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 (and
16 =head2 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.
23 The aim of the implementation is to provide the PerlIO API in a flexible
24 and platform neutral manner. It is also a trial of an "Object Oriented
25 C, with vtables" approach which may be applied to perl6.
27 =head2 Layers vs Disciplines
29 Initial discussion of the ability to modify IO streams behaviour used
30 the term "discipline" for the entities which were added. This came (I
31 believe) from the use of the term in "sfio", which in turn borrowed it
32 from "line disciplines" on Unix terminals. However, this document (and
33 the C code) uses the term "layer".
35 This is, I hope, a natural term given the implementation, and should
36 avoid connotations that are inherent in earlier uses of "discipline"
37 for things which are rather different.
39 =head2 Data Structures
41 The basic data structure is a PerlIOl:
43 typedef struct _PerlIO PerlIOl;
44 typedef struct _PerlIO_funcs PerlIO_funcs;
45 typedef PerlIOl *PerlIO;
49 PerlIOl * next; /* Lower layer */
50 PerlIO_funcs * tab; /* Functions for this layer */
51 IV flags; /* Various flags for state */
54 A C<PerlIOl *> is a pointer to the struct, and the I<application>
55 level C<PerlIO *> is a pointer to a C<PerlIOl *> - i.e. a pointer
56 to a pointer to the struct. This allows the application level C<PerlIO *>
57 to remain constant while the actual C<PerlIOl *> underneath
58 changes. (Compare perl's C<SV *> which remains constant while its
59 C<sv_any> field changes as the scalar's type changes.) An IO stream is
60 then in general represented as a pointer to this linked-list of
63 It should be noted that because of the double indirection in a C<PerlIO *>,
64 a C<< &(perlio->next) >> "is" a C<PerlIO *>, and so to some degree
65 at least one layer can use the "standard" API on the next layer down.
67 A "layer" is composed of two parts:
73 The functions and attributes of the "layer class".
77 The per-instance data for a particular handle.
81 =head2 Functions and Attributes
83 The functions and attributes are accessed via the "tab" (for table)
84 member of C<PerlIOl>. The functions (methods of the layer "class") are
85 fixed, and are defined by the C<PerlIO_funcs> type. They are broadly the
86 same as the public C<PerlIO_xxxxx> functions:
93 IV (*Pushed)(pTHX_ PerlIO *f,const char *mode,SV *arg);
94 IV (*Popped)(pTHX_ PerlIO *f);
95 PerlIO * (*Open)(pTHX_ PerlIO_funcs *tab,
98 int fd, int imode, int perm,
100 int narg, SV **args);
101 SV * (*Getarg)(pTHX_ PerlIO *f, CLONE_PARAMS *param, int flags)
102 IV (*Fileno)(pTHX_ PerlIO *f);
103 PerlIO * (*Dup)(pTHX_ PerlIO *f, PerlIO *o, CLONE_PARAMS *param, int flags)
104 /* Unix-like functions - cf sfio line disciplines */
105 SSize_t (*Read)(pTHX_ PerlIO *f, void *vbuf, Size_t count);
106 SSize_t (*Unread)(pTHX_ PerlIO *f, const void *vbuf, Size_t count);
107 SSize_t (*Write)(pTHX_ PerlIO *f, const void *vbuf, Size_t count);
108 IV (*Seek)(pTHX_ PerlIO *f, Off_t offset, int whence);
109 Off_t (*Tell)(pTHX_ PerlIO *f);
110 IV (*Close)(pTHX_ PerlIO *f);
111 /* Stdio-like buffered IO functions */
112 IV (*Flush)(pTHX_ PerlIO *f);
113 IV (*Fill)(pTHX_ PerlIO *f);
114 IV (*Eof)(pTHX_ PerlIO *f);
115 IV (*Error)(pTHX_ PerlIO *f);
116 void (*Clearerr)(pTHX_ PerlIO *f);
117 void (*Setlinebuf)(pTHX_ PerlIO *f);
118 /* Perl's snooping functions */
119 STDCHAR * (*Get_base)(pTHX_ PerlIO *f);
120 Size_t (*Get_bufsiz)(pTHX_ PerlIO *f);
121 STDCHAR * (*Get_ptr)(pTHX_ PerlIO *f);
122 SSize_t (*Get_cnt)(pTHX_ PerlIO *f);
123 void (*Set_ptrcnt)(pTHX_ PerlIO *f,STDCHAR *ptr,SSize_t cnt);
126 The first few members of the struct give a "name" for the layer, the
127 size to C<malloc> for the per-instance data, and some flags which are
128 attributes of the class as whole (such as whether it is a buffering
129 layer), then follow the functions which fall into four basic groups:
