3 perliol - C API for Perl's implementation of IO in Layers.
7 /* Defining a layer ... */
12 This document describes the behavior and implementation of the PerlIO
13 abstraction described in L<perlapio> when C<USE_PERLIO> is defined.
15 =head2 History and Background
17 The PerlIO abstraction was introduced in perl5.003_02 but languished as
18 just an abstraction until perl5.7.0. However during that time a number
19 of perl extensions switched to using it, so the API is mostly fixed to
20 maintain (source) compatibility.
22 The aim of the implementation is to provide the PerlIO API in a flexible
23 and platform neutral manner. It is also a trial of an "Object Oriented
24 C, with vtables" approach which may be applied to Perl 6.
26 =head2 Basic Structure
28 PerlIO is a stack of layers.
30 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 I<above> and I<below>
34 are used to refer to the relative positioning of the stack layers.
36 A layer contains a "vtable", the table of I/O operations (at C level
37 a table of function pointers), and status flags. The functions in the
38 vtable implement operations like "open", "read", and "write".
40 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
46 The requests do not necessarily go always all the way down to the
47 operating system: that's where PerlIO buffering comes into play.
49 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
52 on the left" and "Perl is on the right".
54 What exact layers are in this default stack depends on a lot of
55 things: your operating system, Perl version, Perl compile time
56 configuration, and Perl runtime configuration. See L<PerlIO>,
57 L<perlrun/PERLIO>, and L<open> for more information.
59 binmode() operates similarly to open(): by default the specified
60 layers are pushed on top of the existing stack.
62 However, note that even as the specified layers are "pushed on top"
63 for open() and binmode(), this doesn't mean that the effects are
64 limited to the "top": PerlIO layers can be very 'active' and inspect
65 and affect layers also deeper in the stack. As an example there
66 is a layer called "raw" which repeatedly "pops" layers until
67 it reaches the first layer that has declared itself capable of
68 handling binary data. The "pushed" layers are processed in left-to-right
71 sysopen() operates (unsurprisingly) at a lower level in the stack than
72 open(). For example in Unix or Unix-like systems sysopen() operates
73 directly at the level of file descriptors: in the terms of PerlIO
74 layers, it uses only the "unix" layer, which is a rather thin wrapper
75 on top of the Unix file descriptors.
77 =head2 Layers vs Disciplines
79 Initial discussion of the ability to modify IO streams behaviour used
80 the term "discipline" for the entities which were added. This came (I
81 believe) from the use of the term in "sfio", which in turn borrowed it
82 from "line disciplines" on Unix terminals. However, this document (and
83 the C code) uses the term "layer".
85 This is, I hope, a natural term given the implementation, and should
86 avoid connotations that are inherent in earlier uses of "discipline"
87 for things which are rather different.
89 =head2 Data Structures
91 The basic data structure is a PerlIOl:
93 typedef struct _PerlIO PerlIOl;
94 typedef struct _PerlIO_funcs PerlIO_funcs;
95 typedef PerlIOl *PerlIO;
99 PerlIOl * next; /* Lower layer */
100 PerlIO_funcs * tab; /* Functions for this layer */
101 U32 flags; /* Various flags for state */
104 A C<PerlIOl *> is a pointer to the struct, and the I<application>
105 level C<PerlIO *> is a pointer to a C<PerlIOl *> - i.e. a pointer
106 to a pointer to the struct. This allows the application level C<PerlIO *>
107 to remain constant while the actual C<PerlIOl *> underneath
108 changes. (Compare perl's C<SV *> which remains constant while its
109 C<sv_any> field changes as the scalar's type changes.) An IO stream is
110 then in general represented as a pointer to this linked-list of
113 It should be noted that because of the double indirection in a C<PerlIO *>,
114 a C<< &(perlio->next) >> "is" a C<PerlIO *>, and so to some degree
115 at least one layer can use the "standard" API on the next layer down.
