Anthony Baxter | 70e8e87 | 2006-07-10 07:41:04 +0000 | [diff] [blame] | 1 | This is Python version 2.5 beta 2 |
Skip Montanaro | bff63f0 | 2004-03-05 14:33:21 +0000 | [diff] [blame] | 2 | ================================== |
Guido van Rossum | 9144763 | 2000-04-11 17:11:09 +0000 | [diff] [blame] | 3 | |
Georg Brandl | b69406d | 2006-02-11 15:30:36 +0000 | [diff] [blame] | 4 | Copyright (c) 2001, 2002, 2003, 2004, 2005, 2006 Python Software Foundation. |
Guido van Rossum | 67345eb | 2001-01-18 14:51:12 +0000 | [diff] [blame] | 5 | All rights reserved. |
| 6 | |
Guido van Rossum | c07d5fa | 2000-09-01 22:50:02 +0000 | [diff] [blame] | 7 | Copyright (c) 2000 BeOpen.com. |
| 8 | All rights reserved. |
| 9 | |
Guido van Rossum | 23f7aed | 2001-04-12 20:53:31 +0000 | [diff] [blame] | 10 | Copyright (c) 1995-2001 Corporation for National Research Initiatives. |
Guido van Rossum | c07d5fa | 2000-09-01 22:50:02 +0000 | [diff] [blame] | 11 | All rights reserved. |
| 12 | |
Guido van Rossum | 6ebd299 | 2000-09-03 04:47:47 +0000 | [diff] [blame] | 13 | Copyright (c) 1991-1995 Stichting Mathematisch Centrum. |
Guido van Rossum | c07d5fa | 2000-09-01 22:50:02 +0000 | [diff] [blame] | 14 | All rights reserved. |
| 15 | |
Guido van Rossum | 4f40c2d | 2000-09-05 16:46:22 +0000 | [diff] [blame] | 16 | |
Guido van Rossum | c07d5fa | 2000-09-01 22:50:02 +0000 | [diff] [blame] | 17 | License information |
| 18 | ------------------- |
| 19 | |
Guido van Rossum | 4f40c2d | 2000-09-05 16:46:22 +0000 | [diff] [blame] | 20 | See the file "LICENSE" for information on the history of this |
| 21 | software, terms & conditions for usage, and a DISCLAIMER OF ALL |
| 22 | WARRANTIES. |
Guido van Rossum | c07d5fa | 2000-09-01 22:50:02 +0000 | [diff] [blame] | 23 | |
Guido van Rossum | 4f40c2d | 2000-09-05 16:46:22 +0000 | [diff] [blame] | 24 | This Python distribution contains no GNU General Public Licensed |
| 25 | (GPLed) code so it may be used in proprietary projects just like prior |
| 26 | Python distributions. There are interfaces to some GNU code but these |
| 27 | are entirely optional. |
| 28 | |
| 29 | All trademarks referenced herein are property of their respective |
| 30 | holders. |
Guido van Rossum | 7980826 | 1997-12-11 18:01:47 +0000 | [diff] [blame] | 31 | |
| 32 | |
Guido van Rossum | 8d7d4ed | 1996-07-30 21:41:07 +0000 | [diff] [blame] | 33 | What's new in this release? |
| 34 | --------------------------- |
Guido van Rossum | 3ff96dd | 1996-07-30 18:05:04 +0000 | [diff] [blame] | 35 | |
Guido van Rossum | 67345eb | 2001-01-18 14:51:12 +0000 | [diff] [blame] | 36 | See the file "Misc/NEWS". |
Guido van Rossum | 6477380 | 1997-11-26 17:07:02 +0000 | [diff] [blame] | 37 | |
Michael W. Hudson | 71dcc3e | 2005-02-22 15:33:26 +0000 | [diff] [blame] | 38 | |
Guido van Rossum | 6477380 | 1997-11-26 17:07:02 +0000 | [diff] [blame] | 39 | If you don't read instructions |
| 40 | ------------------------------ |
| 41 | |
| 42 | Congratulations on getting this far. :-) |
| 43 | |
| 44 | To start building right away (on UNIX): type "./configure" in the |
Guido van Rossum | 2dcec0c | 2002-02-04 01:59:23 +0000 | [diff] [blame] | 45 | current directory and when it finishes, type "make". This creates an |
| 46 | executable "./python"; to install in /usr/local, first do "su root" |
| 47 | and then "make install". |
| 48 | |
Michael W. Hudson | 71dcc3e | 2005-02-22 15:33:26 +0000 | [diff] [blame] | 49 | The section `Build instructions' below is still recommended reading. |
Guido van Rossum | 3ff96dd | 1996-07-30 18:05:04 +0000 | [diff] [blame] | 50 | |
| 51 | |
Guido van Rossum | 91cb9d2 | 1995-04-10 11:47:38 +0000 | [diff] [blame] | 52 | What is Python anyway? |
| 53 | ---------------------- |
| 54 | |
Michael W. Hudson | 71dcc3e | 2005-02-22 15:33:26 +0000 | [diff] [blame] | 55 | Python is an interpreted, interactive object-oriented programming |
| 56 | language suitable (amongst other uses) for distributed application |
| 57 | development, scripting, numeric computing and system testing. Python |
| 58 | is often compared to Tcl, Perl, Java, JavaScript, Visual Basic or |