135 Opening and setup functions
143 Stdio class buffering options.
147 Functions to support Perl's traditional "fast" access to the buffer.
151 A layer does not have to implement all the functions, but the whole
152 table has to be present. Unimplemented slots can be NULL (which will
153 result in an error when called) or can be filled in with stubs to
154 "inherit" behaviour from a "base class". This "inheritance" is fixed
155 for all instances of the layer, but as the layer chooses which stubs
156 to populate the table, limited "multiple inheritance" is possible.
158 =head2 Per-instance Data
160 The per-instance data are held in memory beyond the basic PerlIOl
161 struct, by making a PerlIOl the first member of the layer's struct
166 struct _PerlIO base; /* Base "class" info */
167 STDCHAR * buf; /* Start of buffer */
168 STDCHAR * end; /* End of valid part of buffer */
169 STDCHAR * ptr; /* Current position in buffer */
170 Off_t posn; /* Offset of buf into the file */
171 Size_t bufsiz; /* Real size of buffer */
172 IV oneword; /* Emergency buffer */
175 In this way (as for perl's scalars) a pointer to a PerlIOBuf can be
176 treated as a pointer to a PerlIOl.
178 =head2 Layers in action.
182 +-----------+ +----------+ +--------+
183 PerlIO ->| |--->| next |--->| NULL |
184 +-----------+ +----------+ +--------+
185 | | | buffer | | fd |
186 +-----------+ | | +--------+
190 The above attempts to show how the layer scheme works in a simple case.
191 The application's C<PerlIO *> points to an entry in the table(s)
192 representing open (allocated) handles. For example the first three slots
193 in the table correspond to C<stdin>,C<stdout> and C<stderr>. The table
194 in turn points to the current "top" layer for the handle - in this case
195 an instance of the generic buffering layer "perlio". That layer in turn
196 points to the next layer down - in this case the lowlevel "unix" layer.
198 The above is roughly equivalent to a "stdio" buffered stream, but with
199 much more flexibility:
205 If Unix level C<read>/C<write>/C<lseek> is not appropriate for (say)
206 sockets then the "unix" layer can be replaced (at open time or even
207 dynamically) with a "socket" layer.
211 Different handles can have different buffering schemes. The "top"
212 layer could be the "mmap" layer if reading disk files was quicker
213 using C<mmap> than C<read>. An "unbuffered" stream can be implemented
214 simply by not having a buffer layer.
218 Extra layers can be inserted to process the data as it flows through.
219 This was the driving need for including the scheme in perl 5.7.0+ - we
220 needed a mechanism to allow data to be translated between perl's
221 internal encoding (conceptually at least Unicode as UTF-8), and the
222 "native" format used by the system. This is provided by the
223 ":encoding(xxxx)" layer which typically sits above the buffering layer.
227 A layer can be added that does "\n" to CRLF translation. This layer
228 can be used on any platform, not just those that normally do such
233 =head2 Per-instance flag bits
235 The generic flag bits are a hybrid of C<O_XXXXX> style flags deduced
236 from the mode string passed to C<PerlIO_open()>, and state bits for
237 typical buffer layers.
245 =item PERLIO_F_CANWRITE
247 Writes are permitted, i.e. opened as "w" or "r+" or "a", etc.
249 =item PERLIO_F_CANREAD
251 Reads are permitted i.e. opened "r" or "w+" (or even "a+" - ick).
255 An error has occurred (for C<PerlIO_error()>).
257 =item PERLIO_F_TRUNCATE
259 Truncate file suggested by open mode.
261 =item PERLIO_F_APPEND
263 All writes should be appends.
267 Layer is performing Win32-like "\n" mapped to CR,LF for output and CR,LF
268 mapped to "\n" for input. Normally the provided "crlf" layer is the only
269 layer that need bother about this. C<PerlIO_binmode()> will mess with this
270 flag rather than add/remove layers if the C<PERLIO_K_CANCRLF> bit is set
271 for the layers class.
