information for the opened file, by calling "PerlIO_push". The
431 layers is an array of all the layers destined for the "PerlIO *",
432 and any arguments passed to them, n is the index into that array of
433 the layer being called. The macro "PerlIOArg" will return a
434 (possibly "NULL") SV * for the argument passed to the layer.
435 The mode string is an ""fopen()"-like" string which would match the
437 regular expression "/^[I#]?[rwa]\+?[bt]?$/".
438 The 'I' prefix is used during creation of "stdin".."stderr" via
440 special "PerlIO_fdopen" calls; the '#' prefix means that this is
441 "sysopen" and that imode and perm should be passed to
442 "PerlLIO_open3"; 'r' means read, 'w' means write and 'a' means
443 append. The '+' suffix means that both reading and
444 writing/appending are permitted. The 'b' suffix means file should
445 be binary, and 't' means it is text. (Almost all layers should do
446 the IO in binary mode, and ignore the b/t bits. The ":crlf" layer
447 should be pushed to handle the distinction.)
448 If old is not "NULL" then this is a "PerlIO_reopen". Perl itself
450 does not use this (yet?) and semantics are a little vague.
451 If fd not negative then it is the numeric file descriptor fd, which
453 will be open in a manner compatible with the supplied mode string,
454 the call is thus equivalent to "PerlIO_fdopen". In this case nargs
455 will be zero.
456 If nargs is greater than zero then it gives the number of arguments
458 passed to "open", otherwise it will be 1 if for example
459 "PerlIO_open" was called. In simple cases SvPV_nolen(*args) is the
460 pathname to open.
461 Having said all that translation-only layers do not need to provide
463 "Open()" at all, but rather leave the opening to a lower level
464 layer and wait to be "pushed". If a layer does provide "Open()" it
465 should normally call the "Open()" method of next layer down (if
466 any) and then push itself on top if that succeeds.
467 If "PerlIO_push" was performed and open has failed, it must
469 "PerlIO_pop" itself, since if it's not, the layer won't be removed
470 and may cause bad problems.
471 Returns "NULL" on failure.
473 Binmode
475 IV (*Binmode)(pTHX_ PerlIO *f);
476 Optional. Used when ":raw" layer is pushed (explicitly or as a
478 result of binmode(FH)). If not present layer will be popped. If
479 present should configure layer as binary (or pop itself) and return
480 0. If it returns -1 for error "binmode" will fail with layer still
481 on the stack.
482 Getarg
484 SV * (*Getarg)(pTHX_ PerlIO *f,
485 CLONE_PARAMS *param, int flags);
486 Optional. If present should return an SV * representing the string
488 argument passed to the layer when it was pushed. e.g.
489 ":encoding(ascii)" would return an SvPV with value "ascii". (param
490 and flags arguments can be ignored in most cases)
491 "Dup" uses "Getarg" to retrieve the argument originally passed to
493 "Pushed", so you must implement this function if your layer has an
494 extra argument to "Pushed" and will ever be "Dup"ed.
495 Fileno
497 IV (*Fileno)(pTHX_ PerlIO *f);
498 Returns the Unix/Posix numeric file descriptor for the handle.
500 Normally "PerlIOBase_fileno()" (which just asks next layer down)
501 will suffice for this.
502 Returns -1 on error, which is considered to include the case where
504 the layer cannot provide such a file descriptor.
505 Dup
507 PerlIO * (*Dup)(pTHX_ PerlIO *f, PerlIO *o,
508 CLONE_PARAMS *param, int flags);
509 XXX: Needs more docs.
511 Used as part of the "clone" process when a thread is spawned (in
513 which case param will be non-NULL) and when a stream is being
514 duplicated via '&' in the "open".
515 Similar to "Open", returns PerlIO* on success, "NULL" on failure.
517 Read
519 SSize_t (*Read)(pTHX_ PerlIO *f, void *vbuf, Size_t count);
520 Basic read operation.
522 Typically will call "Fill" and manipulate pointers (possibly via
524 the API). "PerlIOBuf_read()" may be suitable for derived classes
525 which provide "fast gets" methods.
526 Returns actual bytes read, or -1 on an error.
528 Unread
530 SSize_t (*Unread)(pTHX_ PerlIO *f,
531 const void *vbuf, Size_t count);
532 A superset of stdio's "ungetc()". Should arrange for future reads
534 to see the bytes in "vbuf". If there is no obviously better
535 implementation then "PerlIOBase_unread()" provides the function by
536 pushing a "fake" "pending" layer above the calling layer.
