/[pcre]/code/trunk/NON-UNIX-USE
ViewVC logotype

Diff of /code/trunk/NON-UNIX-USE

Parent Directory Parent Directory | Revision Log Revision Log | View Patch Patch

revision 41 by nigel, Sat Feb 24 21:39:17 2007 UTC revision 371 by ph10, Mon Aug 25 18:28:05 2008 UTC
# Line 1  Line 1 
1  Compiling PCRE on non-Unix systems  Compiling PCRE on non-Unix systems
2  ----------------------------------  ----------------------------------
3    
4  If you want to compile PCRE for a non-Unix system, note that it consists  This document contains the following sections:
 entirely of code written in Standard C, and so should compile successfully  
 on any machine with a Standard C compiler and library, using normal compiling  
 commands to do the following:  
   
 (1) Copy or rename the file config.in as config.h. You should not have to  
 change any settings inside it for a Standard C environment.  
   
 (2) Compile dftables.c as a stand-alone program, and then run it with  
 the standard output sent to chartables.c. This generates a set of standard  
 character tables.  
   
 (3) Compile maketables.c, get.c, study.c and pcre.c and link them all  
 together into an object library in whichever form your system keeps such  
 libraries. This is the pcre library (chartables.c gets included by means of an  
 #include directive).  
   
 (4) Similarly, compile pcreposix.c and link it as the pcreposix library.  
   
 (5) Compile the test program pcretest.c. This needs the functions in the  
 pcre and pcreposix libraries when linking.  
   
 (6) Run pcretest on the testinput files in the testdata directory, and check  
 that the output matches the corresponding testoutput files. You must use the  
 -i option when checking testinput2.  
   
 If you have a system without "configure" but where you can use a Makefile, edit  
 Makefile.in to create Makefile, substituting suitable values for the variables  
 at the head of the file.  
   
 Some help in building a Win32 DLL of PCRE in GnuWin32 environments was  
 contributed by Paul.Sokolovsky@technologist.com. These environments are  
 Mingw32 (http://www.xraylith.wisc.edu/~khan/software/gnu-win32/) and  
 CygWin  (http://sourceware.cygnus.com/cygwin/). Paul comments:  
   
