1PerlIO(3pm) Perl Programmers Reference Guide PerlIO(3pm)
2
6 PerlIO - On demand loader for PerlIO layers and root of PerlIO::* name
7 space
8
10 # support platform-native and CRLF text files
11 open(my $fh, "<:crlf", "my.txt") or die "open failed: $!";
12 # append UTF-8 encoded text
14 open(my $fh, ">>:encoding(UTF-8)", "some.log")
15 or die "open failed: $!";
16 # portably open a binary file for reading
18 open(my $fh, "<", "his.jpg") or die "open failed: $!";
19 binmode($fh) or die "binmode failed: $!";
20 Shell:
22 PERLIO=:perlio perl ....
23
25 When an undefined layer 'foo' is encountered in an "open" or "binmode"
26 layer specification then C code performs the equivalent of:
27 use PerlIO 'foo';
29 The Perl code in PerlIO.pm then attempts to locate a layer by doing
31 require PerlIO::foo;
33 Otherwise the "PerlIO" package is a place holder for additional PerlIO
35 related functions.
36 Layers
38 Generally speaking, PerlIO layers (previously sometimes referred to as
39 "disciplines") are an ordered stack applied to a filehandle (specified
40 as a space- or colon-separated list, conventionally written with a
41 leading colon). Each layer performs some operation on any input or
42 output, except when bypassed such as with "sysread" or "syswrite".
43 Read operations go through the stack in the order they are set (left to
44 right), and write operations in the reverse order.
45 There are also layers which actually just set flags on lower layers, or
47 layers that modify the current stack but don't persist on the stack
48 themselves; these are referred to as pseudo-layers.
49 When opening a handle, it will be opened with any layers specified
51 explicitly in the open() call (or the platform defaults, if specified
52 as a colon with no following layers).
53 If layers are not explicitly specified, the handle will be opened with
55 the layers specified by the ${^OPEN} variable (usually set by using the
56 open pragma for a lexical scope, or the "-C" command-line switch or
57 "PERL_UNICODE" environment variable for the main program scope).
58 If layers are not specified in the open() call or "${^OPEN}" variable,
60 the handle will be opened with the default layer stack configured for
61 that architecture; see "Defaults and how to override them".
62 Some layers will automatically insert required lower level layers if
64 not present; for example ":perlio" will insert ":unix" below itself for
65 low level IO, and ":encoding" will insert the platform defaults for
66 buffered IO.
67 The "binmode" function can be called on an opened handle to push
69 additional layers onto the stack, which may also modify the existing
70 layers. "binmode" called with no layers will remove or unset any
71 existing layers which transform the byte stream, making the handle
72 suitable for binary data.
73 The following layers are currently defined:
75 :unix
77 Lowest level layer which provides basic PerlIO operations in terms
78 of UNIX/POSIX numeric file descriptor calls (open(), read(),
79 write(), lseek(), close()). It is used even on non-Unix
80 architectures, and most other layers operate on top of it.
81 :stdio
83 Layer which calls "fread", "fwrite" and "fseek"/"ftell" etc. Note
84 that as this is "real" stdio it will ignore any layers beneath it
85 and go straight to the operating system via the C library as usual.
86 This layer implements both low level IO and buffering, but is
87 rarely used on modern architectures.
88 :perlio
90 A from scratch implementation of buffering for PerlIO. Provides
91 fast access to the buffer for "sv_gets" which implements Perl's
92 readline/<> and in general attempts to minimize data copying.
93 ":perlio" will insert a ":unix" layer below itself to do low level
95 IO.
96 :crlf
98 A layer that implements DOS/Windows like CRLF line endings. On
99 read converts pairs of CR,LF to a single "\n" newline character.
100 On write converts each "\n" to a CR,LF pair. Note that this layer
101 will silently refuse to be pushed on top of itself.
102 It currently does not mimic MS-DOS as far as treating of Control-Z
104 as being an end-of-file marker.
105 On DOS/Windows like architectures where this layer is part of the
107 defaults, it also acts like the ":perlio" layer, and removing the
108 CRLF translation (such as with ":raw") will only unset the CRLF
109 translation flag. Since Perl 5.14, you can also apply another
110 ":crlf" layer later, such as when the CRLF translation must occur
111 after an encoding layer. On other architectures, it is a mundane
112 CRLF translation layer and can be added and removed normally.
113 # translate CRLF after encoding on Perl 5.14 or newer
115 binmode $fh, ":raw:encoding(UTF-16LE):crlf"
116 or die "binmode failed: $!";
117 :utf8
119 Pseudo-layer that declares that the stream accepts Perl's internal
120 upgraded encoding of characters, which is approximately UTF-8 on
121 ASCII machines, but UTF-EBCDIC on EBCDIC machines. This allows any
122 character Perl can represent to be read from or written to the
123 stream.
