source: project/chicken/trunk/pcre/NON-UNIX-USE @ 12117

Last change on this file since 12117 was 12117, checked in by Kon Lovett, 13 years ago

PCRE 7.8, use of "full" flonum-hash, new scheme-complete by Alex Shinn.

File size: 17.8 KB
Line 
1Compiling PCRE on non-Unix systems
2----------------------------------
3
4This document contains the following sections:
5
6  General
7  Generic instructions for the PCRE C library
8  The C++ wrapper functions
9  Building for virtual Pascal
10  Stack size in Windows environments
11  Linking programs in Windows environments
12  Comments about Win32 builds
13  Building PCRE on Windows with CMake
14  Use of relative paths with CMake on Windows
15  Testing with runtest.bat
16  Building under Windows with BCC5.5
17  Building PCRE on OpenVMS
18
19
20GENERAL
21
22I (Philip Hazel) have no experience of Windows or VMS sytems and how their
23libraries work. The items in the PCRE distribution and Makefile that relate to
24anything other than Unix-like systems are untested by me.
25
26There are some other comments and files in the Contrib directory on the ftp
27site that you may find useful. See
28
29  ftp://ftp.csx.cam.ac.uk/pub/software/programming/pcre/Contrib
30
31If you want to compile PCRE for a non-Unix system (especially for a system that
32does not support "configure" and "make" files), note that the basic PCRE
33library consists entirely of code written in Standard C, and so should compile
34successfully on any system that has a Standard C compiler and library. The C++
35wrapper functions are a separate issue (see below).
36
37The PCRE distribution includes a "configure" file for use by the Configure/Make
38build system, as found in many Unix-like environments. There is also support
39support for CMake, which some users prefer, in particular in Windows
40environments. There are some instructions for CMake under Windows in the
41section entitled "Building PCRE with CMake" below. CMake can also be used to
42build PCRE in Unix-like systems.
43
44
45GENERIC INSTRUCTIONS FOR THE PCRE C LIBRARY
46
47The following are generic comments about building the PCRE C library "by hand".
48
49 (1) Copy or rename the file config.h.generic as config.h, and edit the macro
50     settings that it contains to whatever is appropriate for your environment.
51     In particular, if you want to force a specific value for newline, you can
52     define the NEWLINE macro. When you compile any of the PCRE modules, you
53     must specify -DHAVE_CONFIG_H to your compiler so that config.h is included
54     in the sources.
55
56     An alternative approach is not to edit config.h, but to use -D on the
57     compiler command line to make any changes that you need to the
58     configuration options. In this case -DHAVE_CONFIG_H must not be set.
59
60     NOTE: There have been occasions when the way in which certain parameters
61     in config.h are used has changed between releases. (In the configure/make
62     world, this is handled automatically.) When upgrading to a new release,
63     you are strongly advised to review config.h.generic before re-using what
64     you had previously.
65
66 (2) Copy or rename the file pcre.h.generic as pcre.h.
67
68 (3) EITHER:
69       Copy or rename file pcre_chartables.c.dist as pcre_chartables.c.
70
71     OR:
72       Compile dftables.c as a stand-alone program (using -DHAVE_CONFIG_H if
73       you have set up config.h), and then run it with the single argument
74       "pcre_chartables.c". This generates a set of standard character tables
75       and writes them to that file. The tables are generated using the default
76       C locale for your system. If you want to use a locale that is specified
77       by LC_xxx environment variables, add the -L option to the dftables
78       command. You must use this method if you are building on a system that
79       uses EBCDIC code.
80
81     The tables in pcre_chartables.c are defaults. The caller of PCRE can
82     specify alternative tables at run time.
83
84 (4) Ensure that you have the following header files:
85
86       pcre_internal.h
87       ucp.h
88
89 (5) Also ensure that you have the following file, which is #included as source
90     when building a debugging version of PCRE, and is also used by pcretest.
91
92       pcre_printint.src
93
94 (6) Compile the following source files, setting -DHAVE_CONFIG_H as a compiler
95     option if you have set up config.h with your configuration, or else use
96     other -D settings to change the configuration as required.
97
98       pcre_chartables.c
99       pcre_compile.c
100       pcre_config.c
101       pcre_dfa_exec.c
102       pcre_exec.c
103       pcre_fullinfo.c
104       pcre_get.c
105       pcre_globals.c
106       pcre_info.c
107       pcre_maketables.c
108       pcre_newline.c
109       pcre_ord2utf8.c
110       pcre_refcount.c
111       pcre_study.c
112       pcre_tables.c
113       pcre_try_flipped.c
114       pcre_ucd.c
115       pcre_valid_utf8.c
116       pcre_version.c
117       pcre_xclass.c
118
119     Make sure that you include -I. in the compiler command (or equivalent for
120     an unusual compiler) so that all included PCRE header files are first
121     sought in the current directory. Otherwise you run the risk of picking up
122     a previously-installed file from somewhere else.