537 Returns the number of unread chars.
539 Write
541 SSize_t (*Write)(PerlIO *f, const void *vbuf, Size_t count);
542 Basic write operation.
544 Returns bytes written or -1 on an error.
546 Seek
548 IV (*Seek)(pTHX_ PerlIO *f, Off_t offset, int whence);
549 Position the file pointer. Should normally call its own "Flush"
551 method and then the "Seek" method of next layer down.
552 Returns 0 on success, -1 on failure.
554 Tell
556 Off_t (*Tell)(pTHX_ PerlIO *f);
557 Return the file pointer. May be based on layers cached concept of
559 position to avoid overhead.
560 Returns -1 on failure to get the file pointer.
562 Close
564 IV (*Close)(pTHX_ PerlIO *f);
565 Close the stream. Should normally call "PerlIOBase_close()" to
567 flush itself and close layers below, and then deallocate any data
568 structures (buffers, translation tables, ...) not held directly in
569 the data structure.
570 Returns 0 on success, -1 on failure.
572 Flush
574 IV (*Flush)(pTHX_ PerlIO *f);
575 Should make stream's state consistent with layers below. That is,
577 any buffered write data should be written, and file position of
578 lower layers adjusted for data read from below but not actually
579 consumed. (Should perhaps "Unread()" such data to the lower
580 layer.)
581 Returns 0 on success, -1 on failure.
583 Fill
585 IV (*Fill)(pTHX_ PerlIO *f);
586 The buffer for this layer should be filled (for read) from layer
588 below. When you "subclass" PerlIOBuf layer, you want to use its
589 _read method and to supply your own fill method, which fills the
590 PerlIOBuf's buffer.
591 Returns 0 on success, -1 on failure.
593 Eof
595 IV (*Eof)(pTHX_ PerlIO *f);
596 Return end-of-file indicator. "PerlIOBase_eof()" is normally
598 sufficient.
599 Returns 0 on end-of-file, 1 if not end-of-file, -1 on error.
601 Error
603 IV (*Error)(pTHX_ PerlIO *f);
604 Return error indicator. "PerlIOBase_error()" is normally
606 sufficient.
607 Returns 1 if there is an error (usually when "PERLIO_F_ERROR" is
609 set, 0 otherwise.
610 Clearerr
612 void (*Clearerr)(pTHX_ PerlIO *f);
613 Clear end-of-file and error indicators. Should call
615 "PerlIOBase_clearerr()" to set the "PERLIO_F_XXXXX" flags, which
616 may suffice.
617 Setlinebuf
619 void (*Setlinebuf)(pTHX_ PerlIO *f);
620 Mark the stream as line buffered. "PerlIOBase_setlinebuf()" sets
622 the PERLIO_F_LINEBUF flag and is normally sufficient.
623 Get_base
625 STDCHAR * (*Get_base)(pTHX_ PerlIO *f);
626 Allocate (if not already done so) the read buffer for this layer
628 and return pointer to it. Return NULL on failure.
629 Get_bufsiz
631 Size_t (*Get_bufsiz)(pTHX_ PerlIO *f);
632 Return the number of bytes that last "Fill()" put in the buffer.
634 Get_ptr
636 STDCHAR * (*Get_ptr)(pTHX_ PerlIO *f);
637 Return the current read pointer relative to this layer's buffer.
639 Get_cnt
641 SSize_t (*Get_cnt)(pTHX_ PerlIO *f);
642 Return the number of bytes left to be read in the current buffer.
644 Set_ptrcnt
646 void (*Set_ptrcnt)(pTHX_ PerlIO *f,
647 STDCHAR *ptr, SSize_t cnt);
648 Adjust the read pointer and count of bytes to match "ptr" and/or
650 "cnt". The application (or layer above) must ensure they are
651 consistent. (Checking is allowed by the paranoid.)
652 Utilities
654 To ask for the next layer down use PerlIONext(PerlIO *f).
655 To check that a PerlIO* is valid use PerlIOValid(PerlIO *f). (All this
657 does is really just to check that the pointer is non-NULL and that the
658 pointer behind that is non-NULL.)
659 PerlIOBase(PerlIO *f) returns the "Base" pointer, or in other words,
661 the "PerlIOl*" pointer.
662 PerlIOSelf(PerlIO* f, type) return the PerlIOBase cast to a type.
664 Perl_PerlIO_or_Base(PerlIO* f, callback, base, failure, args) either
666 calls the callback from the functions of the layer f (just by the name
667 of the IO function, like "Read") with the args, or if there is no such
668 callback, calls the base version of the callback with the same args, or
669 if the f is invalid, set errno to EBADF and return failure.