117 A "layer" is composed of two parts:
123 The functions and attributes of the "layer class".
127 The per-instance data for a particular handle.
131 =head2 Functions and Attributes
133 The functions and attributes are accessed via the "tab" (for table)
134 member of C<PerlIOl>. The functions (methods of the layer "class") are
135 fixed, and are defined by the C<PerlIO_funcs> type. They are broadly the
136 same as the public C<PerlIO_xxxxx> functions:
144 IV (*Pushed)(pTHX_ PerlIO *f,const char *mode,SV *arg, PerlIO_funcs *tab);
145 IV (*Popped)(pTHX_ PerlIO *f);
146 PerlIO * (*Open)(pTHX_ PerlIO_funcs *tab,
147 PerlIO_list_t *layers, IV n,
149 int fd, int imode, int perm,
151 int narg, SV **args);
152 IV (*Binmode)(pTHX_ PerlIO *f);
153 SV * (*Getarg)(pTHX_ PerlIO *f, CLONE_PARAMS *param, int flags)
154 IV (*Fileno)(pTHX_ PerlIO *f);
155 PerlIO * (*Dup)(pTHX_ PerlIO *f, PerlIO *o, CLONE_PARAMS *param, int flags)
156 /* Unix-like functions - cf sfio line disciplines */
157 SSize_t (*Read)(pTHX_ PerlIO *f, void *vbuf, Size_t count);
158 SSize_t (*Unread)(pTHX_ PerlIO *f, const void *vbuf, Size_t count);
159 SSize_t (*Write)(pTHX_ PerlIO *f, const void *vbuf, Size_t count);
160 IV (*Seek)(pTHX_ PerlIO *f, Off_t offset, int whence);
161 Off_t (*Tell)(pTHX_ PerlIO *f);
162 IV (*Close)(pTHX_ PerlIO *f);
163 /* Stdio-like buffered IO functions */
164 IV (*Flush)(pTHX_ PerlIO *f);
165 IV (*Fill)(pTHX_ PerlIO *f);
166 IV (*Eof)(pTHX_ PerlIO *f);
167 IV (*Error)(pTHX_ PerlIO *f);
168 void (*Clearerr)(pTHX_ PerlIO *f);
169 void (*Setlinebuf)(pTHX_ PerlIO *f);
170 /* Perl's snooping functions */
171 STDCHAR * (*Get_base)(pTHX_ PerlIO *f);
172 Size_t (*Get_bufsiz)(pTHX_ PerlIO *f);
173 STDCHAR * (*Get_ptr)(pTHX_ PerlIO *f);
174 SSize_t (*Get_cnt)(pTHX_ PerlIO *f);
175 void (*Set_ptrcnt)(pTHX_ PerlIO *f,STDCHAR *ptr,SSize_t cnt);
178 The first few members of the struct give a function table size for
179 compatibility check "name" for the layer, the size to C<malloc> for the per-instance data,
180 and some flags which are attributes of the class as whole (such as whether it is a buffering
181 layer), then follow the functions which fall into four basic groups:
187 Opening and setup functions
195 Stdio class buffering options.
199 Functions to support Perl's traditional "fast" access to the buffer.
203 A layer does not have to implement all the functions, but the whole
204 table has to be present. Unimplemented slots can be NULL (which will
205 result in an error when called) or can be filled in with stubs to
206 "inherit" behaviour from a "base class". This "inheritance" is fixed
207 for all instances of the layer, but as the layer chooses which stubs
208 to populate the table, limited "multiple inheritance" is possible.
210 =head2 Per-instance Data
212 The per-instance data are held in memory beyond the basic PerlIOl
213 struct, by making a PerlIOl the first member of the layer's struct
218 struct _PerlIO base; /* Base "class" info */
219 STDCHAR * buf; /* Start of buffer */
220 STDCHAR * end; /* End of valid part of buffer */
221 STDCHAR * ptr; /* Current position in buffer */
222 Off_t posn; /* Offset of buf into the file */
223 Size_t bufsiz; /* Real size of buffer */
224 IV oneword; /* Emergency buffer */
227 In this way (as for perl's scalars) a pointer to a PerlIOBuf can be
228 treated as a pointer to a PerlIOl.
230 =head2 Layers in action.