123
124 (7) Now link all the compiled code into an object library in whichever form
125     your system keeps such libraries. This is the basic PCRE C library. If
126     your system has static and shared libraries, you may have to do this once
127     for each type.
128
129 (8) Similarly, compile pcreposix.c (remembering -DHAVE_CONFIG_H if necessary)
130     and link the result (on its own) as the pcreposix library.
131
132 (9) Compile the test program pcretest.c (again, don't forget -DHAVE_CONFIG_H).
133     This needs the functions in the pcre and pcreposix libraries when linking.
134     It also needs the pcre_printint.src source file, which it #includes.
135
136(10) Run pcretest on the testinput files in the testdata directory, and check
137     that the output matches the corresponding testoutput files. Note that the
138     supplied files are in Unix format, with just LF characters as line
139     terminators. You may need to edit them to change this if your system uses
140     a different convention. If you are using Windows, you probably should use
141     the wintestinput3 file instead of testinput3 (and the corresponding output
142     file). This is a locale test; wintestinput3 sets the locale to "french"
143     rather than "fr_FR", and there some minor output differences.
144
145(11) If you want to use the pcregrep command, compile and link pcregrep.c; it
146     uses only the basic PCRE library (it does not need the pcreposix library).
147
148
149THE C++ WRAPPER FUNCTIONS
150
151The PCRE distribution also contains some C++ wrapper functions and tests,
152contributed by Google Inc. On a system that can use "configure" and "make",
153the functions are automatically built into a library called pcrecpp. It should
154be straightforward to compile the .cc files manually on other systems. The
155files called xxx_unittest.cc are test programs for each of the corresponding
156xxx.cc files.
157
158
159BUILDING FOR VIRTUAL PASCAL
160
161A script for building PCRE using Borland's C++ compiler for use with VPASCAL
162was contributed by Alexander Tokarev. Stefan Weber updated the script and added
163additional files. The following files in the distribution are for building PCRE
164for use with VP/Borland: makevp_c.txt, makevp_l.txt, makevp.bat, pcregexp.pas.
165
166
167STACK SIZE IN WINDOWS ENVIRONMENTS
168
169The default processor stack size of 1Mb in some Windows environments is too
170small for matching patterns that need much recursion. In particular, test 2 may
171fail because of this. Normally, running out of stack causes a crash, but there
172have been cases where the test program has just died silently. See your linker
173documentation for how to increase stack size if you experience problems. The
174Linux default of 8Mb is a reasonable choice for the stack, though even that can
175be too small for some pattern/subject combinations.
176
177PCRE has a compile configuration option to disable the use of stack for
178recursion so that heap is used instead. However, pattern matching is
179significantly slower when this is done. There is more about stack usage in the
180"pcrestack" documentation.
181
182
183LINKING PROGRAMS IN WINDOWS ENVIRONMENTS
184
185If you want to statically link a program against a PCRE library in the form of
186a non-dll .a file, you must define PCRE_STATIC before including pcre.h,
187otherwise the pcre_malloc() and pcre_free() exported functions will be declared
188__declspec(dllimport), with unwanted results.
189
190
191CALLING CONVENTIONS IN WINDOWS ENVIRONMENTS
192
193It is possible to compile programs to use different calling conventions using
194MSVC. Search the web for "calling conventions" for more information. To make it
195easier to change the calling convention for the exported functions in the
196PCRE library, the macro PCRE_CALL_CONVENTION is present in all the external
197definitions. It can be set externally when compiling (e.g. in CFLAGS). If it is
198not set, it defaults to empty; the default calling convention is then used
199(which is what is wanted most of the time).
200
201
202COMMENTS ABOUT WIN32 BUILDS (see also "BUILDING PCRE WITH CMAKE" below)
203
204There are two ways of building PCRE using the "configure, make, make install"
205paradigm on Windows systems: using MinGW or using Cygwin. These are not at all
206the same thing; they are completely different from each other. There is also
207support for building using CMake, which some users find a more straightforward
208way of building PCRE under Windows. However, the tests are not run
209automatically when CMake is used.
210
211The MinGW home page (http://www.mingw.org/) says this:
212
213  MinGW: A collection of freely available and freely distributable Windows
214  specific header files and import libraries combined with GNU toolsets that
215  allow one to produce native Windows programs that do not rely on any
216  3rd-party C runtime DLLs.
217
218The Cygwin home page (http://www.cygwin.com/) says this:
219
220  Cygwin is a Linux-like environment for Windows. It consists of two parts:
221
222  . A DLL (cygwin1.dll) which acts as a Linux API emulation layer providing
223    substantial Linux API functionality
224
225  . A collection of tools which provide Linux look and feel.