124 This layer (which actually sets a flag on the preceding layer, and
126 is implicitly set by any ":encoding" layer) does not translate or
127 validate byte sequences. It instead indicates that the byte stream
128 will have been arranged by other layers to be provided in Perl's
129 internal upgraded encoding, which Perl code (and correctly written
130 XS code) will interpret as decoded Unicode characters.
131 CAUTION: Do not use this layer to translate from UTF-8 bytes, as
133 invalid UTF-8 or binary data will result in malformed Perl strings.
134 It is unlikely to produce invalid UTF-8 when used for output,
135 though it will instead produce UTF-EBCDIC on EBCDIC systems. The
136 ":encoding(UTF-8)" layer (hyphen is significant) is preferred as it
137 will ensure translation between valid UTF-8 bytes and valid Unicode
138 characters.
139 :bytes
141 This is the inverse of the ":utf8" pseudo-layer. It turns off the
142 flag on the layer below so that data read from it is considered to
143 be Perl's internal downgraded encoding, thus interpreted as the
144 native single-byte encoding of Latin-1 or EBCDIC. Likewise on
145 output Perl will warn if a "wide" character (a codepoint not in the
146 range 0..255) is written to a such a stream.
147 This is very dangerous to push on a handle using an ":encoding"
149 layer, as such a layer assumes to be working with Perl's internal
150 upgraded encoding, so you will likely get a mangled result.
151 Instead use ":raw" or ":pop" to remove encoding layers.
152 :raw
154 The ":raw" pseudo-layer is defined as being identical to calling
155 "binmode($fh)" - the stream is made suitable for passing binary
156 data, i.e. each byte is passed as-is. The stream will still be
157 buffered (but this was not always true before Perl 5.14).
158 In Perl 5.6 and some books the ":raw" layer is documented as the
160 inverse of the ":crlf" layer. That is no longer the case - other
161 layers which would alter the binary nature of the stream are also
162 disabled. If you want UNIX line endings on a platform that
163 normally does CRLF translation, but still want UTF-8 or encoding
164 defaults, the appropriate thing to do is to add ":perlio" to the
165 PERLIO environment variable, or open the handle explicitly with
166 that layer, to replace the platform default of ":crlf".
167 The implementation of ":raw" is as a pseudo-layer which when
169 "pushed" pops itself and then any layers which would modify the
170 binary data stream. (Undoing ":utf8" and ":crlf" may be
171 implemented by clearing flags rather than popping layers but that
172 is an implementation detail.)
173 As a consequence of the fact that ":raw" normally pops layers, it
175 usually only makes sense to have it as the only or first element in
176 a layer specification. When used as the first element it provides
177 a known base on which to build e.g.
178 open(my $fh,">:raw:encoding(UTF-8)",...)
180 or die "open failed: $!";
181 will construct a "binary" stream regardless of the platform
183 defaults, but then enable UTF-8 translation.
184 :pop
186 A pseudo-layer that removes the top-most layer. Gives Perl code a
187 way to manipulate the layer stack. Note that ":pop" only works on
188 real layers and will not undo the effects of pseudo-layers or flags
189 like ":utf8". An example of a possible use might be:
190 open(my $fh,...) or die "open failed: $!";
192 ...
193 binmode($fh,":encoding(...)") or die "binmode failed: $!";
194 # next chunk is encoded
195 ...
196 binmode($fh,":pop") or die "binmode failed: $!";
197 # back to un-encoded
198 A more elegant (and safer) interface is needed.
200 :win32
202 On Win32 platforms this experimental layer uses the native "handle"
203 IO rather than the unix-like numeric file descriptor layer. Known
204 to be buggy as of Perl 5.8.2.
205 Custom Layers
207 It is possible to write custom layers in addition to the above builtin
208 ones, both in C/XS and Perl, as a module named "PerlIO::<layer name>".
209 Some custom layers come with the Perl distribution.
210 :encoding
212 Use ":encoding(ENCODING)" to transparently do character set and
213 encoding transformations, for example from Shift-JIS to Unicode.
214 Note that an ":encoding" also enables ":utf8". See
215 PerlIO::encoding for more information.
216 :mmap
218 A layer which implements "reading" of files by using "mmap()" to
219 make a (whole) file appear in the process's address space, and then
220 using that as PerlIO's "buffer". This may be faster in certain
221 circumstances for large files, and may result in less physical
222 memory use when multiple processes are reading the same file.