234 +-----------+ +----------+ +--------+
235 PerlIO ->| |--->| next |--->| NULL |
236 +-----------+ +----------+ +--------+
237 | | | buffer | | fd |
238 +-----------+ | | +--------+
242 The above attempts to show how the layer scheme works in a simple case.
243 The application's C<PerlIO *> points to an entry in the table(s)
244 representing open (allocated) handles. For example the first three slots
245 in the table correspond to C<stdin>,C<stdout> and C<stderr>. The table
246 in turn points to the current "top" layer for the handle - in this case
247 an instance of the generic buffering layer "perlio". That layer in turn
248 points to the next layer down - in this case the low-level "unix" layer.
250 The above is roughly equivalent to a "stdio" buffered stream, but with
251 much more flexibility:
257 If Unix level C<read>/C<write>/C<lseek> is not appropriate for (say)
258 sockets then the "unix" layer can be replaced (at open time or even
259 dynamically) with a "socket" layer.
263 Different handles can have different buffering schemes. The "top"
264 layer could be the "mmap" layer if reading disk files was quicker
265 using C<mmap> than C<read>. An "unbuffered" stream can be implemented
266 simply by not having a buffer layer.
270 Extra layers can be inserted to process the data as it flows through.
271 This was the driving need for including the scheme in perl 5.7.0+ - we
272 needed a mechanism to allow data to be translated between perl's
273 internal encoding (conceptually at least Unicode as UTF-8), and the
274 "native" format used by the system. This is provided by the
275 ":encoding(xxxx)" layer which typically sits above the buffering layer.
279 A layer can be added that does "\n" to CRLF translation. This layer
280 can be used on any platform, not just those that normally do such
285 =head2 Per-instance flag bits
287 The generic flag bits are a hybrid of C<O_XXXXX> style flags deduced
288 from the mode string passed to C<PerlIO_open()>, and state bits for
289 typical buffer layers.
297 =item PERLIO_F_CANWRITE
299 Writes are permitted, i.e. opened as "w" or "r+" or "a", etc.
301 =item PERLIO_F_CANREAD
303 Reads are permitted i.e. opened "r" or "w+" (or even "a+" - ick).
307 An error has occurred (for C<PerlIO_error()>).
309 =item PERLIO_F_TRUNCATE
311 Truncate file suggested by open mode.
313 =item PERLIO_F_APPEND
315 All writes should be appends.
319 Layer is performing Win32-like "\n" mapped to CR,LF for output and CR,LF
320 mapped to "\n" for input. Normally the provided "crlf" layer is the only
321 layer that need bother about this. C<PerlIO_binmode()> will mess with this
322 flag rather than add/remove layers if the C<PERLIO_K_CANCRLF> bit is set
323 for the layers class.
327 Data written to this layer should be UTF-8 encoded; data provided
328 by this layer should be considered UTF-8 encoded. Can be set on any layer
329 by ":utf8" dummy layer. Also set on ":encoding" layer.
333 Layer is unbuffered - i.e. write to next layer down should occur for
334 each write to this layer.
338 The buffer for this layer currently holds data written to it but not sent
343 The buffer for this layer currently holds unconsumed data read from
346 =item PERLIO_F_LINEBUF
348 Layer is line buffered. Write data should be passed to next layer down
349 whenever a "\n" is seen. Any data beyond the "\n" should then be
354 File has been C<unlink()>ed, or should be deleted on C<close()>.
360 =item PERLIO_F_FASTGETS
362 This instance of this layer supports the "fast C<gets>" interface.
363 Normally set based on C<PERLIO_K_FASTGETS> for the class and by the
364 existence of the function(s) in the table. However a class that
365 normally provides that interface may need to avoid it on a
366 particular instance. The "pending" layer needs to do this when
367 it is pushed above a layer which does not support the interface.
368 (Perl's C<sv_gets()> does not expect the streams fast C<gets> behaviour
369 to change during one "get".)
373 =head2 Methods in Detail
381 Size of the function table. This is compared against the value PerlIO
382 code "knows" as a compatibility check. Future versions I<may> be able
383 to tolerate layers compiled against an old version of the headers.
389 The name of the layer whose open() method Perl should invoke on
390 open(). For example if the layer is called APR, you will call:
392 open $fh, ">:APR", ...