   For CygWin, set CFLAGS=-mno-cygwin, and do 'make dll'. You'll get  
   pcre.dll (containing pcreposix also), libpcre.dll.a, and dynamically  
   linked pgrep and pcretest. If you have /bin/sh, run RunTest (three  
   main test go ok, locale not supported).  
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    
20    GENERAL
21    
22    I (Philip Hazel) have no experience of Windows or VMS sytems and how their
23    libraries work. The items in the PCRE distribution and Makefile that relate to
24    anything other than Unix-like systems are untested by me.
25    
26    There are some other comments and files in the Contrib directory on the ftp
27    site that you may find useful. See
28    
29      ftp://ftp.csx.cam.ac.uk/pub/software/programming/pcre/Contrib
30    
31    If you want to compile PCRE for a non-Unix system (especially for a system that
32    does not support "configure" and "make" files), note that the basic PCRE
33    library consists entirely of code written in Standard C, and so should compile
34    successfully on any system that has a Standard C compiler and library. The C++
35    wrapper functions are a separate issue (see below).
36    
37    The PCRE distribution includes a "configure" file for use by the Configure/Make
38    build system, as found in many Unix-like environments. There is also support
39    support for CMake, which some users prefer, in particular in Windows
40    environments. There are some instructions for CMake under Windows in the
41    section entitled "Building PCRE with CMake" below. CMake can also be used to
42    build PCRE in Unix-like systems.
43    
44    
45    GENERIC INSTRUCTIONS FOR THE PCRE C LIBRARY
46    
47    The 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           ucpinternal.h
89           ucptable.h
90    
91     (5) Also ensure that you have the following file, which is #included as source
92         when building a debugging version of PCRE, and is also used by pcretest.
93    
94           pcre_printint.src
95    
96     (6) Compile the following source files, setting -DHAVE_CONFIG_H as a compiler
97         option if you have set up config.h with your configuration, or else use
98         other -D settings to change the configuration as required.
99    
100           pcre_chartables.c
101           pcre_compile.c
102           pcre_config.c
103           pcre_dfa_exec.c
104           pcre_exec.c
105           pcre_fullinfo.c
106           pcre_get.c
107           pcre_globals.c
108           pcre_info.c
109           pcre_maketables.c
110           pcre_newline.c
111           pcre_ord2utf8.c
112           pcre_refcount.c
113           pcre_study.c
114           pcre_tables.c
115           pcre_try_flipped.c
116           pcre_ucp_searchfuncs.c
117           pcre_valid_utf8.c
118           pcre_version.c
119           pcre_xclass.c
120    
121         Make sure that you include -I. in the compiler command (or equivalent for
122         an unusual compiler) so that all included PCRE header files are first
123         sought in the current directory. Otherwise you run the risk of picking up
124         a previously-installed file from somewhere else.
125    
126     (7) Now link all the compiled code into an object library in whichever form
127         your system keeps such libraries. This is the basic PCRE C library. If
128         your system has static and shared libraries, you may have to do this once
129         for each type.
130    
131     (8) Similarly, compile pcreposix.c (remembering -DHAVE_CONFIG_H if necessary)
132         and link the result (on its own) as the pcreposix library.
133    
134     (9) Compile the test program pcretest.c (again, don't forget -DHAVE_CONFIG_H).
135         This needs the functions in the pcre and pcreposix libraries when linking.
136         It also needs the pcre_printint.src source file, which it #includes.
137    
138    (10) Run pcretest on the testinput files in the testdata directory, and check
139         that the output matches the corresponding testoutput files. Note that the
140         supplied files are in Unix format, with just LF characters as line
141         terminators. You may need to edit them to change this if your system uses
142         a different convention. If you are using Windows, you probably should use
143         the wintestinput3 file instead of testinput3 (and the corresponding output
144         file). This is a locale test; wintestinput3 sets the locale to "french"
145         rather than "fr_FR", and there some minor output differences.
146    
147    (11) If you want to use the pcregrep command, compile and link pcregrep.c; it
148         uses only the basic PCRE library (it does not need the pcreposix library).