275 Data written to this layer should be UTF-8 encoded; data provided
276 by this layer should be considered UTF-8 encoded. Can be set on any layer
277 by ":utf8" dummy layer. Also set on ":encoding" layer.
281 Layer is unbuffered - i.e. write to next layer down should occur for
282 each write to this layer.
286 The buffer for this layer currently holds data written to it but not sent
291 The buffer for this layer currently holds unconsumed data read from
294 =item PERLIO_F_LINEBUF
296 Layer is line buffered. Write data should be passed to next layer down
297 whenever a "\n" is seen. Any data beyond the "\n" should then be
302 File has been C<unlink()>ed, or should be deleted on C<close()>.
308 =item PERLIO_F_FASTGETS
310 This instance of this layer supports the "fast C<gets>" interface.
311 Normally set based on C<PERLIO_K_FASTGETS> for the class and by the
312 existence of the function(s) in the table. However a class that
313 normally provides that interface may need to avoid it on a
314 particular instance. The "pending" layer needs to do this when
315 it is pushed above a layer which does not support the interface.
316 (Perl's C<sv_gets()> does not expect the streams fast C<gets> behaviour
317 to change during one "get".)
321 =head2 Methods in Detail
327 The name of the layer whose open() method Perl should invoke on
328 open(). For example if the layer is called APR, you will call:
330 open $fh, ">:APR", ...
332 and Perl knows that it has to invoke the PerlIOAPR_open() method
333 implemented by the APR layer.
337 The size of the per-instance data structure, e.g.:
343 XXX: explain all the available flags here
347 =item * PERLIO_K_BUFFERED
349 =item * PERLIO_K_CANCRLF
351 =item * PERLIO_K_FASTGETS
353 =item * PERLIO_K_MULTIARG
355 Used when the layer's open() accepts more arguments than usual. The
356 extra arguments should come not before the C<MODE> argument. When this
357 flag is used it's up to the layer to validate the args.
363 =item IV (*Pushed)(pTHX_ PerlIO *f,const char *mode, SV *arg);
365 The only absolutely mandatory method. Called when the layer is pushed
366 onto the stack. The C<mode> argument may be NULL if this occurs
367 post-open. The C<arg> will be non-C<NULL> if an argument string was
368 passed. In most cases this should call C<PerlIOBase_pushed()> to
369 convert C<mode> into the appropriate C<PERLIO_F_XXXXX> flags in
370 addition to any actions the layer itself takes. If a layer is not
371 expecting an argument it need neither save the one passed to it, nor
372 provide C<Getarg()> (it could perhaps C<Perl_warn> that the argument
375 Returns 0 on success. On failure returns -1 and should set errno.
377 =item IV (*Popped)(pTHX_ PerlIO *f);
379 Called when the layer is popped from the stack. A layer will normally
380 be popped after C<Close()> is called. But a layer can be popped
381 without being closed if the program is dynamically managing layers on
382 the stream. In such cases C<Popped()> should free any resources
383 (buffers, translation tables, ...) not held directly in the layer's
384 struct. It should also C<Unread()> any unconsumed data that has been
385 read and buffered from the layer below back to that layer, so that it
386 can be re-provided to what ever is now above.
388 Returns 0 on success and failure.
390 =item PerlIO * (*Open)(...);
392 The C<Open()> method has lots of arguments because it combines the
393 functions of perl's C<open>, C<PerlIO_open>, perl's C<sysopen>,
394 C<PerlIO_fdopen> and C<PerlIO_reopen>. The full prototype is as
397 PerlIO * (*Open)(pTHX_ PerlIO_funcs *tab,
400 int fd, int imode, int perm,
402 int narg, SV **args);
404 Open should (perhaps indirectly) call C<PerlIO_allocate()> to allocate
405 a slot in the table and associate it with the layers information for
406 the opened file, by calling C<PerlIO_push>. The I<layers> AV is an
407 array of all the layers destined for the C<PerlIO *>, and any
408 arguments passed to them, I<n> is the index into that array of the
409 layer being called. The macro C<PerlIOArg> will return a (possibly
410 C<NULL>) SV * for the argument passed to the layer.
412 The I<mode> string is an "C<fopen()>-like" string which would match
413 the regular expression C</^[I#]?[rwa]\+?[bt]?$/>.