394 and Perl knows that it has to invoke the PerlIOAPR_open() method
395 implemented by the APR layer.
401 The size of the per-instance data structure, e.g.:
405 If this field is zero then C<PerlIO_pushed> does not malloc anything
406 and assumes layer's Pushed function will do any required layer stack
407 manipulation - used to avoid malloc/free overhead for dummy layers.
408 If the field is non-zero it must be at least the size of C<PerlIOl>,
409 C<PerlIO_pushed> will allocate memory for the layer's data structures
410 and link new layer onto the stream's stack. (If the layer's Pushed
411 method returns an error indication the layer is popped again.)
419 =item * PERLIO_K_BUFFERED
421 The layer is buffered.
425 The layer is acceptable to have in a binmode(FH) stack - i.e. it does not
426 (or will configure itself not to) transform bytes passing through it.
428 =item * PERLIO_K_CANCRLF
430 Layer can translate between "\n" and CRLF line ends.
432 =item * PERLIO_K_FASTGETS
434 Layer allows buffer snooping.
436 =item * PERLIO_K_MULTIARG
438 Used when the layer's open() accepts more arguments than usual. The
439 extra arguments should come not before the C<MODE> argument. When this
440 flag is used it's up to the layer to validate the args.
446 IV (*Pushed)(pTHX_ PerlIO *f,const char *mode, SV *arg);
448 The only absolutely mandatory method. Called when the layer is pushed
449 onto the stack. The C<mode> argument may be NULL if this occurs
450 post-open. The C<arg> will be non-C<NULL> if an argument string was
451 passed. In most cases this should call C<PerlIOBase_pushed()> to
452 convert C<mode> into the appropriate C<PERLIO_F_XXXXX> flags in
453 addition to any actions the layer itself takes. If a layer is not
454 expecting an argument it need neither save the one passed to it, nor
455 provide C<Getarg()> (it could perhaps C<Perl_warn> that the argument
458 Returns 0 on success. On failure returns -1 and should set errno.
462 IV (*Popped)(pTHX_ PerlIO *f);
464 Called when the layer is popped from the stack. A layer will normally
465 be popped after C<Close()> is called. But a layer can be popped
466 without being closed if the program is dynamically managing layers on
467 the stream. In such cases C<Popped()> should free any resources
468 (buffers, translation tables, ...) not held directly in the layer's
469 struct. It should also C<Unread()> any unconsumed data that has been
470 read and buffered from the layer below back to that layer, so that it
471 can be re-provided to what ever is now above.
473 Returns 0 on success and failure. If C<Popped()> returns I<true> then
474 I<perlio.c> assumes that either the layer has popped itself, or the
475 layer is super special and needs to be retained for other reasons.
476 In most cases it should return I<false>.
480 PerlIO * (*Open)(...);
482 The C<Open()> method has lots of arguments because it combines the
483 functions of perl's C<open>, C<PerlIO_open>, perl's C<sysopen>,
484 C<PerlIO_fdopen> and C<PerlIO_reopen>. The full prototype is as
487 PerlIO * (*Open)(pTHX_ PerlIO_funcs *tab,
488 PerlIO_list_t *layers, IV n,
490 int fd, int imode, int perm,
492 int narg, SV **args);
494 Open should (perhaps indirectly) call C<PerlIO_allocate()> to allocate
495 a slot in the table and associate it with the layers information for
496 the opened file, by calling C<PerlIO_push>. The I<layers> is an
497 array of all the layers destined for the C<PerlIO *>, and any
498 arguments passed to them, I<n> is the index into that array of the
499 layer being called. The macro C<PerlIOArg> will return a (possibly
500 C<NULL>) SV * for the argument passed to the layer.
502 The I<mode> string is an "C<fopen()>-like" string which would match
503 the regular expression C</^[I#]?[rwa]\+?[bt]?$/>.