149    
150    
151    THE C++ WRAPPER FUNCTIONS
152    
153    The PCRE distribution also contains some C++ wrapper functions and tests,
154    contributed by Google Inc. On a system that can use "configure" and "make",
155    the functions are automatically built into a library called pcrecpp. It should
156    be straightforward to compile the .cc files manually on other systems. The
157    files called xxx_unittest.cc are test programs for each of the corresponding
158    xxx.cc files.
159    
160    
161    BUILDING FOR VIRTUAL PASCAL
162    
163    A script for building PCRE using Borland's C++ compiler for use with VPASCAL
164    was contributed by Alexander Tokarev. Stefan Weber updated the script and added
165    additional files. The following files in the distribution are for building PCRE
166    for use with VP/Borland: makevp_c.txt, makevp_l.txt, makevp.bat, pcregexp.pas.
167    
168    
169    STACK SIZE IN WINDOWS ENVIRONMENTS
170    
171    The default processor stack size of 1Mb in some Windows environments is too
172    small for matching patterns that need much recursion. In particular, test 2 may
173    fail because of this. Normally, running out of stack causes a crash, but there
174    have been cases where the test program has just died silently. See your linker
175    documentation for how to increase stack size if you experience problems. The
176    Linux default of 8Mb is a reasonable choice for the stack, though even that can
177    be too small for some pattern/subject combinations.
178    
179    PCRE has a compile configuration option to disable the use of stack for
180    recursion so that heap is used instead. However, pattern matching is
181    significantly slower when this is done. There is more about stack usage in the
182    "pcrestack" documentation.
183    
184    
185    LINKING PROGRAMS IN WINDOWS ENVIRONMENTS
186    
187    If you want to statically link a program against a PCRE library in the form of
188    a non-dll .a file, you must define PCRE_STATIC before including pcre.h,
189    otherwise the pcre_malloc() and pcre_free() exported functions will be declared
190    __declspec(dllimport), with unwanted results.
191    
192    
193    CALLING CONVENTIONS IN WINDOWS ENVIRONMENTS
194    
195    It is possible to compile programs to use different calling conventions using
196    MSVC. Search the web for "calling conventions" for more information. To make it
197    easier to change the calling convention for the exported functions in the
198    PCRE library, the macro PCRE_CALL_CONVENTION is present in all the external
199    definitions. It can be set externally when compiling (e.g. in CFLAGS). If it is
200    not set, it defaults to empty; the default calling convention is then used
201    (which is what is wanted most of the time).
202    
203    
204    COMMENTS ABOUT WIN32 BUILDS (see also "BUILDING PCRE WITH CMAKE" below)
205    
206    There are two ways of building PCRE using the "configure, make, make install"
207    paradigm on Windows systems: using MinGW or using Cygwin. These are not at all
208    the same thing; they are completely different from each other. There is also
209    support for building using CMake, which some users find a more straightforward
210    way of building PCRE under Windows. However, the tests are not run
211    automatically when CMake is used.
212    
213    The MinGW home page (http://www.mingw.org/) says this:
214    
215      MinGW: A collection of freely available and freely distributable Windows
216      specific header files and import libraries combined with GNU toolsets that
217      allow one to produce native Windows programs that do not rely on any
218      3rd-party C runtime DLLs.
219    
220    The Cygwin home page (http://www.cygwin.com/) says this:
221    
222      Cygwin is a Linux-like environment for Windows. It consists of two parts:
223    
224      . A DLL (cygwin1.dll) which acts as a Linux API emulation layer providing
225        substantial Linux API functionality
226    
227      . A collection of tools which provide Linux look and feel.
228    
229      The Cygwin DLL currently works with all recent, commercially released x86 32
230      bit and 64 bit versions of Windows, with the exception of Windows CE.
231    
232    On both MinGW and Cygwin, PCRE should build correctly using:
233    
234      ./configure && make && make install
235    
236    This should create two libraries called libpcre and libpcreposix, and, if you
237    have enabled building the C++ wrapper, a third one called libpcrecpp. These are
238    independent libraries: when you like with libpcreposix or libpcrecpp you must