415 The C<'I'> prefix is used during creation of C<stdin>..C<stderr> via
416 special C<PerlIO_fdopen> calls; the C<'#'> prefix means that this is
417 C<sysopen> and that I<imode> and I<perm> should be passed to
418 C<PerlLIO_open3>; C<'r'> means B<r>ead, C<'w'> means B<w>rite and
419 C<'a'> means B<a>ppend. The C<'+'> suffix means that both reading and
420 writing/appending are permitted. The C<'b'> suffix means file should
421 be binary, and C<'t'> means it is text. (Binary/Text should be ignored
422 by almost all layers and binary IO done, with PerlIO. The C<:crlf>
423 layer should be pushed to handle the distinction.)
425 If I<old> is not C<NULL> then this is a C<PerlIO_reopen>. Perl itself
426 does not use this (yet?) and semantics are a little vague.
428 If I<fd> not negative then it is the numeric file descriptor I<fd>,
429 which will be open in a manner compatible with the supplied mode
430 string, the call is thus equivalent to C<PerlIO_fdopen>. In this case
431 I<nargs> will be zero.
433 If I<nargs> is greater than zero then it gives the number of arguments
434 passed to C<open>, otherwise it will be 1 if for example
435 C<PerlIO_open> was called. In simple cases SvPV_nolen(*args) is the
438 Having said all that translation-only layers do not need to provide
439 C<Open()> at all, but rather leave the opening to a lower level layer
440 and wait to be "pushed". If a layer does provide C<Open()> it should
441 normally call the C<Open()> method of next layer down (if any) and
442 then push itself on top if that succeeds.
444 Returns C<NULL> on failure.
446 =item SV * (*Getarg)(pTHX_ PerlIO *f, CLONE_PARAMS *param, int flags)
448 Optional. If present should return an SV * representing the string
449 argument passed to the layer when it was
450 pushed. e.g. ":encoding(ascii)" would return an SvPV with value
451 "ascii". (I<param> and I<flags> arguments can be ignored in most
454 =item IV (*Fileno)(pTHX_ PerlIO *f);
456 Returns the Unix/Posix numeric file descriptor for the handle. Normally
457 C<PerlIOBase_fileno()> (which just asks next layer down) will suffice
460 Returns -1 if the layer cannot provide such a file descriptor, or in
461 the case of the error.
463 XXX: two possible results end up in -1, one is an error the other is
466 =item PerlIO * (*Dup)(pTHX_ PerlIO *f, PerlIO *o, CLONE_PARAMS *param, int flags)
470 Similar to C<Open>, returns PerlIO* on success, C<NULL> on failure.
472 =item SSize_t (*Read)(pTHX_ PerlIO *f, void *vbuf, Size_t count);
474 Basic read operation.
476 Typically will call C<Fill> and manipulate pointers (possibly via the
477 API). C<PerlIOBuf_read()> may be suitable for derived classes which
478 provide "fast gets" methods.
480 Returns actual bytes read, or -1 on an error.
482 =item SSize_t (*Unread)(pTHX_ PerlIO *f, const void *vbuf, Size_t count);
484 A superset of stdio's C<ungetc()>. Should arrange for future reads to
485 see the bytes in C<vbuf>. If there is no obviously better implementation
486 then C<PerlIOBase_unread()> provides the function by pushing a "fake"
487 "pending" layer above the calling layer.
489 Returns the number of unread chars.
491 =item SSize_t (*Write)(PerlIO *f, const void *vbuf, Size_t count);
493 Basic write operation.
495 Returns bytes written or -1 on an error.
497 =item IV (*Seek)(pTHX_ PerlIO *f, Off_t offset, int whence);
499 Position the file pointer. Should normally call its own C<Flush>
500 method and then the C<Seek> method of next layer down.