505 The C<'I'> prefix is used during creation of C<stdin>..C<stderr> via
506 special C<PerlIO_fdopen> calls; the C<'#'> prefix means that this is
507 C<sysopen> and that I<imode> and I<perm> should be passed to
508 C<PerlLIO_open3>; C<'r'> means B<r>ead, C<'w'> means B<w>rite and
509 C<'a'> means B<a>ppend. The C<'+'> suffix means that both reading and
510 writing/appending are permitted. The C<'b'> suffix means file should
511 be binary, and C<'t'> means it is text. (Almost all layers should do
512 the IO in binary mode, and ignore the b/t bits. The C<:crlf> layer
513 should be pushed to handle the distinction.)
515 If I<old> is not C<NULL> then this is a C<PerlIO_reopen>. Perl itself
516 does not use this (yet?) and semantics are a little vague.
518 If I<fd> not negative then it is the numeric file descriptor I<fd>,
519 which will be open in a manner compatible with the supplied mode
520 string, the call is thus equivalent to C<PerlIO_fdopen>. In this case
521 I<nargs> will be zero.
523 If I<nargs> is greater than zero then it gives the number of arguments
524 passed to C<open>, otherwise it will be 1 if for example
525 C<PerlIO_open> was called. In simple cases SvPV_nolen(*args) is the
528 If a layer provides C<Open()> it should normally call the C<Open()>
529 method of next layer down (if any) and then push itself on top if that
530 succeeds. C<PerlIOBase_open> is provided to do exactly that, so in
531 most cases you don't have to write your own C<Open()> method. If this
532 method is not defined, other layers may have difficulty pushing
533 themselves on top of it during open.
535 If C<PerlIO_push> was performed and open has failed, it must
536 C<PerlIO_pop> itself, since if it's not, the layer won't be removed
537 and may cause bad problems.
539 Returns C<NULL> on failure.
543 IV (*Binmode)(pTHX_ PerlIO *f);
545 Optional. Used when C<:raw> layer is pushed (explicitly or as a result
546 of binmode(FH)). If not present layer will be popped. If present
547 should configure layer as binary (or pop itself) and return 0.
548 If it returns -1 for error C<binmode> will fail with layer
553 SV * (*Getarg)(pTHX_ PerlIO *f,
554 CLONE_PARAMS *param, int flags);
556 Optional. If present should return an SV * representing the string
557 argument passed to the layer when it was
558 pushed. e.g. ":encoding(ascii)" would return an SvPV with value
559 "ascii". (I<param> and I<flags> arguments can be ignored in most
562 C<Dup> uses C<Getarg> to retrieve the argument originally passed to
563 C<Pushed>, so you must implement this function if your layer has an
564 extra argument to C<Pushed> and will ever be C<Dup>ed.
568 IV (*Fileno)(pTHX_ PerlIO *f);
570 Returns the Unix/Posix numeric file descriptor for the handle. Normally
571 C<PerlIOBase_fileno()> (which just asks next layer down) will suffice
574 Returns -1 on error, which is considered to include the case where the
575 layer cannot provide such a file descriptor.
579 PerlIO * (*Dup)(pTHX_ PerlIO *f, PerlIO *o,
580 CLONE_PARAMS *param, int flags);
582 XXX: Needs more docs.
584 Used as part of the "clone" process when a thread is spawned (in which
585 case param will be non-NULL) and when a stream is being duplicated via
588 Similar to C<Open>, returns PerlIO* on success, C<NULL> on failure.
592 SSize_t (*Read)(pTHX_ PerlIO *f, void *vbuf, Size_t count);
594 Basic read operation.
596 Typically will call C<Fill> and manipulate pointers (possibly via the
597 API). C<PerlIOBuf_read()> may be suitable for derived classes which
598 provide "fast gets" methods.
600 Returns actual bytes read, or -1 on an error.
604 SSize_t (*Unread)(pTHX_ PerlIO *f,
605 const void *vbuf, Size_t count);
607 A superset of stdio's C<ungetc()>. Should arrange for future reads to
608 see the bytes in C<vbuf>. If there is no obviously better implementation
609 then C<PerlIOBase_unread()> provides the function by pushing a "fake"
610 "pending" layer above the calling layer.
612 Returns the number of unread chars.
616 SSize_t (*Write)(PerlIO *f, const void *vbuf, Size_t count);
618 Basic write operation.
620 Returns bytes written or -1 on an error.
624 IV (*Seek)(pTHX_ PerlIO *f, Off_t offset, int whence);
626 Position the file pointer. Should normally call its own C<Flush>
627 method and then the C<Seek> method of next layer down.