| 59 | Scheme. To find out more about what Python can do for you, point your |
| 60 | browser to http://www.python.org/. |
Guido van Rossum | 91cb9d2 | 1995-04-10 11:47:38 +0000 | [diff] [blame] | 61 | |
| 62 | |
Guido van Rossum | f501b4e | 1996-10-25 14:32:48 +0000 | [diff] [blame] | 63 | How do I learn Python? |
| 64 | ---------------------- |
| 65 | |
Fred Drake | 0e6444c | 1999-05-17 19:35:01 +0000 | [diff] [blame] | 66 | The official tutorial is still a good place to start; see |
Michael W. Hudson | 71dcc3e | 2005-02-22 15:33:26 +0000 | [diff] [blame] | 67 | http://docs.python.org/ for online and downloadable versions, as well |
| 68 | as a list of other introductions, and reference documentation. |
Guido van Rossum | f501b4e | 1996-10-25 14:32:48 +0000 | [diff] [blame] | 69 | |
Guido van Rossum | d0a42e2 | 2000-03-31 20:16:45 +0000 | [diff] [blame] | 70 | There's a quickly growing set of books on Python. See |
Michael W. Hudson | fad4640 | 2005-06-16 15:51:20 +0000 | [diff] [blame] | 71 | http://wiki.python.org/moin/PythonBooks for a list. |
Guido van Rossum | 8d90f9d | 1997-05-22 20:13:25 +0000 | [diff] [blame] | 72 | |
Guido van Rossum | f501b4e | 1996-10-25 14:32:48 +0000 | [diff] [blame] | 73 | |
Guido van Rossum | c07d5fa | 2000-09-01 22:50:02 +0000 | [diff] [blame] | 74 | Documentation |
| 75 | ------------- |
Guido van Rossum | 91cb9d2 | 1995-04-10 11:47:38 +0000 | [diff] [blame] | 76 | |
Guido van Rossum | c07d5fa | 2000-09-01 22:50:02 +0000 | [diff] [blame] | 77 | All documentation is provided online in a variety of formats. In |
| 78 | order of importance for new users: Tutorial, Library Reference, |
Guido van Rossum | 4f40c2d | 2000-09-05 16:46:22 +0000 | [diff] [blame] | 79 | Language Reference, Extending & Embedding, and the Python/C API. The |
| 80 | Library Reference is especially of immense value since much of |
| 81 | Python's power is described there, including the built-in data types |
| 82 | and functions! |
Guido van Rossum | 91cb9d2 | 1995-04-10 11:47:38 +0000 | [diff] [blame] | 83 | |
Guido van Rossum | 4f40c2d | 2000-09-05 16:46:22 +0000 | [diff] [blame] | 84 | All documentation is also available online at the Python web site |
Michael W. Hudson | 71dcc3e | 2005-02-22 15:33:26 +0000 | [diff] [blame] | 85 | (http://docs.python.org/, see below). It is available online for |
Guido van Rossum | 4f40c2d | 2000-09-05 16:46:22 +0000 | [diff] [blame] | 86 | occasional reference, or can be downloaded in many formats for faster |
Fred Drake | 6caae14 | 2000-10-25 17:51:02 +0000 | [diff] [blame] | 87 | access. The documentation is available in HTML, PostScript, PDF, and |
| 88 | LaTeX formats; the LaTeX version is primarily for documentation |
| 89 | authors, translators, and people with special formatting requirements. |
Guido van Rossum | c07d5fa | 2000-09-01 22:50:02 +0000 | [diff] [blame] | 90 | |
Michael W. Hudson | 71dcc3e | 2005-02-22 15:33:26 +0000 | [diff] [blame] | 91 | Unfortunately, new-style classes (new in Python 2.2) have not yet been |
Anthony Baxter | 70e0796 | 2006-04-03 14:16:27 +0000 | [diff] [blame] | 92 | integrated into Python's standard documentation. A collection of |
Michael W. Hudson | 71dcc3e | 2005-02-22 15:33:26 +0000 | [diff] [blame] | 93 | pointers to what has been written is at: |
Barry Warsaw | 0655745 | 2001-12-14 20:47:12 +0000 | [diff] [blame] | 94 | |
Michael W. Hudson | 71dcc3e | 2005-02-22 15:33:26 +0000 | [diff] [blame] | 95 | http://www.python.org/doc/newstyle.html |
Barry Warsaw | 0655745 | 2001-12-14 20:47:12 +0000 | [diff] [blame] | 96 | |
Guido van Rossum | c07d5fa | 2000-09-01 22:50:02 +0000 | [diff] [blame] | 97 | |
Guido van Rossum | 4f40c2d | 2000-09-05 16:46:22 +0000 | [diff] [blame] | 98 | Web sites |
| 99 | --------- |
Guido van Rossum | c07d5fa | 2000-09-01 22:50:02 +0000 | [diff] [blame] | 100 | |
Guido van Rossum | 4f40c2d | 2000-09-05 16:46:22 +0000 | [diff] [blame] | 101 | New Python releases and related technologies are published at |
Guido van Rossum | 67345eb | 2001-01-18 14:51:12 +0000 | [diff] [blame] | 102 | http://www.python.org/. Come visit us! |