239    also link with libpcre, which contains the basic functions. (Some earlier
240    releases of PCRE included the basic libpcre functions in libpcreposix. This no
241    longer happens.)
242    
243    A user submitted a special-purpose patch that makes it easy to create
244    "pcre.dll" under mingw32 using the "msys" environment. It provides "pcre.dll"
245    as a special target. If you use this target, no other files are built, and in
246    particular, the pcretest and pcregrep programs are not built. An example of how
247    this might be used is:
248    
249      ./configure --enable-utf --disable-cpp CFLAGS="-03 -s"; make pcre.dll
250    
251    Using Cygwin's compiler generates libraries and executables that depend on
252    cygwin1.dll. If a library that is generated this way is distributed,
253    cygwin1.dll has to be distributed as well. Since cygwin1.dll is under the GPL
254    licence, this forces not only PCRE to be under the GPL, but also the entire
255    application. A distributor who wants to keep their own code proprietary must
256    purchase an appropriate Cygwin licence.
257    
258    MinGW has no such restrictions. The MinGW compiler generates a library or
259    executable that can run standalone on Windows without any third party dll or
260    licensing issues.
261    
262    But there is more complication:
263    
264    If a Cygwin user uses the -mno-cygwin Cygwin gcc flag, what that really does is
265    to tell Cygwin's gcc to use the MinGW gcc. Cygwin's gcc is only acting as a
266    front end to MinGW's gcc (if you install Cygwin's gcc, you get both Cygwin's
267    gcc and MinGW's gcc). So, a user can:
268    
269    . Build native binaries by using MinGW or by getting Cygwin and using
270      -mno-cygwin.
271    
272    . Build binaries that depend on cygwin1.dll by using Cygwin with the normal
273      compiler flags.
274    
275    The test files that are supplied with PCRE are in Unix format, with LF
276    characters as line terminators. It may be necessary to change the line
277    terminators in order to get some of the tests to work. We hope to improve
278    things in this area in future.
279    
280    
281    BUILDING PCRE ON WINDOWS WITH CMAKE
282    
283    CMake is an alternative build facility that can be used instead of the
284    traditional Unix "configure". CMake version 2.4.7 supports Borland makefiles,
285    MinGW makefiles, MSYS makefiles, NMake makefiles, UNIX makefiles, Visual Studio
286    6, Visual Studio 7, Visual Studio 8, and Watcom W8. The following instructions
287    were contributed by a PCRE user.
288    
289    1.  Download CMake 2.4.7 or above from http://www.cmake.org/, install and ensure
290        that cmake\bin is on your path.
291    
292    2.  Unzip (retaining folder structure) the PCRE source tree into a source
293        directory such as C:\pcre.
294    
295    3.  Create a new, empty build directory: C:\pcre\build\
296    
297    4.  Run CMakeSetup from the Shell envirornment of your build tool, e.g., Msys
298        for Msys/MinGW or Visual Studio Command Prompt for VC/VC++
299    
300    5.  Enter C:\pcre\pcre-xx and C:\pcre\build for the source and build
301        directories, respectively
302    
303    6.  Hit the "Configure" button.
304    
305    7.  Select the particular IDE / build tool that you are using (Visual Studio,
306        MSYS makefiles, MinGW makefiles, etc.)
307    
308    8.  The GUI will then list several configuration options. This is where you can
309        enable UTF-8 support, etc.
310    
311    9.  Hit "Configure" again. The adjacent "OK" button should now be active.
312    
313    10. Hit "OK".
314    
315    11. The build directory should now contain a usable build system, be it a
316        solution file for Visual Studio, makefiles for MinGW, etc.
317    
318    
319    USE OF RELATIVE PATHS WITH CMAKE ON WINDOWS
320    
321    A PCRE user comments as follows:
322    
323    I thought that others may want to know the current state of
324    CMAKE_USE_RELATIVE_PATHS support on Windows.
325    
326    Here it is:
327    -- AdditionalIncludeDirectories is only partially modified (only the
328    first path - see below)
329    -- Only some of the contained file paths are modified - shown below for
330    pcre.vcproj
331    -- It properly modifies
332    
333    I am sure CMake people can fix that if they want to. Until then one will
334    need to replace existing absolute paths in project files with relative
335    paths manually (e.g. from VS) - relative to project file location. I did
336    just that before being told to try CMAKE_USE_RELATIVE_PATHS. Not a big
337    deal.
338    
339    AdditionalIncludeDirectories="E:\builds\pcre\build;E:\builds\pcre\pcre-7.5;"
