996
|
1 CLC-INTERCAL 1.-94.-2
|
|
2
|
|
3 "human beings are given free will in order to choose between
|
|
4 insanity on the one hand and lunacy on the other"
|
|
5 Aldous Huxley
|
|
6
|
|
7 -- - --------
|
|
8 -2 - Contents
|
|
9 -- - --------
|
|
10
|
|
11 -2 - Contents
|
|
12 -1 - Foreword
|
|
13 0 - Introduction to CLC-INTERCAL pre-pre-escape 1.-94.-2
|
|
14 1 - What has changed since CLC-INTERCAL 0.05
|
|
15 2 - How to install this escape
|
|
16 3 - How to modify the compiler(s)
|
|
17 4 - Important notes about the new INTERNET library
|
|
18 5 - Licence
|
|
19
|
|
20
|
|
21 -- - --------
|
|
22 -1 - Foreword
|
|
23 -- - --------
|
|
24
|
|
25 A few years ago, I had, for several reasons, decided to completely GIVE UP
|
|
26 working on INTERCAL or, if you prefer, ABSTAIN FROM developing CLC-INTERCAL
|
|
27 any farther.
|
|
28
|
|
29 Of course, the flesh is weak and here I am at it again. It just so happens
|
|
30 that I am using the "compiler compiler compiler compiler" for a serious
|
|
31 project, so I am going to develop it further.
|
|
32
|
|
33
|
|
34 - - ------------ -- ------------ -------------- --------
|
|
35 0 - Introduction to CLC-INTERCAL pre-pre-escape 1.-94.-2
|
|
36 - - ------------ -- ------------ -------------- --------
|
|
37
|
|
38 It has been decided during the development of CLC-INTERCAL 0.06 that there
|
|
39 would be no further releases of CLC-INTERCAL. However, the compiler has
|
|
40 been left unattended on an FTP server, and appears to have escaped. For this
|
|
41 reason, we refer to this as an "escape" rather than a "release".
|
|
42
|
|
43 Additionally, development of CLC-INTERCAL version 0 has effectively stopped
|
|
44 before the escape of 0.06; a new versioning scheme has been introduced for
|
|
45 CLC-INTERCAL version 1. The main points to note are:
|
|
46
|
|
47 * The term "version" has been replaced by "perversion" for correctness
|
|
48 * The perversion number consists of a floating-point number with
|
|
49 independent signs for the integer and fractional part. Negative
|
|
50 fractions indicate pre-escapes (so 1.-94 means "94 pre-escapes to
|
|
51 go before 1.00". Or you can just add the numbers together and get
|
|
52 0.06, which is entirely a coincidence since 0.06 is not being
|
|
53 developed)
|
|
54 * The fractional part of a perversion number can be integer or floating
|
|
55 point, with a similar meaning for the parts. The current pre-escape is
|
|
56 1.-94.-2 which means "2 pre-pre-escapes to go before pre-escape 1.-94".
|
|
57
|
|
58 The fact that the compiler has escaped does not allow you to use without
|
|
59 reference to its licence: see the end of this document or the file COPYING.
|
|
60
|
|
61 Since this is a pre-pre-escape, there are parts which are unimplemented
|
|
62 or partially implemented. The file "MISSING" describes what is missing.
|
|
63 There might also parts which are implemented but are intended to be left
|
|
64 implemented: these will be removed from the final escapes.
|
|
65
|
|
66
|
|
67 - - ---- --- ------- ----- ------------ ----
|
|
68 1 - What has changed since CLC-INTERCAL 0.05
|
|
69 - - ---- --- ------- ----- ------------ ----
|
|
70
|
|
71 Just about everything. See the file "Changes" for a complete list, or see
|
|
72 the following executive summary.
|
|
73
|
|
74 The code is a complete rewrite. The "CREATE" statement which was documented
|
|
75 but not implemented has now been implemented (in a different form) and not
|
|
76 documented. This means that the compiler only needs to know about one
|
|
77 statement, "CREATE", and everything else can be created on-the-fly. This is
|
|
78 why the compiler itself and the various compatibility modules now include
|
|
79 INTERCAL source code consisting almost exclusively of "CREATE" statements.
|
|
80
|
|
81 This made it a lot easier to extend the language by specifying new features
|
|
82 in INTERCAL instead of Perl. Most statements have been modified to allow
|
|
83 creation of quantum bits. For example, it is possible to create 16 quantum
|
|
84 bits in one operation by assigning to a register while not assigning to it.