Guido van Rossum | c07d5fa | 2000-09-01 22:50:02 +0000 | [diff] [blame] | 103 | |
Michael W. Hudson | be6e0f4 | 2002-06-23 16:29:36 +0000 | [diff] [blame] | 104 | There's also a Python community web site at |
| 105 | http://starship.python.net/. |
Guido van Rossum | c07d5fa | 2000-09-01 22:50:02 +0000 | [diff] [blame] | 106 | |
Guido van Rossum | 4f40c2d | 2000-09-05 16:46:22 +0000 | [diff] [blame] | 107 | |
| 108 | Newsgroups and Mailing Lists |
| 109 | ---------------------------- |
Guido van Rossum | c07d5fa | 2000-09-01 22:50:02 +0000 | [diff] [blame] | 110 | |
| 111 | Read comp.lang.python, a high-volume discussion newsgroup about |
| 112 | Python, or comp.lang.python.announce, a low-volume moderated newsgroup |
| 113 | for Python-related announcements. These are also accessible as |
Michael W. Hudson | 71dcc3e | 2005-02-22 15:33:26 +0000 | [diff] [blame] | 114 | mailing lists: see http://www.python.org/community/lists.html for an |
| 115 | overview of these and many other Python-related mailing lists. |
Guido van Rossum | c07d5fa | 2000-09-01 22:50:02 +0000 | [diff] [blame] | 116 | |
Anthony Baxter | 70e0796 | 2006-04-03 14:16:27 +0000 | [diff] [blame] | 117 | Archives are accessible via the Google Groups Usenet archive; see |
Michael W. Hudson | be6e0f4 | 2002-06-23 16:29:36 +0000 | [diff] [blame] | 118 | http://groups.google.com/. The mailing lists are also archived, see |
Michael W. Hudson | 71dcc3e | 2005-02-22 15:33:26 +0000 | [diff] [blame] | 119 | http://www.python.org/community/lists.html for details. |
Guido van Rossum | c07d5fa | 2000-09-01 22:50:02 +0000 | [diff] [blame] | 120 | |
| 121 | |
| 122 | Bug reports |
| 123 | ----------- |
| 124 | |
Guido van Rossum | 4f40c2d | 2000-09-05 16:46:22 +0000 | [diff] [blame] | 125 | To report or search for bugs, please use the Python Bug |
| 126 | Tracker at http://sourceforge.net/bugs/?group_id=5470. |
Guido van Rossum | c07d5fa | 2000-09-01 22:50:02 +0000 | [diff] [blame] | 127 | |
| 128 | |
| 129 | Patches and contributions |
| 130 | ------------------------- |
| 131 | |
Fred Drake | 6caae14 | 2000-10-25 17:51:02 +0000 | [diff] [blame] | 132 | To submit a patch or other contribution, please use the Python Patch |
| 133 | Manager at http://sourceforge.net/patch/?group_id=5470. Guidelines |
| 134 | for patch submission may be found at http://www.python.org/patches/. |
Guido van Rossum | c07d5fa | 2000-09-01 22:50:02 +0000 | [diff] [blame] | 135 | |
| 136 | If you have a proposal to change Python, it's best to submit a Python |
| 137 | Enhancement Proposal (PEP) first. All current PEPs, as well as |
Walter Dörwald | 7dc65d6 | 2003-02-02 23:37:05 +0000 | [diff] [blame] | 138 | guidelines for submitting a new PEP, are listed at |
Michael W. Hudson | 71dcc3e | 2005-02-22 15:33:26 +0000 | [diff] [blame] | 139 | http://www.python.org/peps/. |
Guido van Rossum | c07d5fa | 2000-09-01 22:50:02 +0000 | [diff] [blame] | 140 | |
| 141 | |
| 142 | Questions |
| 143 | --------- |
| 144 | |
| 145 | For help, if you can't find it in the manuals or on the web site, it's |
| 146 | best to post to the comp.lang.python or the Python mailing list (see |
| 147 | above). If you specifically don't want to involve the newsgroup or |
Guido van Rossum | 67345eb | 2001-01-18 14:51:12 +0000 | [diff] [blame] | 148 | mailing list, send questions to help@python.org (a group of volunteers |
| 149 | who answer questions as they can). The newsgroup is the most |
| 150 | efficient way to ask public questions. |
Guido van Rossum | 627b2d7 | 1993-12-24 10:39:16 +0000 | [diff] [blame] | 151 | |
Guido van Rossum | 901454e | 2000-06-29 22:28:44 +0000 | [diff] [blame] | 152 | |
Guido van Rossum | 433c8ad | 1994-08-01 12:07:07 +0000 | [diff] [blame] | 153 | Build instructions |
Guido van Rossum | 91cb9d2 | 1995-04-10 11:47:38 +0000 | [diff] [blame] | 154 | ================== |
Guido van Rossum | 627b2d7 | 1993-12-24 10:39:16 +0000 | [diff] [blame] | 155 | |
Guido van Rossum | 59b0b60 | 2002-02-08 22:22:35 +0000 | [diff] [blame] | 156 | Before you can build Python, you must first configure it. |