223 Files which are not "mmap()"-able revert to behaving like the
225 ":perlio" layer. Writes also behave like the ":perlio" layer, as
226 "mmap()" for write needs extra house-keeping (to extend the file)
227 which negates any advantage.
228 The ":mmap" layer will not exist if the platform does not support
230 "mmap()". See PerlIO::mmap for more information.
231 :via
233 ":via(MODULE)" allows a transformation to be applied by an
234 arbitrary Perl module, for example compression / decompression,
235 encryption / decryption. See PerlIO::via for more information.
236 :scalar
238 A layer implementing "in memory" files using scalar variables,
239 automatically used in place of the platform defaults for IO when
240 opening such a handle. As such, the scalar is expected to act like
241 a file, only containing or storing bytes. See PerlIO::scalar for
242 more information.
243 Alternatives to raw
245 To get a binary stream an alternate method is to use:
246 open(my $fh,"<","whatever") or die "open failed: $!";
248 binmode($fh) or die "binmode failed: $!";
249 This has the advantage of being backward compatible with older versions
251 of Perl that did not use PerlIO or where ":raw" was buggy (as it was
252 before Perl 5.14).
253 To get an unbuffered stream specify an unbuffered layer (e.g. ":unix")
255 in the open call:
256 open(my $fh,"<:unix",$path) or die "open failed: $!";
258 Defaults and how to override them
260 If the platform is MS-DOS like and normally does CRLF to "\n"
261 translation for text files then the default layers are:
262 :unix:crlf
264 Otherwise if "Configure" found out how to do "fast" IO using the
266 system's stdio (not common on modern architectures), then the default
267 layers are:
268 :stdio
270 Otherwise the default layers are
272 :unix:perlio
274 Note that the "default stack" depends on the operating system and on
276 the Perl version, and both the compile-time and runtime configurations
277 of Perl. The default can be overridden by setting the environment
278 variable PERLIO to a space or colon separated list of layers, however
279 this cannot be used to set layers that require loading modules like
280 ":encoding".
281 This can be used to see the effect of/bugs in the various layers e.g.
283 cd .../perl/t
285 PERLIO=:stdio ./perl harness
286 PERLIO=:perlio ./perl harness
287 For the various values of PERLIO see "PERLIO" in perlrun.
289 The following table summarizes the default layers on UNIX-like and DOS-
291 like platforms and depending on the setting of $ENV{PERLIO}:
292 PERLIO UNIX-like DOS-like
294 ------ --------- --------
295 unset / "" :unix:perlio / :stdio [1] :unix:crlf
296 :stdio :stdio :stdio
297 :perlio :unix:perlio :unix:perlio
298 # [1] ":stdio" if Configure found out how to do "fast stdio" (depends
300 # on the stdio implementation) and in Perl 5.8, else ":unix:perlio"
301 Querying the layers of filehandles
303 The following returns the names of the PerlIO layers on a filehandle.
304 my @layers = PerlIO::get_layers($fh); # Or FH, *FH, "FH".
306 The layers are returned in the order an open() or binmode() call would
308 use them, and without colons.
309 By default the layers from the input side of the filehandle are
311 returned; to get the output side, use the optional "output" argument:
312 my @layers = PerlIO::get_layers($fh, output => 1);
314 (Usually the layers are identical on either side of a filehandle but
316 for example with sockets there may be differences.)
317 There is no set_layers(), nor does get_layers() return a tied array
319 mirroring the stack, or anything fancy like that. This is not
320 accidental or unintentional. The PerlIO layer stack is a bit more
321 complicated than just a stack (see for example the behaviour of
322 ":raw"). You are supposed to use open() and binmode() to manipulate
323 the stack.
324 Implementation details follow, please close your eyes.
326 The arguments to layers are by default returned in parentheses after
328 the name of the layer, and certain layers (like ":utf8") are not real
329 layers but instead flags on real layers; to get all of these returned
330 separately, use the optional "details" argument:
331 my @layer_and_args_and_flags = PerlIO::get_layers($fh, details => 1);
333 The result will be up to be three times the number of layers: the first
335 element will be a name, the second element the arguments (unspecified
336 arguments will be "undef"), the third element the flags, the fourth
337 element a name again, and so forth.
338 You may open your eyes now.
340
342 Nick Ing-Simmons <nick@ing-simmons.net>
343
345 "binmode" in perlfunc, "open" in perlfunc, perlunicode, perliol, Encode
346perl v5.32.1 2021-03-31 PerlIO(3pm)