502 Returns 0 on success, -1 on failure.
504 =item Off_t (*Tell)(pTHX_ PerlIO *f);
506 Return the file pointer. May be based on layers cached concept of
507 position to avoid overhead.
509 Returns -1 on failure to get the file pointer.
511 =item IV (*Close)(pTHX_ PerlIO *f);
513 Close the stream. Should normally call C<PerlIOBase_close()> to flush
514 itself and close layers below, and then deallocate any data structures
515 (buffers, translation tables, ...) not held directly in the data
518 Returns 0 on success, -1 on failure.
520 =item IV (*Flush)(pTHX_ PerlIO *f);
522 Should make stream's state consistent with layers below. That is, any
523 buffered write data should be written, and file position of lower layers
524 adjusted for data read from below but not actually consumed.
525 (Should perhaps C<Unread()> such data to the lower layer.)
527 Returns 0 on success, -1 on failure.
529 =item IV (*Fill)(pTHX_ PerlIO *f);
531 The buffer for this layer should be filled (for read) from layer
532 below. When you "subclass" PerlIOBuf layer, you want to use its
533 I<_read> method and to supply your own fill method, which fills the
536 Returns 0 on success, -1 on failure.
538 =item IV (*Eof)(pTHX_ PerlIO *f);
540 Return end-of-file indicator. C<PerlIOBase_eof()> is normally sufficient.
542 Returns 0 on end-of-file, 1 if not end-of-file, -1 on error.
544 =item IV (*Error)(pTHX_ PerlIO *f);
546 Return error indicator. C<PerlIOBase_error()> is normally sufficient.
548 Returns 1 if there is an error (usually when C<PERLIO_F_ERROR> is set,
551 =item void (*Clearerr)(pTHX_ PerlIO *f);
553 Clear end-of-file and error indicators. Should call C<PerlIOBase_clearerr()>
554 to set the C<PERLIO_F_XXXXX> flags, which may suffice.
556 =item void (*Setlinebuf)(pTHX_ PerlIO *f);
558 Mark the stream as line buffered. C<PerlIOBase_setlinebuf()> sets the
559 PERLIO_F_LINEBUF flag and is normally sufficient.
561 =item STDCHAR * (*Get_base)(pTHX_ PerlIO *f);
563 Allocate (if not already done so) the read buffer for this layer and
564 return pointer to it. Return NULL on failure.
566 =item Size_t (*Get_bufsiz)(pTHX_ PerlIO *f);
568 Return the number of bytes that last C<Fill()> put in the buffer.
570 =item STDCHAR * (*Get_ptr)(pTHX_ PerlIO *f);
572 Return the current read pointer relative to this layer's buffer.
574 =item SSize_t (*Get_cnt)(pTHX_ PerlIO *f);
576 Return the number of bytes left to be read in the current buffer.
578 =item void (*Set_ptrcnt)(pTHX_ PerlIO *f,STDCHAR *ptr,SSize_t cnt);
580 Adjust the read pointer and count of bytes to match C<ptr> and/or C<cnt>.
581 The application (or layer above) must ensure they are consistent.
582 (Checking is allowed by the paranoid.)
589 The file C<perlio.c> provides the following layers:
595 A basic non-buffered layer which calls Unix/POSIX C<read()>, C<write()>,
596 C<lseek()>, C<close()>. No buffering. Even on platforms that distinguish
597 between O_TEXT and O_BINARY this layer is always O_BINARY.
601 A very complete generic buffering layer which provides the whole of
602 PerlIO API. It is also intended to be used as a "base class" for other
603 layers. (For example its C<Read()> method is implemented in terms of
604 the C<Get_cnt()>/C<Get_ptr()>/C<Set_ptrcnt()> methods).
606 "perlio" over "unix" provides a complete replacement for stdio as seen
607 via PerlIO API. This is the default for USE_PERLIO when system's stdio
608 does not permit perl's "fast gets" access, and which do not
609 distinguish between C<O_TEXT> and C<O_BINARY>.
613 A layer which provides the PerlIO API via the layer scheme, but
614 implements it by calling system's stdio. This is (currently) the default
615 if system's stdio provides sufficient access to allow perl's "fast gets"
616 access and which do not distinguish between C<O_TEXT> and C<O_BINARY>.