| 157 | Fortunately, the configuration and build process has been automated |
| 158 | for Unix and Linux installations, so all you usually have to do is |
| 159 | type a few commands and sit back. There are some platforms where |
| 160 | things are not quite as smooth; see the platform specific notes below. |
| 161 | If you want to build for multiple platforms sharing the same source |
| 162 | tree, see the section on VPATH below. |
Guido van Rossum | 627b2d7 | 1993-12-24 10:39:16 +0000 | [diff] [blame] | 163 | |
Guido van Rossum | 59b0b60 | 2002-02-08 22:22:35 +0000 | [diff] [blame] | 164 | Start by running the script "./configure", which determines your |
| 165 | system configuration and creates the Makefile. (It takes a minute or |
| 166 | two -- please be patient!) You may want to pass options to the |
| 167 | configure script -- see the section below on configuration options and |
| 168 | variables. When it's done, you are ready to run make. |
Guido van Rossum | 6d9cc80 | 1996-08-01 17:31:22 +0000 | [diff] [blame] | 169 | |
Guido van Rossum | 59b0b60 | 2002-02-08 22:22:35 +0000 | [diff] [blame] | 170 | To build Python, you normally type "make" in the toplevel directory. |
| 171 | If you have changed the configuration, the Makefile may have to be |
| 172 | rebuilt. In this case you may have to run make again to correctly |
| 173 | build your desired target. The interpreter executable is built in the |
| 174 | top level directory. |
Guido van Rossum | 6d9cc80 | 1996-08-01 17:31:22 +0000 | [diff] [blame] | 175 | |
Guido van Rossum | 4f40c2d | 2000-09-05 16:46:22 +0000 | [diff] [blame] | 176 | Once you have built a Python interpreter, see the subsections below on |
Guido van Rossum | 59b0b60 | 2002-02-08 22:22:35 +0000 | [diff] [blame] | 177 | testing and installation. If you run into trouble, see the next |
| 178 | section. |
| 179 | |
| 180 | Previous versions of Python used a manual configuration process that |
| 181 | involved editing the file Modules/Setup. While this file still exists |
| 182 | and manual configuration is still supported, it is rarely needed any |
| 183 | more: almost all modules are automatically built as appropriate under |
| 184 | guidance of the setup.py script, which is run by Make after the |
| 185 | interpreter has been built. |
Guido van Rossum | 627b2d7 | 1993-12-24 10:39:16 +0000 | [diff] [blame] | 186 | |
Guido van Rossum | 0a516c9 | 1994-09-12 10:58:40 +0000 | [diff] [blame] | 187 | |
Guido van Rossum | 91cb9d2 | 1995-04-10 11:47:38 +0000 | [diff] [blame] | 188 | Troubleshooting |
| 189 | --------------- |
Guido van Rossum | 627b2d7 | 1993-12-24 10:39:16 +0000 | [diff] [blame] | 190 | |
Guido van Rossum | 6d9cc80 | 1996-08-01 17:31:22 +0000 | [diff] [blame] | 191 | See also the platform specific notes in the next section. |
| 192 | |
Michael W. Hudson | 71dcc3e | 2005-02-22 15:33:26 +0000 | [diff] [blame] | 193 | If you run into other trouble, see the FAQ |
| 194 | (http://www.python.org/doc/faq) for hints on what can go wrong, and |
| 195 | how to fix it. |
Guido van Rossum | 91cb9d2 | 1995-04-10 11:47:38 +0000 | [diff] [blame] | 196 | |
Guido van Rossum | 6d9cc80 | 1996-08-01 17:31:22 +0000 | [diff] [blame] | 197 | If you rerun the configure script with different options, remove all |
| 198 | object files by running "make clean" before rebuilding. Believe it or |
| 199 | not, "make clean" sometimes helps to clean up other inexplicable |
| 200 | problems as well. Try it before sending in a bug report! |
| 201 | |
Guido van Rossum | d0fe845 | 1996-08-26 03:02:37 +0000 | [diff] [blame] | 202 | If the configure script fails or doesn't seem to find things that |
Michael W. Hudson | 71dcc3e | 2005-02-22 15:33:26 +0000 | [diff] [blame] | 203 | should be there, inspect the config.log file. |
Guido van Rossum | c0be2f5 | 1997-10-08 05:05:28 +0000 | [diff] [blame] | 204 | |
| 205 | If you get a warning for every file about the -Olimit option being no |
| 206 | longer supported, you can ignore it. There's no foolproof way to know |