629 Returns 0 on success, -1 on failure.
633 Off_t (*Tell)(pTHX_ PerlIO *f);
635 Return the file pointer. May be based on layers cached concept of
636 position to avoid overhead.
638 Returns -1 on failure to get the file pointer.
642 IV (*Close)(pTHX_ PerlIO *f);
644 Close the stream. Should normally call C<PerlIOBase_close()> to flush
645 itself and close layers below, and then deallocate any data structures
646 (buffers, translation tables, ...) not held directly in the data
649 Returns 0 on success, -1 on failure.
653 IV (*Flush)(pTHX_ PerlIO *f);
655 Should make stream's state consistent with layers below. That is, any
656 buffered write data should be written, and file position of lower layers
657 adjusted for data read from below but not actually consumed.
658 (Should perhaps C<Unread()> such data to the lower layer.)
660 Returns 0 on success, -1 on failure.
664 IV (*Fill)(pTHX_ PerlIO *f);
666 The buffer for this layer should be filled (for read) from layer
667 below. When you "subclass" PerlIOBuf layer, you want to use its
668 I<_read> method and to supply your own fill method, which fills the
671 Returns 0 on success, -1 on failure.
675 IV (*Eof)(pTHX_ PerlIO *f);
677 Return end-of-file indicator. C<PerlIOBase_eof()> is normally sufficient.
679 Returns 0 on end-of-file, 1 if not end-of-file, -1 on error.
683 IV (*Error)(pTHX_ PerlIO *f);
685 Return error indicator. C<PerlIOBase_error()> is normally sufficient.
687 Returns 1 if there is an error (usually when C<PERLIO_F_ERROR> is set),
692 void (*Clearerr)(pTHX_ PerlIO *f);
694 Clear end-of-file and error indicators. Should call C<PerlIOBase_clearerr()>
695 to set the C<PERLIO_F_XXXXX> flags, which may suffice.
699 void (*Setlinebuf)(pTHX_ PerlIO *f);
701 Mark the stream as line buffered. C<PerlIOBase_setlinebuf()> sets the
702 PERLIO_F_LINEBUF flag and is normally sufficient.
706 STDCHAR * (*Get_base)(pTHX_ PerlIO *f);
708 Allocate (if not already done so) the read buffer for this layer and
709 return pointer to it. Return NULL on failure.
713 Size_t (*Get_bufsiz)(pTHX_ PerlIO *f);
715 Return the number of bytes that last C<Fill()> put in the buffer.
719 STDCHAR * (*Get_ptr)(pTHX_ PerlIO *f);
721 Return the current read pointer relative to this layer's buffer.
725 SSize_t (*Get_cnt)(pTHX_ PerlIO *f);
727 Return the number of bytes left to be read in the current buffer.
731 void (*Set_ptrcnt)(pTHX_ PerlIO *f,
732 STDCHAR *ptr, SSize_t cnt);
734 Adjust the read pointer and count of bytes to match C<ptr> and/or C<cnt>.
735 The application (or layer above) must ensure they are consistent.
736 (Checking is allowed by the paranoid.)
742 To ask for the next layer down use PerlIONext(PerlIO *f).
744 To check that a PerlIO* is valid use PerlIOValid(PerlIO *f). (All
745 this does is really just to check that the pointer is non-NULL and
746 that the pointer behind that is non-NULL.)
748 PerlIOBase(PerlIO *f) returns the "Base" pointer, or in other words,
749 the C<PerlIOl*> pointer.
751 PerlIOSelf(PerlIO* f, type) return the PerlIOBase cast to a type.
753 Perl_PerlIO_or_Base(PerlIO* f, callback, base, failure, args) either
754 calls the I<callback> from the functions of the layer I<f> (just by
755 the name of the IO function, like "Read") with the I<args>, or if
756 there is no such callback, calls the I<base> version of the callback
757 with the same args, or if the f is invalid, set errno to EBADF and
760 Perl_PerlIO_or_fail(PerlIO* f, callback, failure, args) either calls
761 the I<callback> of the functions of the layer I<f> with the I<args>,
762 or if there is no such callback, set errno to EINVAL. Or if the f is
763 invalid, set errno to EBADF and return I<failure>.