|
|
85 It goes without saying that the practice is not recommended if you don't
|
|
86 actually possess a quantum computer.
|
|
87
|
|
88 A new statement, "NEXT FROM", combines all the advantages of "COME FROM" and
|
|
89 "NEXT". It has been implemented essentially to provide a system call
|
|
90 interface, where the operating system is supposed to have a "NEXT FROM"
|
|
91 lurking in its darkest corner. See the last entry in the file "Changes"
|
|
92 for a complete description of the system call interface. Since this is
|
|
93 INTERCAL, the syscall interface is entirely system independent. System
|
|
94 dependent system calls are not supported.
|
|
95
|
|
96 All old programs should work unchanged (modulo the unimplemented parts, but
|
|
97 this is due to change as the perversion number progresses towards 1.00).
|
|
98 Programs intended for different compilers can also be compiled by selecting
|
|
99 a different compiler object (for example, "ick.iacc" provides compatibility
|
|
100 with the C-INTERCAL compiler "ick", and "1972.iacc" with the original
|
|
101 Princeton compiler). All these compiler object must be first compiled using
|
|
102 "iacc.iacc", the INTERCAL Astonishing Compiler Compiler, which is also used to
|
|
103 compile itself before compiling other compilers. A pre-compiled version is
|
|
104 provided in this escape otherwise you'll never get started.
|
|
105
|
|
106 The command-line compiler "oo, ick" and the inline module "Language::INTERCAL"
|
|
107 are not currently provided. They have been replaced by the compiler "sick"
|
|
108 and the module "Language::INTERCAL::Sick" respectively, which have a rather
|
|
109 different usage. A compatibility "oo, ick" and "Language::INTERCAL" is
|
|
110 intended to appear in a future pre-escape, and certainly before the escape.
|
|
111
|
|
112 The former "on-line reference", has been included in the distribution in
|
|
113 directory doc/html; it has been adapted for CLC-INTERCAL 1.-94; however
|
|
114 some new features might be still undocumented or underdocumented. Some of
|
|
115 the HTML documents are included as templates: the actual documentation
|
|
116 will be produced in blib/htmldoc
|
|
117
|
|
118
|
|
119 - - --- -- ------- ---- ------
|
|
120 2 - How to install this escape
|
|
121 - - --- -- ------- ---- ------
|
|
122
|
|
123 To unpack the distribution you need the following:
|
|
124
|
|
125 * A command interpreter compatible with the Bourne shell
|
|
126 * the "dd" program
|
|
127 * optionally, gzip and/or bzip2
|
|
128
|
|
129 If necessary, uncompress the distribution:
|
|
130
|
|
131 gzip -d CLC-INTERCAL-1.-94.-2.ddsh.gz
|
|
132
|
|
133 or:
|
|
134
|
|
135 bzip2 -d CLC-INTERCAL-1.-94.-2.ddsh.bz2
|
|
136
|
|
137 Unpack the distribution:
|
|
138
|
|
139 sh CLC-INTERCAL-1.-94.-2.ddsh
|
|
140
|
|
141 Create the Makefile:
|
|
142
|
|
143 cd CLC-INTERCAL-1.-94.-2
|
|
144 perl Makefile.PL
|
|
145
|
|
146 Build the compilers:
|
|
147
|
|
148 make
|
|
149
|
|
150 See if everything works:
|
|
151
|
|
152 make test
|
|
153
|
|
154 And install:
|
|
155
|
|
156 make install
|
|
157
|
|
158 It is possible that, by the time you read this, we have provided two
|
|
159 alternative versions of some modules: pure Perl and Perl-with-C. The latter
|
|
160 is likely to be better in some way (speed, size), otherwise we would not
|
|
161 provide it. If Makefile.PL finds that alternative versions are available,
|
|
162 it will ask which one you want to install. In general, the pure Perl versions
|
|
163 will work as long as your Perl installation works; the Perl-with-C should
|
|
164 work but might break because your C compiler is not the same as mine, or
|
|
165 simply because it's C and it's more self-obfuscating than INTERCAL. If it
|
|
166 does not work, install the pure Perl version. If that does not work either,
|
|
167 you have found an unimplemented feature. Congratulations.