Guido van Rossum | 4f40c2d | 2000-09-05 16:46:22 +0000 | [diff] [blame] | 207 | whether this option is needed; all we can do is test whether it is |
Guido van Rossum | c0be2f5 | 1997-10-08 05:05:28 +0000 | [diff] [blame] | 208 | accepted without error. On some systems, e.g. older SGI compilers, it |
| 209 | is essential for performance (specifically when compiling ceval.c, |
| 210 | which has more basic blocks than the default limit of 1000). If the |
| 211 | warning bothers you, edit the Makefile to remove "-Olimit 1500" from |
| 212 | the OPT variable. |
Guido van Rossum | d0fe845 | 1996-08-26 03:02:37 +0000 | [diff] [blame] | 213 | |
Guido van Rossum | 7ac1153 | 2000-08-29 21:36:40 +0000 | [diff] [blame] | 214 | If you get failures in test_long, or sys.maxint gets set to -1, you |
| 215 | are probably experiencing compiler bugs, usually related to |
Martin v. Löwis | 9cea5f5 | 2002-11-15 07:28:05 +0000 | [diff] [blame] | 216 | optimization. This is a common problem with some versions of gcc, and |
| 217 | some vendor-supplied compilers, which can sometimes be worked around |
| 218 | by turning off optimization. Consider switching to stable versions |
Anthony Baxter | e9d719b | 2004-11-30 01:49:18 +0000 | [diff] [blame] | 219 | (gcc 2.95.2, gcc 3.x, or contact your vendor.) |
Guido van Rossum | 7ac1153 | 2000-08-29 21:36:40 +0000 | [diff] [blame] | 220 | |
| 221 | From Python 2.0 onward, all Python C code is ANSI C. Compiling using |
| 222 | old K&R-C-only compilers is no longer possible. ANSI C compilers are |
| 223 | available for all modern systems, either in the form of updated |
Martin v. Löwis | 9cea5f5 | 2002-11-15 07:28:05 +0000 | [diff] [blame] | 224 | compilers from the vendor, or one of the free compilers (gcc). |
Guido van Rossum | 91cb9d2 | 1995-04-10 11:47:38 +0000 | [diff] [blame] | 225 | |
Martin v. Löwis | 779ffc0 | 2002-12-02 22:17:01 +0000 | [diff] [blame] | 226 | Unsupported systems |
| 227 | ------------------- |
| 228 | |
Martin v. Löwis | 8cab8b0 | 2006-07-24 05:05:22 +0000 | [diff] [blame] | 229 | A number of features are not supported in Python 2.5 anymore. Some |
| 230 | support code is still present, but will be removed in Python 2.6. |
Martin v. Löwis | 779ffc0 | 2002-12-02 22:17:01 +0000 | [diff] [blame] | 231 | If you still need to use current Python versions on these systems, |
| 232 | please send a message to python-dev@python.org indicating that you |
Martin v. Löwis | 8cab8b0 | 2006-07-24 05:05:22 +0000 | [diff] [blame] | 233 | volunteer to support this system. For a more detailed discussion |
| 234 | regarding no-longer-supported and resupporting platforms, as well |
| 235 | as a list of platforms that became or will be unsupported, see PEP 11. |
Martin v. Löwis | 779ffc0 | 2002-12-02 22:17:01 +0000 | [diff] [blame] | 236 | |
| 237 | More specifically, the following systems are not supported any |
| 238 | longer: |
| 239 | - SunOS 4 |
| 240 | - DYNIX |
| 241 | - dgux |
| 242 | - Minix |
Martin v. Löwis | 8cab8b0 | 2006-07-24 05:05:22 +0000 | [diff] [blame] | 243 | - NeXT |
Martin v. Löwis | 779ffc0 | 2002-12-02 22:17:01 +0000 | [diff] [blame] | 244 | - Irix 4 and --with-sgi-dl |
| 245 | - Linux 1 |
| 246 | - Systems defining __d6_pthread_create (configure.in) |
| 247 | - Systems defining PY_PTHREAD_D4, PY_PTHREAD_D6, |
| 248 | or PY_PTHREAD_D7 in thread_pthread.h |
| 249 | - Systems using --with-dl-dld |
Martin v. Löwis | 7e4cfcb | 2002-12-19 16:21:49 +0000 | [diff] [blame] | 250 | - Systems using --without-universal-newlines |
Martin v. Löwis | 8cab8b0 | 2006-07-24 05:05:22 +0000 | [diff] [blame] | 251 | - MacOS 9 |
| 252 | |
| 253 | The following systems are still supported in Python 2.5, but |
| 254 | support will be dropped in 2.6: |
| 255 | - Systems using --with-wctype-functions |
| 256 | - Win9x, WinME |
| 257 | |
| 258 | Warning on install in Windows 98 and Windows Me |
| 259 | ----------------------------------------------- |
| 260 | |
| 261 | Following Microsoft's closing of Extended Support for |
| 262 | Windows 98/ME (July 11, 2006), Python 2.6 will stop |
| 263 | supporting these platforms. Python development and |