765 Perl_PerlIO_or_Base_void(PerlIO* f, callback, base, args) either calls
766 the I<callback> of the functions of the layer I<f> with the I<args>,
767 or if there is no such callback, calls the I<base> version of the
768 callback with the same args, or if the f is invalid, set errno to
771 Perl_PerlIO_or_fail_void(PerlIO* f, callback, args) either calls the
772 I<callback> of the functions of the layer I<f> with the I<args>, or if
773 there is no such callback, set errno to EINVAL. Or if the f is
774 invalid, set errno to EBADF.
776 =head2 Implementing PerlIO Layers
778 If you find the implementation document unclear or not sufficient,
779 look at the existing PerlIO layer implementations, which include:
783 =item * C implementations
785 The F<perlio.c> and F<perliol.h> in the Perl core implement the
786 "unix", "perlio", "stdio", "crlf", "utf8", "byte", "raw", "pending"
787 layers, and also the "mmap" and "win32" layers if applicable.
788 (The "win32" is currently unfinished and unused, to see what is used
789 instead in Win32, see L<PerlIO/"Querying the layers of filehandles"> .)
791 PerlIO::encoding, PerlIO::scalar, PerlIO::via in the Perl core.
793 PerlIO::gzip and APR::PerlIO (mod_perl 2.0) on CPAN.
795 =item * Perl implementations
797 PerlIO::via::QuotedPrint in the Perl core and PerlIO::via::* on CPAN.
801 If you are creating a PerlIO layer, you may want to be lazy, in other
802 words, implement only the methods that interest you. The other methods
803 you can either replace with the "blank" methods
808 (which do nothing, and return zero and -1, respectively) or for
809 certain methods you may assume a default behaviour by using a NULL
810 method. The Open method looks for help in the 'parent' layer.
811 The following table summarizes the behaviour:
813 method behaviour with NULL
815 Clearerr PerlIOBase_clearerr
816 Close PerlIOBase_close
819 Error PerlIOBase_error
820 Fileno PerlIOBase_fileno
835 Setlinebuf PerlIOBase_setlinebuf
837 Unread PerlIOBase_unread
840 FAILURE Set errno (to EINVAL in Unixish, to LIB$_INVARG in VMS) and
841 return -1 (for numeric return values) or NULL (for pointers)
842 INHERITED Inherited from the layer below
843 SUCCESS Return 0 (for numeric return values) or a pointer
847 The file C<perlio.c> provides the following layers:
853 A basic non-buffered layer which calls Unix/POSIX C<read()>, C<write()>,
854 C<lseek()>, C<close()>. No buffering. Even on platforms that distinguish
855 between O_TEXT and O_BINARY this layer is always O_BINARY.
859 A very complete generic buffering layer which provides the whole of
860 PerlIO API. It is also intended to be used as a "base class" for other
861 layers. (For example its C<Read()> method is implemented in terms of
862 the C<Get_cnt()>/C<Get_ptr()>/C<Set_ptrcnt()> methods).
864 "perlio" over "unix" provides a complete replacement for stdio as seen
865 via PerlIO API. This is the default for USE_PERLIO when system's stdio
866 does not permit perl's "fast gets" access, and which do not
867 distinguish between C<O_TEXT> and C<O_BINARY>.
871 A layer which provides the PerlIO API via the layer scheme, but
872 implements it by calling system's stdio. This is (currently) the default
873 if system's stdio provides sufficient access to allow perl's "fast gets"
874 access and which do not distinguish between C<O_TEXT> and C<O_BINARY>.
878 A layer derived using "perlio" as a base class. It provides Win32-like
879 "\n" to CR,LF translation. Can either be applied above "perlio" or serve
880 as the buffer layer itself. "crlf" over "unix" is the default if system
881 distinguishes between C<O_TEXT> and C<O_BINARY> opens. (At some point
882 "unix" will be replaced by a "native" Win32 IO layer on that platform,
883 as Win32's read/write layer has various drawbacks.) The "crlf" layer is
884 a reasonable model for a layer which transforms data in some way.