670 Perl_PerlIO_or_fail(PerlIO* f, callback, failure, args) either calls
672 the callback of the functions of the layer f with the args, or if there
673 is no such callback, set errno to EINVAL. Or if the f is invalid, set
674 errno to EBADF and return failure.
675 Perl_PerlIO_or_Base_void(PerlIO* f, callback, base, args) either calls
677 the callback of the functions of the layer f with the args, or if there
678 is no such callback, calls the base version of the callback with the
679 same args, or if the f is invalid, set errno to EBADF.
680 Perl_PerlIO_or_fail_void(PerlIO* f, callback, args) either calls the
682 callback of the functions of the layer f with the args, or if there is
683 no such callback, set errno to EINVAL. Or if the f is invalid, set
684 errno to EBADF.
685 Implementing PerlIO Layers
687 If you find the implementation document unclear or not sufficient, look
688 at the existing PerlIO layer implementations, which include:
689 · C implementations
691 The perlio.c and perliol.h in the Perl core implement the "unix",
693 "perlio", "stdio", "crlf", "utf8", "byte", "raw", "pending" layers,
694 and also the "mmap" and "win32" layers if applicable. (The "win32"
695 is currently unfinished and unused, to see what is used instead in
696 Win32, see "Querying the layers of filehandles" in PerlIO .)
697 PerlIO::encoding, PerlIO::scalar, PerlIO::via in the Perl core.
699 PerlIO::gzip and APR::PerlIO (mod_perl 2.0) on CPAN.
701 · Perl implementations
703 PerlIO::via::QuotedPrint in the Perl core and PerlIO::via::* on
705 CPAN.
706 If you are creating a PerlIO layer, you may want to be lazy, in other
708 words, implement only the methods that interest you. The other methods
709 you can either replace with the "blank" methods
710 PerlIOBase_noop_ok
712 PerlIOBase_noop_fail
713 (which do nothing, and return zero and -1, respectively) or for certain
715 methods you may assume a default behaviour by using a NULL method. The
716 Open method looks for help in the 'parent' layer. The following table
717 summarizes the behaviour:
718 method behaviour with NULL
720 Clearerr PerlIOBase_clearerr
722 Close PerlIOBase_close
723 Dup PerlIOBase_dup
724 Eof PerlIOBase_eof
725 Error PerlIOBase_error
726 Fileno PerlIOBase_fileno
727 Fill FAILURE
728 Flush SUCCESS
729 Getarg SUCCESS
730 Get_base FAILURE
731 Get_bufsiz FAILURE
732 Get_cnt FAILURE
733 Get_ptr FAILURE
734 Open INHERITED
735 Popped SUCCESS
736 Pushed SUCCESS
737 Read PerlIOBase_read
738 Seek FAILURE
739 Set_cnt FAILURE
740 Set_ptrcnt FAILURE
741 Setlinebuf PerlIOBase_setlinebuf
742 Tell FAILURE
743 Unread PerlIOBase_unread
744 Write FAILURE
745 FAILURE Set errno (to EINVAL in UNIXish, to LIB$_INVARG in VMS) and
747 return -1 (for numeric return values) or NULL (for pointers)
748 INHERITED Inherited from the layer below
749 SUCCESS Return 0 (for numeric return values) or a pointer
750 Core Layers
752 The file "perlio.c" provides the following layers:
753 "unix"
755 A basic non-buffered layer which calls Unix/POSIX "read()",
756 "write()", "lseek()", "close()". No buffering. Even on platforms
757 that distinguish between O_TEXT and O_BINARY this layer is always
758 O_BINARY.
759 "perlio"
761 A very complete generic buffering layer which provides the whole of
762 PerlIO API. It is also intended to be used as a "base class" for
763 other layers. (For example its "Read()" method is implemented in
764 terms of the "Get_cnt()"/"Get_ptr()"/"Set_ptrcnt()" methods).
765 "perlio" over "unix" provides a complete replacement for stdio as
767 seen via PerlIO API. This is the default for USE_PERLIO when
768 system's stdio does not permit perl's "fast gets" access, and which
769 do not distinguish between "O_TEXT" and "O_BINARY".
770 "stdio"
772 A layer which provides the PerlIO API via the layer scheme, but
773 implements it by calling system's stdio. This is (currently) the
774 default if system's stdio provides sufficient access to allow
775 perl's "fast gets" access and which do not distinguish between
776 "O_TEXT" and "O_BINARY".
777 "crlf"
779 A layer derived using "perlio" as a base class. It provides
780 Win32-like "\n" to CR,LF translation. Can either be applied above
781 "perlio" or serve as the buffer layer itself. "crlf" over "unix" is
782 the default if system distinguishes between "O_TEXT" and "O_BINARY"
783 opens. (At some point "unix" will be replaced by a "native" Win32
784 IO layer on that platform, as Win32's read/write layer has various
785 drawbacks.) The "crlf" layer is a reasonable model for a layer
786 which transforms data in some way.