| 264 | maintainability becomes easier (and more reliable) when |
| 265 | platform specific code targeting OSes with few users |
| 266 | and no dedicated expert developers is taken out. The |
| 267 | vendor also warns that the OS versions listed above |
| 268 | "can expose customers to security risks" and recommends |
| 269 | upgrade. |
Martin v. Löwis | 779ffc0 | 2002-12-02 22:17:01 +0000 | [diff] [blame] | 270 | |
Guido van Rossum | 91cb9d2 | 1995-04-10 11:47:38 +0000 | [diff] [blame] | 271 | Platform specific notes |
| 272 | ----------------------- |
| 273 | |
Guido van Rossum | 0447a32 | 1995-10-08 01:22:33 +0000 | [diff] [blame] | 274 | (Some of these may no longer apply. If you find you can build Python |
Guido van Rossum | 67345eb | 2001-01-18 14:51:12 +0000 | [diff] [blame] | 275 | on these platforms without the special directions mentioned here, |
| 276 | submit a documentation bug report to SourceForge (see Bug Reports |
| 277 | above) so we can remove them!) |
Guido van Rossum | 91cb9d2 | 1995-04-10 11:47:38 +0000 | [diff] [blame] | 278 | |
Skip Montanaro | 6dc4a8e | 2003-05-02 15:27:59 +0000 | [diff] [blame] | 279 | Unix platforms: If your vendor still ships (and you still use) Berkeley DB |
Anthony Baxter | 70e0796 | 2006-04-03 14:16:27 +0000 | [diff] [blame] | 280 | 1.85 you will need to edit Modules/Setup to build the bsddb185 |
| 281 | module and add a line to sitecustomize.py which makes it the |
| 282 | default. In Modules/Setup a line like |
Skip Montanaro | 6dc4a8e | 2003-05-02 15:27:59 +0000 | [diff] [blame] | 283 | |
Anthony Baxter | 70e0796 | 2006-04-03 14:16:27 +0000 | [diff] [blame] | 284 | bsddb185 bsddbmodule.c |
Skip Montanaro | 6dc4a8e | 2003-05-02 15:27:59 +0000 | [diff] [blame] | 285 | |
Anthony Baxter | 70e0796 | 2006-04-03 14:16:27 +0000 | [diff] [blame] | 286 | should work. (You may need to add -I, -L or -l flags to direct the |
| 287 | compiler and linker to your include files and libraries.) |
Skip Montanaro | 6dc4a8e | 2003-05-02 15:27:59 +0000 | [diff] [blame] | 288 | |
Michael W. Hudson | 71dcc3e | 2005-02-22 15:33:26 +0000 | [diff] [blame] | 289 | XXX I think this next bit is out of date: |
Skip Montanaro | 6dc4a8e | 2003-05-02 15:27:59 +0000 | [diff] [blame] | 290 | |
Guido van Rossum | c0be2f5 | 1997-10-08 05:05:28 +0000 | [diff] [blame] | 291 | 64-bit platforms: The modules audioop, imageop and rgbimg don't work. |
Anthony Baxter | 70e0796 | 2006-04-03 14:16:27 +0000 | [diff] [blame] | 292 | The setup.py script disables them on 64-bit installations. |
| 293 | Don't try to enable them in the Modules/Setup file. They |
| 294 | contain code that is quite wordsize sensitive. (If you have a |
| 295 | fix, let us know!) |
Guido van Rossum | c0be2f5 | 1997-10-08 05:05:28 +0000 | [diff] [blame] | 296 | |
Guido van Rossum | 4462e93 | 1997-01-22 21:00:32 +0000 | [diff] [blame] | 297 | Solaris: When using Sun's C compiler with threads, at least on Solaris |
Anthony Baxter | 70e0796 | 2006-04-03 14:16:27 +0000 | [diff] [blame] | 298 | 2.5.1, you need to add the "-mt" compiler option (the simplest |
| 299 | way is probably to specify the compiler with this option as |
| 300 | the "CC" environment variable when running the configure |
| 301 | script). |
Guido van Rossum | 4462e93 | 1997-01-22 21:00:32 +0000 | [diff] [blame] | 302 | |
Guido van Rossum | 4299231 | 2002-09-23 14:02:18 +0000 | [diff] [blame] | 303 | When using GCC on Solaris, beware of binutils 2.13 or GCC |
| 304 | versions built using it. This mistakenly enables the |
| 305 | -zcombreloc option which creates broken shared libraries on |
| 306 | Solaris. binutils 2.12 works, and the binutils maintainers |
Skip Montanaro | 4de9cba | 2003-01-03 16:26:23 +0000 | [diff] [blame] | 307 | are aware of the problem. Binutils 2.13.1 only partially |
| 308 | fixed things. It appears that 2.13.2 solves the problem |
| 309 | completely. This problem is known to occur with Solaris 2.7 |
| 310 | and 2.8, but may also affect earlier and later versions of the |
| 311 | OS. |
Guido van Rossum | 4299231 | 2002-09-23 14:02:18 +0000 | [diff] [blame] | 312 | |