620 A layer derived using "perlio" as a base class. It provides Win32-like
621 "\n" to CR,LF translation. Can either be applied above "perlio" or serve
622 as the buffer layer itself. "crlf" over "unix" is the default if system
623 distinguishes between C<O_TEXT> and C<O_BINARY> opens. (At some point
624 "unix" will be replaced by a "native" Win32 IO layer on that platform,
625 as Win32's read/write layer has various drawbacks.) The "crlf" layer is
626 a reasonable model for a layer which transforms data in some way.
630 If Configure detects C<mmap()> functions this layer is provided (with
631 "perlio" as a "base") which does "read" operations by mmap()ing the
632 file. Performance improvement is marginal on modern systems, so it is
633 mainly there as a proof of concept. It is likely to be unbundled from
634 the core at some point. The "mmap" layer is a reasonable model for a
635 minimalist "derived" layer.
639 An "internal" derivative of "perlio" which can be used to provide
640 Unread() function for layers which have no buffer or cannot be
641 bothered. (Basically this layer's C<Fill()> pops itself off the stack
642 and so resumes reading from layer below.)
646 A dummy layer which never exists on the layer stack. Instead when
647 "pushed" it actually pops the stack(!), removing itself, and any other
648 layers until it reaches a layer with the class C<PERLIO_K_RAW> bit set.
652 Another dummy layer. When pushed it pops itself and sets the
653 C<PERLIO_F_UTF8> flag on the layer which was (and now is once more)
654 the top of the stack.
658 In addition F<perlio.c> also provides a number of C<PerlIOBase_xxxx()>
659 functions which are intended to be used in the table slots of classes
660 which do not need to do anything special for a particular method.
662 =head2 Extension Layers
664 Layers can made available by extension modules. When an unknown layer
665 is encountered the PerlIO code will perform the equivalent of :
669 Where I<layer> is the unknown layer. F<PerlIO.pm> will then attempt to:
671 require PerlIO::layer;
673 If after that process the layer is still not defined then the C<open>
676 The following extension layers are bundled with perl:
684 makes this layer available, although F<PerlIO.pm> "knows" where to
685 find it. It is an example of a layer which takes an argument as it is
688 open($fh,"<:encoding(iso-8859-7)",$pathname)
694 open($fh,"...",\$scalar)
696 When a handle is so opened, then reads get bytes from the string value
697 of I<$scalar>, and writes change the value. In both cases the position
698 in I<$scalar> starts as zero but can be altered via C<seek>, and
699 determined via C<tell>.
701 =item ":Object" or ":Perl"
703 May be provided to allow layers to be implemented as perl code -
704 implementation is being investigated.
710 Things that need to be done to improve this document.
716 Explain how to make a valid fh without going through open()(i.e. apply
717 a layer). For example if the file is not opened through perl, but we
718 want to get back a fh, like it was opened by Perl.
720 How PerlIO_apply_layera fits in, where its docs, was it made public?
722 Currently the example could be something like this:
724 PerlIO *foo_to_PerlIO(pTHX_ char *mode, ...)
726 char *mode; /* "w", "r", etc */
727 const char *layers = ":APR"; /* the layer name */
728 PerlIO *f = PerlIO_allocate(aTHX);
733 PerlIO_apply_layers(aTHX_ f, mode, layers);
736 PerlIOAPR *st = PerlIOSelf(f, PerlIOAPR);
737 /* fill in the st struct, as in _open() */
739 PerlIOBase(f)->flags |= PERLIO_F_OPEN;
748 fix/add the documentation in places marked as XXX.
752 The handling of errors by the layer is not specified. e.g. when $!
753 should be set explicitly, when the error handling should be just
754 delegated to the top layer.
756 Probably give some hints on using SETERRNO() or pointers to where they
761 I think it would help to give some concrete examples to make it easier
762 to understand the API. Of course I agree that the API has to be
763 concise, but since there is no second document that is more of a
764 guide, I think that it'd make it easier to start with the doc which is
765 an API, but has examples in it in places where things are unclear, to
766 a person who is not a PerlIO guru (yet).