340    AdditionalIncludeDirectories=".;E:\builds\pcre\pcre-7.5;"
341    
342    RelativePath="pcre.h">
343    RelativePath="pcre_chartables.c">
344    RelativePath="pcre_chartables.c.rule">
345    
346    
347    TESTING WITH RUNTEST.BAT
348    
349    1. Copy RunTest.bat into the directory where pcretest.exe has been created.
350    
351    2. Edit RunTest.bat and insert a line that indentifies the relative location of
352       the pcre source, e.g.:
353    
354       set srcdir=..\pcre-7.4-RC3
355    
356    3. Run RunTest.bat from a command shell environment. Test outputs will
357       automatically be compared to expected results, and discrepancies will
358       identified in the console output.
359    
360    4. To test pcrecpp, run pcrecpp_unittest.exe, pcre_stringpiece_unittest.exe and
361       pcre_scanner_unittest.exe.
362    
363    
364    BUILDING UNDER WINDOWS WITH BCC5.5
365    
366    Michael Roy sent these comments about building PCRE under Windows with BCC5.5:
367    
368      Some of the core BCC libraries have a version of PCRE from 1998 built in,
369      which can lead to pcre_exec() giving an erroneous PCRE_ERROR_NULL from a
370      version mismatch. I'm including an easy workaround below, if you'd like to
371      include it in the non-unix instructions:
372    
373      When linking a project with BCC5.5, pcre.lib must be included before any of
374      the libraries cw32.lib, cw32i.lib, cw32mt.lib, and cw32mti.lib on the command
375      line.
376    
377    
378    BUILDING PCRE ON OPENVMS
379    
380    Dan Mooney sent the following comments about building PCRE on OpenVMS. They
381    relate to an older version of PCRE that used fewer source files, so the exact
382    commands will need changing. See the current list of source files above.
383    
384    "It was quite easy to compile and link the library. I don't have a formal
385    make file but the attached file [reproduced below] contains the OpenVMS DCL
386    commands I used to build the library. I had to add #define
387    POSIX_MALLOC_THRESHOLD 10 to pcre.h since it was not defined anywhere.
388    
389    The library was built on:
390    O/S: HP OpenVMS v7.3-1
391    Compiler: Compaq C v6.5-001-48BCD
392    Linker: vA13-01
393    
394    The test results did not match 100% due to the issues you mention in your
395    documentation regarding isprint(), iscntrl(), isgraph() and ispunct(). I
396    modified some of the character tables temporarily and was able to get the
397    results to match. Tests using the fr locale did not match since I don't have
398    that locale loaded. The study size was always reported to be 3 less than the
399    value in the standard test output files."
400    
401    =========================
402    $! This DCL procedure builds PCRE on OpenVMS
403    $!
404    $! I followed the instructions in the non-unix-use file in the distribution.
405    $!
406    $ COMPILE == "CC/LIST/NOMEMBER_ALIGNMENT/PREFIX_LIBRARY_ENTRIES=ALL_ENTRIES
407    $ COMPILE DFTABLES.C
408    $ LINK/EXE=DFTABLES.EXE DFTABLES.OBJ
409    $ RUN DFTABLES.EXE/OUTPUT=CHARTABLES.C
410    $ COMPILE MAKETABLES.C
411    $ COMPILE GET.C
412    $ COMPILE STUDY.C
413    $! I had to set POSIX_MALLOC_THRESHOLD to 10 in PCRE.H since the symbol
414    $! did not seem to be defined anywhere.
415    $! I edited pcre.h and added #DEFINE SUPPORT_UTF8 to enable UTF8 support.
416    $ COMPILE PCRE.C
417    $ LIB/CREATE PCRE MAKETABLES.OBJ, GET.OBJ, STUDY.OBJ, PCRE.OBJ
418    $! I had to set POSIX_MALLOC_THRESHOLD to 10 in PCRE.H since the symbol
419    $! did not seem to be defined anywhere.
420    $ COMPILE PCREPOSIX.C
421    $ LIB/CREATE PCREPOSIX PCREPOSIX.OBJ
422    $ COMPILE PCRETEST.C
423    $ LINK/EXE=PCRETEST.EXE PCRETEST.OBJ, PCRE/LIB, PCREPOSIX/LIB
424    $! C programs that want access to command line arguments must be
425    $! defined as a symbol
426    $ PCRETEST :== "$ SYS$ROADSUSERS:[DMOONEY.REGEXP]PCRETEST.EXE"
427    $! Arguments must be enclosed in quotes.
428    $ PCRETEST "-C"
429    $! Test results:
430    $!
431    $!   The test results did not match 100%. The functions isprint(), iscntrl(),
432    $!   isgraph() and ispunct() on OpenVMS must not produce the same results
433    $!   as the system that built the test output files provided with the
434    $!   distribution.
435    $!
436    $!   The study size did not match and was always 3 less on OpenVMS.
437    $!
438    $!   Locale could not be set to fr
439    $!
440    =========================
441    
442    Last Updated: 09 July 2008
443  ****  ****

Legend:
Removed from v.41  
changed lines
  Added in v.371

webmaster@exim.org
ViewVC Help
Powered by ViewVC 1.1.12