Anthony Baxter | 70e0796 | 2006-04-03 14:16:27 +0000 | [diff] [blame] | 313 | When the dynamic loader complains about errors finding shared |
| 314 | libraries, such as |
Martin v. Löwis | 9cea5f5 | 2002-11-15 07:28:05 +0000 | [diff] [blame] | 315 | |
Anthony Baxter | 70e0796 | 2006-04-03 14:16:27 +0000 | [diff] [blame] | 316 | ld.so.1: ./python: fatal: libstdc++.so.5: open failed: |
| 317 | No such file or directory |
Martin v. Löwis | 9cea5f5 | 2002-11-15 07:28:05 +0000 | [diff] [blame] | 318 | |
Anthony Baxter | 70e0796 | 2006-04-03 14:16:27 +0000 | [diff] [blame] | 319 | you need to first make sure that the library is available on |
| 320 | your system. Then, you need to instruct the dynamic loader how |
| 321 | to find it. You can choose any of the following strategies: |
Martin v. Löwis | 9cea5f5 | 2002-11-15 07:28:05 +0000 | [diff] [blame] | 322 | |
Anthony Baxter | 70e0796 | 2006-04-03 14:16:27 +0000 | [diff] [blame] | 323 | 1. When compiling Python, set LD_RUN_PATH to the directories |
| 324 | containing missing libraries. |
| 325 | 2. When running Python, set LD_LIBRARY_PATH to these directories. |
| 326 | 3. Use crle(8) to extend the search path of the loader. |
| 327 | 4. Modify the installed GCC specs file, adding -R options into the |
| 328 | *link: section. |
Martin v. Löwis | 9cea5f5 | 2002-11-15 07:28:05 +0000 | [diff] [blame] | 329 | |
Skip Montanaro | 9f6606a | 2005-11-17 18:19:39 +0000 | [diff] [blame] | 330 | The complex object fails to compile on Solaris 10 with gcc 3.4 (at |
| 331 | least up to 3.4.3). To work around it, define Py_HUGE_VAL as |
| 332 | HUGE_VAL(), e.g.: |
| 333 | |
| 334 | make CPPFLAGS='-D"Py_HUGE_VAL=HUGE_VAL()" -I. -I$(srcdir)/Include' |
Anthony Baxter | 70e0796 | 2006-04-03 14:16:27 +0000 | [diff] [blame] | 335 | ./python setup.py CPPFLAGS='-D"Py_HUGE_VAL=HUGE_VAL()"' |
Skip Montanaro | 9f6606a | 2005-11-17 18:19:39 +0000 | [diff] [blame] | 336 | |
Guido van Rossum | 7ac1153 | 2000-08-29 21:36:40 +0000 | [diff] [blame] | 337 | Linux: A problem with threads and fork() was tracked down to a bug in |
Anthony Baxter | 70e0796 | 2006-04-03 14:16:27 +0000 | [diff] [blame] | 338 | the pthreads code in glibc version 2.0.5; glibc version 2.0.7 |
| 339 | solves the problem. This causes the popen2 test to fail; |
| 340 | problem and solution reported by Pablo Bleyer. |
Guido van Rossum | 4462e93 | 1997-01-22 21:00:32 +0000 | [diff] [blame] | 341 | |
Martin v. Löwis | 3efa04a | 2003-06-14 05:29:27 +0000 | [diff] [blame] | 342 | Red Hat Linux: Red Hat 9 built Python2.2 in UCS-4 mode and hacked |
Anthony Baxter | 70e0796 | 2006-04-03 14:16:27 +0000 | [diff] [blame] | 343 | Tcl to support it. To compile Python2.3 with Tkinter, you will |
| 344 | need to pass --enable-unicode=ucs4 flag to ./configure. |
Martin v. Löwis | 3efa04a | 2003-06-14 05:29:27 +0000 | [diff] [blame] | 345 | |
Anthony Baxter | 70e0796 | 2006-04-03 14:16:27 +0000 | [diff] [blame] | 346 | There's an executable /usr/bin/python which is Python |
| 347 | 1.5.2 on most older Red Hat installations; several key Red Hat tools |
| 348 | require this version. Python 2.1.x may be installed as |
| 349 | /usr/bin/python2. The Makefile installs Python as |
| 350 | /usr/local/bin/python, which may or may not take precedence |
| 351 | over /usr/bin/python, depending on how you have set up $PATH. |
Guido van Rossum | 2dcec0c | 2002-02-04 01:59:23 +0000 | [diff] [blame] | 352 | |
Guido van Rossum | 5923050 | 1999-06-21 20:51:46 +0000 | [diff] [blame] | 353 | FreeBSD 3.x and probably platforms with NCurses that use libmytinfo or |
Anthony Baxter | 70e0796 | 2006-04-03 14:16:27 +0000 | [diff] [blame] | 354 | similar: When using cursesmodule, the linking is not done in |
| 355 | the correct order with the defaults. Remove "-ltermcap" from |
| 356 | the readline entry in Setup, and use as curses entry: "curses |
| 357 | cursesmodule.c -lmytinfo -lncurses -ltermcap" - "mytinfo" (so |
| 358 | called on FreeBSD) should be the name of the auxiliary library |
| 359 | required on your platform. Normally, it would be linked |