226
227  The Cygwin DLL currently works with all recent, commercially released x86 32
228  bit and 64 bit versions of Windows, with the exception of Windows CE.
229
230On both MinGW and Cygwin, PCRE should build correctly using:
231
232  ./configure && make && make install
233
234This should create two libraries called libpcre and libpcreposix, and, if you
235have enabled building the C++ wrapper, a third one called libpcrecpp. These are
236independent libraries: when you like with libpcreposix or libpcrecpp you must
237also link with libpcre, which contains the basic functions. (Some earlier
238releases of PCRE included the basic libpcre functions in libpcreposix. This no
239longer happens.)
240
241A user submitted a special-purpose patch that makes it easy to create
242"pcre.dll" under mingw32 using the "msys" environment. It provides "pcre.dll"
243as a special target. If you use this target, no other files are built, and in
244particular, the pcretest and pcregrep programs are not built. An example of how
245this might be used is:
246
247  ./configure --enable-utf --disable-cpp CFLAGS="-03 -s"; make pcre.dll
248
249Using Cygwin's compiler generates libraries and executables that depend on
250cygwin1.dll. If a library that is generated this way is distributed,
251cygwin1.dll has to be distributed as well. Since cygwin1.dll is under the GPL
252licence, this forces not only PCRE to be under the GPL, but also the entire
253application. A distributor who wants to keep their own code proprietary must
254purchase an appropriate Cygwin licence.
255
256MinGW has no such restrictions. The MinGW compiler generates a library or
257executable that can run standalone on Windows without any third party dll or
258licensing issues.
259
260But there is more complication:
261
262If a Cygwin user uses the -mno-cygwin Cygwin gcc flag, what that really does is
263to tell Cygwin's gcc to use the MinGW gcc. Cygwin's gcc is only acting as a
264front end to MinGW's gcc (if you install Cygwin's gcc, you get both Cygwin's
265gcc and MinGW's gcc). So, a user can:
266
267. Build native binaries by using MinGW or by getting Cygwin and using
268  -mno-cygwin.
269
270. Build binaries that depend on cygwin1.dll by using Cygwin with the normal
271  compiler flags.
272
273The test files that are supplied with PCRE are in Unix format, with LF
274characters as line terminators. It may be necessary to change the line
275terminators in order to get some of the tests to work. We hope to improve
276things in this area in future.
277
278
279BUILDING PCRE ON WINDOWS WITH CMAKE
280
281CMake is an alternative build facility that can be used instead of the
282traditional Unix "configure". CMake version 2.4.7 supports Borland makefiles,
283MinGW makefiles, MSYS makefiles, NMake makefiles, UNIX makefiles, Visual Studio
2846, Visual Studio 7, Visual Studio 8, and Watcom W8. The following instructions
285were contributed by a PCRE user.
286
2871.  Download CMake 2.4.7 or above from http://www.cmake.org/, install and ensure
288    that cmake\bin is on your path.
289
2902.  Unzip (retaining folder structure) the PCRE source tree into a source
291    directory such as C:\pcre.
292
2933.  Create a new, empty build directory: C:\pcre\build\
294
2954.  Run CMakeSetup from the Shell envirornment of your build tool, e.g., Msys
296    for Msys/MinGW or Visual Studio Command Prompt for VC/VC++
297
2985.  Enter C:\pcre\pcre-xx and C:\pcre\build for the source and build
299    directories, respectively
300
3016.  Hit the "Configure" button.
302
3037.  Select the particular IDE / build tool that you are using (Visual Studio,
304    MSYS makefiles, MinGW makefiles, etc.)
305
3068.  The GUI will then list several configuration options. This is where you can
307    enable UTF-8 support, etc.
308
3099.  Hit "Configure" again. The adjacent "OK" button should now be active.
310
31110. Hit "OK".
312
31311. The build directory should now contain a usable build system, be it a
314    solution file for Visual Studio, makefiles for MinGW, etc.
315
316
317USE OF RELATIVE PATHS WITH CMAKE ON WINDOWS
318
319A PCRE user comments as follows:
320
321I thought that others may want to know the current state of
322CMAKE_USE_RELATIVE_PATHS support on Windows.
323
324Here it is:
325-- AdditionalIncludeDirectories is only partially modified (only the
326first path - see below)
327-- Only some of the contained file paths are modified - shown below for
328pcre.vcproj
329-- It properly modifies
330
331I am sure CMake people can fix that if they want to. Until then one will
332need to replace existing absolute paths in project files with relative
333paths manually (e.g. from VS) - relative to project file location. I did
334just that before being told to try CMAKE_USE_RELATIVE_PATHS. Not a big
335deal.