|
|
168
|
|
169 The "make" and "make test" will take some time: "make" will run IACC (the
|
|
170 INTERCAL Astonishing Compiler Compiler) to build itself and all the other
|
|
171 compilers, and this is not quite a quick job. The "make test" will test
|
|
172 all parts of the compiler, including the Virual Machine, the Quantum
|
|
173 Emulator and the Just-Too-Late compiler. Have a coffee break, or read the
|
|
174 source code and try to figure out how it works. If you can figure out how
|
|
175 it works, please tell me, because I cannot.
|
|
176
|
|
177
|
|
178 - - --- -- ------ --- -----------
|
|
179 3 - How to modify the compiler(s)
|
|
180 - - --- -- ------ --- -----------
|
|
181
|
|
182 The compiler and the compiler compiler are currently underdocumented
|
|
183 (which means that there is no documentation whatsoever). If you feel
|
|
184 the urge to modify the compiler itself, ignore the Perl runtime library
|
|
185 and edit the file "sick.iacc", which would have been installed somewhere
|
|
186 system dependent. If you have write permission to this system dependent
|
|
187 location and your command interpreter is compatible with the Bourne Shell,
|
|
188 you can locate the file and edit it by simply typing (all in one line, we
|
|
189 show it in separate lines for readability):
|
|
190
|
|
191 sh -c '${EDITOR-vi} `sick -vo/dev/null sick.iacc 2>&1 |
|
|
192 head -1 | sed -e "s/\.\.\. \[.*$//"`'
|
|
193
|
|
194 If you would rather not type the above, or you have no write permission to
|
|
195 the appropriate place (or if you have an inferior command interpreter), try:
|
|
196
|
|
197 sick -vo/dev/null sick.iacc
|
|
198
|
|
199 The complete path will appear in the first line of the output, so you can copy
|
|
200 or edit it. You do want to copy it in case you break it:
|
|
201
|
|
202 sh -c 'cp `sick -vo/dev/null sick.iacc 2>&1 |\
|
|
203 head -1 | sed -e "s/\.\.\. \[.*$//"` iacc.io.SAFE'
|
|
204
|
|
205 When you have edited the file, rebuild the compiler object with:
|
|
206
|
|
207 sick sick.iacc
|
|
208
|
|
209 If you have edited a copy, type that from the directory where you left the
|
|
210 edited copy, or "sick" won't find it (but see the "-I" command line option)
|
|
211
|
|
212 A similar consideration applies to rebuilding IACC (the INTERCAL Astonishing
|
|
213 Compiler Compiler). However, in this case you need to make a copy of "iacc.io"
|
|
214 as well as "iacc.iacc":
|
|
215
|
|
216 sh -c 'cp `sick -vo/dev/null iacc.iacc 2>&1 |\
|
|
217 head -1 | sed -e "s/\.\.\. \[.*$//"` iacc.io.SAFE'
|
|
218 sh -c 'cp `sick -vo/dev/null iacc.io 2>&1 |\
|
|
219 head -1 | sed -e "s/\.\.\. \[.*$//"` iacc.io.SAFE'
|
|
220 sh -c '${EDITOR-vi} `sick -vo/dev/null iacc.iacc 2>&1 |\
|
|
221 head -1 | sed -e "s/\.\.\. \[.*$//"`'
|
|
222 sick iacc.iacc
|
|
223
|
|
224 If you break "iacc.io" you won't be able to recompile "iacc" unless you
|
|
225 re-install CLC-INTERCAL or use the copy you just made.
|
|
226
|
|
227
|
|
228 - - --------- ----- ----- --- --- -------- -------
|
|
229 4 - Important notes about the new INTERNET library
|
|
230 - - --------- ----- ----- --- --- -------- -------
|
|
231
|
|
232 A new INTERNET library is now available as an extension to the "sick"
|
|
233 compiler (it may work with other compilers, but this is not very likely
|
|
234 for now). See directory doc/INTERNET for details.