787 "mmap"
789 If Configure detects "mmap()" functions this layer is provided
790 (with "perlio" as a "base") which does "read" operations by
791 mmap()ing the file. Performance improvement is marginal on modern
792 systems, so it is mainly there as a proof of concept. It is likely
793 to be unbundled from the core at some point. The "mmap" layer is a
794 reasonable model for a minimalist "derived" layer.
795 "pending"
797 An "internal" derivative of "perlio" which can be used to provide
798 Unread() function for layers which have no buffer or cannot be
799 bothered. (Basically this layer's "Fill()" pops itself off the
800 stack and so resumes reading from layer below.)
801 "raw"
803 A dummy layer which never exists on the layer stack. Instead when
804 "pushed" it actually pops the stack removing itself, it then calls
805 Binmode function table entry on all the layers in the stack -
806 normally this (via PerlIOBase_binmode) removes any layers which do
807 not have "PERLIO_K_RAW" bit set. Layers can modify that behaviour
808 by defining their own Binmode entry.
809 "utf8"
811 Another dummy layer. When pushed it pops itself and sets the
812 "PERLIO_F_UTF8" flag on the layer which was (and now is once more)
813 the top of the stack.
814 In addition perlio.c also provides a number of "PerlIOBase_xxxx()"
816 functions which are intended to be used in the table slots of classes
817 which do not need to do anything special for a particular method.
818 Extension Layers
820 Layers can made available by extension modules. When an unknown layer
821 is encountered the PerlIO code will perform the equivalent of :
822 use PerlIO 'layer';
824 Where layer is the unknown layer. PerlIO.pm will then attempt to:
826 require PerlIO::layer;
828 If after that process the layer is still not defined then the "open"
830 will fail.
831 The following extension layers are bundled with perl:
833 ":encoding"
835 use Encoding;
836 makes this layer available, although PerlIO.pm "knows" where to
838 find it. It is an example of a layer which takes an argument as it
839 is called thus:
840 open( $fh, "<:encoding(iso-8859-7)", $pathname );
842 ":scalar"
844 Provides support for reading data from and writing data to a
845 scalar.
846 open( $fh, "+<:scalar", \$scalar );
848 When a handle is so opened, then reads get bytes from the string
850 value of $scalar, and writes change the value. In both cases the
851 position in $scalar starts as zero but can be altered via "seek",
852 and determined via "tell".
853 Please note that this layer is implied when calling open() thus:
855 open( $fh, "+<", \$scalar );
857 ":via"
859 Provided to allow layers to be implemented as Perl code. For
860 instance:
861 use PerlIO::via::StripHTML;
863 open( my $fh, "<:via(StripHTML)", "index.html" );
864 See PerlIO::via for details.
866
868 Things that need to be done to improve this document.
869 · Explain how to make a valid fh without going through open()(i.e.
871 apply a layer). For example if the file is not opened through perl,
872 but we want to get back a fh, like it was opened by Perl.
873 How PerlIO_apply_layera fits in, where its docs, was it made
875 public?
876 Currently the example could be something like this:
878 PerlIO *foo_to_PerlIO(pTHX_ char *mode, ...)
880 {
881 char *mode; /* "w", "r", etc */
882 const char *layers = ":APR"; /* the layer name */
883 PerlIO *f = PerlIO_allocate(aTHX);
884 if (!f) {
885 return NULL;
886 }
887 PerlIO_apply_layers(aTHX_ f, mode, layers);
889 if (f) {
891 PerlIOAPR *st = PerlIOSelf(f, PerlIOAPR);
892 /* fill in the st struct, as in _open() */
893 st->file = file;
894 PerlIOBase(f)->flags |= PERLIO_F_OPEN;
895 return f;
897 }
898 return NULL;
899 }
900 · fix/add the documentation in places marked as XXX.
902 · The handling of errors by the layer is not specified. e.g. when $!
904 should be set explicitly, when the error handling should be just
905 delegated to the top layer.
906 Probably give some hints on using SETERRNO() or pointers to where
908 they can be found.
909 · I think it would help to give some concrete examples to make it
911 easier to understand the API. Of course I agree that the API has to
912 be concise, but since there is no second document that is more of a
913 guide, I think that it'd make it easier to start with the doc which
914 is an API, but has examples in it in places where things are
915 unclear, to a person who is not a PerlIO guru (yet).
916perl v5.10.1 2009-02-12 PERLIOL(1)