336
337AdditionalIncludeDirectories="E:\builds\pcre\build;E:\builds\pcre\pcre-7.5;"
338AdditionalIncludeDirectories=".;E:\builds\pcre\pcre-7.5;"
339
340RelativePath="pcre.h">
341RelativePath="pcre_chartables.c">
342RelativePath="pcre_chartables.c.rule">
343
344
345TESTING WITH RUNTEST.BAT
346
3471. Copy RunTest.bat into the directory where pcretest.exe has been created.
348
3492. Edit RunTest.bat and insert a line that indentifies the relative location of
350   the pcre source, e.g.:
351
352   set srcdir=..\pcre-7.4-RC3
353
3543. Run RunTest.bat from a command shell environment. Test outputs will
355   automatically be compared to expected results, and discrepancies will
356   identified in the console output.
357
3584. To test pcrecpp, run pcrecpp_unittest.exe, pcre_stringpiece_unittest.exe and
359   pcre_scanner_unittest.exe.
360
361
362BUILDING UNDER WINDOWS WITH BCC5.5
363
364Michael Roy sent these comments about building PCRE under Windows with BCC5.5:
365
366  Some of the core BCC libraries have a version of PCRE from 1998 built in,
367  which can lead to pcre_exec() giving an erroneous PCRE_ERROR_NULL from a
368  version mismatch. I'm including an easy workaround below, if you'd like to
369  include it in the non-unix instructions:
370
371  When linking a project with BCC5.5, pcre.lib must be included before any of
372  the libraries cw32.lib, cw32i.lib, cw32mt.lib, and cw32mti.lib on the command
373  line.
374
375
376BUILDING PCRE ON OPENVMS
377
378Dan Mooney sent the following comments about building PCRE on OpenVMS. They
379relate to an older version of PCRE that used fewer source files, so the exact
380commands will need changing. See the current list of source files above.
381
382"It was quite easy to compile and link the library. I don't have a formal
383make file but the attached file [reproduced below] contains the OpenVMS DCL
384commands I used to build the library. I had to add #define
385POSIX_MALLOC_THRESHOLD 10 to pcre.h since it was not defined anywhere.
386
387The library was built on:
388O/S: HP OpenVMS v7.3-1
389Compiler: Compaq C v6.5-001-48BCD
390Linker: vA13-01
391
392The test results did not match 100% due to the issues you mention in your
393documentation regarding isprint(), iscntrl(), isgraph() and ispunct(). I
394modified some of the character tables temporarily and was able to get the
395results to match. Tests using the fr locale did not match since I don't have
396that locale loaded. The study size was always reported to be 3 less than the
397value in the standard test output files."
398
399=========================
400$! This DCL procedure builds PCRE on OpenVMS
401$!
402$! I followed the instructions in the non-unix-use file in the distribution.
403$!
404$ COMPILE == "CC/LIST/NOMEMBER_ALIGNMENT/PREFIX_LIBRARY_ENTRIES=ALL_ENTRIES
405$ COMPILE DFTABLES.C
406$ LINK/EXE=DFTABLES.EXE DFTABLES.OBJ
407$ RUN DFTABLES.EXE/OUTPUT=CHARTABLES.C
408$ COMPILE MAKETABLES.C
409$ COMPILE GET.C
410$ COMPILE STUDY.C
411$! I had to set POSIX_MALLOC_THRESHOLD to 10 in PCRE.H since the symbol
412$! did not seem to be defined anywhere.
413$! I edited pcre.h and added #DEFINE SUPPORT_UTF8 to enable UTF8 support.
414$ COMPILE PCRE.C
415$ LIB/CREATE PCRE MAKETABLES.OBJ, GET.OBJ, STUDY.OBJ, PCRE.OBJ
416$! I had to set POSIX_MALLOC_THRESHOLD to 10 in PCRE.H since the symbol
417$! did not seem to be defined anywhere.
418$ COMPILE PCREPOSIX.C
419$ LIB/CREATE PCREPOSIX PCREPOSIX.OBJ
420$ COMPILE PCRETEST.C
421$ LINK/EXE=PCRETEST.EXE PCRETEST.OBJ, PCRE/LIB, PCREPOSIX/LIB
422$! C programs that want access to command line arguments must be
423$! defined as a symbol
424$ PCRETEST :== "$ SYS$ROADSUSERS:[DMOONEY.REGEXP]PCRETEST.EXE"
425$! Arguments must be enclosed in quotes.
426$ PCRETEST "-C"
427$! Test results:
428$!
429$!   The test results did not match 100%. The functions isprint(), iscntrl(),
430$!   isgraph() and ispunct() on OpenVMS must not produce the same results
431$!   as the system that built the test output files provided with the
432$!   distribution.
433$!
434$!   The study size did not match and was always 3 less on OpenVMS.
435$!
436$!   Locale could not be set to fr
437$!
438=========================
439
440Last Updated: 05 September 2008
441****
Note: See TracBrowser for help on using the repository browser.