|
|
235
|
|
236 When this extension is loaded, the program will automatically start a
|
|
237 theft server if one isn't found already running on the local computer.
|
|
238 As there isn't yet an official port assignment for this, the port is
|
|
239 stored in the file system.sickrc, and should be changed when an official
|
|
240 assignment is available. Look for the line:
|
|
241
|
|
242 I CAN UNDERSTAND "..INTERNET.PORT" AS "64928"
|
|
243
|
|
244 and change the number as appropriate. You may also want to run your
|
|
245 INTERNET applications using a nonstandard port, but doing so will make
|
|
246 locally running programs incompatible with programs running elsewhere,
|
|
247 so stealing from remote systems will not succeed - unless of course one
|
|
248 has agreed on a port number with the remote system's admin.
|
|
249
|
|
250 Some of the forms of the CASE statement, and the STEAL/SMUGGLE without an
|
|
251 IP address, require knowledge of the IP address and netmask of local LANs.
|
|
252 In some cases, it may be possible to obtain these automatically, however this
|
|
253 is not possible in some systems. If the information is not available, no
|
|
254 broadcasts are possible. To work around this problem it is possible to specify
|
|
255 the list of interface in sickrc (user or system) with options like:
|
|
256
|
|
257 I CAN UNDERSTAND "..INTERNET.DEVICE.eth0" AS "192.168.1.2" WITH 24
|
|
258 I CAN UNDERSTAND "..INTERNET.DEVICE.eth1" AS "10.3.6.9" WITH 8
|
|
259
|
|
260 Note that the option specifies the IP address and netmask of each interface;
|
|
261 in this case, 192.168.1.2/24 and 10.3.6.9/8; the device name is ignored at
|
|
262 present, but it may be a good idea to leave it the same as the name known to
|
|
263 the kernel. Also note that you must install Net::Netmask to be able to use
|
|
264 the list of devices from sickrc.
|
|
265
|
|
266 An unusual feature of the INTERNET is that one can steal filehandles and they
|
|
267 will continue working. For example, if an INTERCAL program with the INTERNET
|
|
268 extension runs on computer A and has standard write redirected from a file,
|
|
269 then another program on computer B can steal that standard write filehandle
|
|
270 and will be able to write in from that file, even if the file is not on a
|
|
271 networked filesystem (because it's been stolen, however, the program which
|
|
272 opened the file can no longer use it - smuggling, on the other hand will allow
|
|
273 to share a file between computers). This feature has been used to create a
|
|
274 network-wide clipboard in which a server (a simple INTERCAL program) just
|
|
275 opens a file and then does nothing; clients then SMUGGLE the filehandle
|
|
276 and store the clipboard contents on it - automagically sharing it over the
|
|
277 network. Moreover, the server can be found using broadcasts, so there is
|
|
278 never any need for the user to know where it is running.
|
|
279
|
|
280
|
|
281 - - -------
|
|
282 5 - Licence
|
|
283 - - -------
|
|
284
|
|
285 Copyright (c) 2006-2008 Claudio Calvelli <intercal@sdf.lonestar.org>
|
|
286 (Please include the word INTERLEAVING in the subject when emailing that
|
|
287 address, or the email may be ignored)
|
|
288
|
|
289 In addition to the above, permission is hereby granted to use, misuse,
|
|
290 modify, distribute, break, fix again, etcetera CLC-INTERCAL-1.-94.-2
|
|
291 provided that the following conditions are met:
|
|
292
|
|
293 1. Redistributions of source code must retain the above copyright
|
|
294 notice, this list of conditions and the following disclaimer.
|
|
295 2. Redistributions in binary form must reproduce the above copyright
|
|
296 notice, this list of conditions and the following disclaimer in the
|
|
297 documentation and/or other materials provided with the distribution.
|
|
298 3. Neither the name of the Author nor the names of its contributors
|
|
299 may be used to endorse or promote products derived from this software
|
|
300 without specific prior written permission.
|
|
301
|
|
302 THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
|
|
303 ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
|
|
304 IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
|
|
305 ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
|
|
306 FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
|
|
307 DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
|
|
308 OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
|
|
309 HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
|
|
310 LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
|
|
311 OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
|
|
312 SUCH DAMAGE.
|
|
313
|