888 If Configure detects C<mmap()> functions this layer is provided (with
889 "perlio" as a "base") which does "read" operations by mmap()ing the
890 file. Performance improvement is marginal on modern systems, so it is
891 mainly there as a proof of concept. It is likely to be unbundled from
892 the core at some point. The "mmap" layer is a reasonable model for a
893 minimalist "derived" layer.
897 An "internal" derivative of "perlio" which can be used to provide
898 Unread() function for layers which have no buffer or cannot be
899 bothered. (Basically this layer's C<Fill()> pops itself off the stack
900 and so resumes reading from layer below.)
904 A dummy layer which never exists on the layer stack. Instead when
905 "pushed" it actually pops the stack removing itself, it then calls
906 Binmode function table entry on all the layers in the stack - normally
907 this (via PerlIOBase_binmode) removes any layers which do not have
908 C<PERLIO_K_RAW> bit set. Layers can modify that behaviour by defining
909 their own Binmode entry.
913 Another dummy layer. When pushed it pops itself and sets the
914 C<PERLIO_F_UTF8> flag on the layer which was (and now is once more)
915 the top of the stack.
919 In addition F<perlio.c> also provides a number of C<PerlIOBase_xxxx()>
920 functions which are intended to be used in the table slots of classes
921 which do not need to do anything special for a particular method.
923 =head2 Extension Layers
925 Layers can be made available by extension modules. When an unknown layer
926 is encountered the PerlIO code will perform the equivalent of :
930 Where I<layer> is the unknown layer. F<PerlIO.pm> will then attempt to:
932 require PerlIO::layer;
934 If after that process the layer is still not defined then the C<open>
937 The following extension layers are bundled with perl:
945 makes this layer available, although F<PerlIO.pm> "knows" where to
946 find it. It is an example of a layer which takes an argument as it is
949 open( $fh, "<:encoding(iso-8859-7)", $pathname );
953 Provides support for reading data from and writing data to a scalar.
955 open( $fh, "+<:scalar", \$scalar );
957 When a handle is so opened, then reads get bytes from the string value
958 of I<$scalar>, and writes change the value. In both cases the position
959 in I<$scalar> starts as zero but can be altered via C<seek>, and
960 determined via C<tell>.
962 Please note that this layer is implied when calling open() thus:
964 open( $fh, "+<", \$scalar );
968 Provided to allow layers to be implemented as Perl code. For instance:
970 use PerlIO::via::StripHTML;
971 open( my $fh, "<:via(StripHTML)", "index.html" );
973 See L<PerlIO::via> for details.
979 Things that need to be done to improve this document.
985 Explain how to make a valid fh without going through open()(i.e. apply
986 a layer). For example if the file is not opened through perl, but we
987 want to get back a fh, like it was opened by Perl.
989 How PerlIO_apply_layera fits in, where its docs, was it made public?
991 Currently the example could be something like this:
993 PerlIO *foo_to_PerlIO(pTHX_ char *mode, ...)
995 char *mode; /* "w", "r", etc */
996 const char *layers = ":APR"; /* the layer name */
997 PerlIO *f = PerlIO_allocate(aTHX);
1002 PerlIO_apply_layers(aTHX_ f, mode, layers);
1005 PerlIOAPR *st = PerlIOSelf(f, PerlIOAPR);
1006 /* fill in the st struct, as in _open() */
1008 PerlIOBase(f)->flags |= PERLIO_F_OPEN;
1017 fix/add the documentation in places marked as XXX.
1021 The handling of errors by the layer is not specified. e.g. when $!
1022 should be set explicitly, when the error handling should be just
1023 delegated to the top layer.
1025 Probably give some hints on using SETERRNO() or pointers to where they
1030 I think it would help to give some concrete examples to make it easier
1031 to understand the API. Of course I agree that the API has to be
1032 concise, but since there is no second document that is more of a
1033 guide, I think that it'd make it easier to start with the doc which is
1034 an API, but has examples in it in places where things are unclear, to
1035 a person who is not a PerlIO guru (yet).