| 360 | automatically, but not necessarily in the correct order. |
Guido van Rossum | 5923050 | 1999-06-21 20:51:46 +0000 | [diff] [blame] | 361 | |
Anthony Baxter | 70e0796 | 2006-04-03 14:16:27 +0000 | [diff] [blame] | 362 | BSDI: BSDI versions before 4.1 have known problems with threads, |
| 363 | which can cause strange errors in a number of modules (for |
| 364 | instance, the 'test_signal' test script will hang forever.) |
| 365 | Turning off threads (with --with-threads=no) or upgrading to |
| 366 | BSDI 4.1 solves this problem. |
Guido van Rossum | 7ac1153 | 2000-08-29 21:36:40 +0000 | [diff] [blame] | 367 | |
| 368 | DEC Unix: Run configure with --with-dec-threads, or with |
Anthony Baxter | 70e0796 | 2006-04-03 14:16:27 +0000 | [diff] [blame] | 369 | --with-threads=no if no threads are desired (threads are on by |
| 370 | default). When using GCC, it is possible to get an internal |
| 371 | compiler error if optimization is used. This was reported for |
| 372 | GCC 2.7.2.3 on selectmodule.c. Manually compile the affected |
| 373 | file without optimization to solve the problem. |
Guido van Rossum | 8eca2c2 | 1996-02-14 18:37:46 +0000 | [diff] [blame] | 374 | |
Guido van Rossum | e1cdb16 | 1999-04-10 16:04:07 +0000 | [diff] [blame] | 375 | DEC Ultrix: compile with GCC to avoid bugs in the native compiler, |
Anthony Baxter | 70e0796 | 2006-04-03 14:16:27 +0000 | [diff] [blame] | 376 | and pass SHELL=/bin/sh5 to Make when installing. |
Guido van Rossum | e1cdb16 | 1999-04-10 16:04:07 +0000 | [diff] [blame] | 377 | |
Anthony Baxter | 70e0796 | 2006-04-03 14:16:27 +0000 | [diff] [blame] | 378 | AIX: A complete overhaul of the shared library support is now in |
| 379 | place. See Misc/AIX-NOTES for some notes on how it's done. |
| 380 | (The optimizer bug reported at this place in previous releases |
| 381 | has been worked around by a minimal code change.) If you get |
| 382 | errors about pthread_* functions, during compile or during |
| 383 | testing, try setting CC to a thread-safe (reentrant) compiler, |
| 384 | like "cc_r". For full C++ module support, set CC="xlC_r" (or |
| 385 | CC="xlC" without thread support). |
Guido van Rossum | 91cb9d2 | 1995-04-10 11:47:38 +0000 | [diff] [blame] | 386 | |
Guido van Rossum | a2534c8 | 2005-09-20 19:06:23 +0000 | [diff] [blame] | 387 | AIX 5.3: To build a 64-bit version with IBM's compiler, I used the |
| 388 | following: |
| 389 | |
| 390 | export PATH=/usr/bin:/usr/vacpp/bin |
Anthony Baxter | 70e0796 | 2006-04-03 14:16:27 +0000 | [diff] [blame] | 391 | ./configure --with-gcc="xlc_r -q64" --with-cxx="xlC_r -q64" \ |
Guido van Rossum | a2534c8 | 2005-09-20 19:06:23 +0000 | [diff] [blame] | 392 | --disable-ipv6 AR="ar -X64" |
Anthony Baxter | 70e0796 | 2006-04-03 14:16:27 +0000 | [diff] [blame] | 393 | make |
Guido van Rossum | a2534c8 | 2005-09-20 19:06:23 +0000 | [diff] [blame] | 394 | |
Anthony Baxter | 7c49747 | 2003-10-01 07:24:54 +0000 | [diff] [blame] | 395 | HP-UX: When using threading, you may have to add -D_REENTRANT to the |
Anthony Baxter | 70e0796 | 2006-04-03 14:16:27 +0000 | [diff] [blame] | 396 | OPT variable in the top-level Makefile; reported by Pat Knight, |
| 397 | this seems to make a difference (at least for HP-UX 10.20) |
| 398 | even though pyconfig.h defines it. This seems unnecessary when |
| 399 | using HP/UX 11 and later - threading seems to work "out of the |
| 400 | box". |
Anthony Baxter | 7c49747 | 2003-10-01 07:24:54 +0000 | [diff] [blame] | 401 | |
Michael W. Hudson | 71dcc3e | 2005-02-22 15:33:26 +0000 | [diff] [blame] | 402 | HP-UX ia64: When building on the ia64 (Itanium) platform using HP's |
Anthony Baxter | 70e0796 | 2006-04-03 14:16:27 +0000 | [diff] [blame] | 403 | compiler, some experience has shown that the compiler's |
| 404 | optimiser produces a completely broken version of python |
| 405 | (see http://www.python.org/sf/814976). To work around this, |
| 406 | edit the Makefile and remove -O from the OPT line. |
Guido van Rossum | 539c662 | 2005-09-14 17:49:54 +0000 | [diff] [blame] | 407 | |
Anthony Baxter | 70e0796 | 2006-04-03 14:16:27 +0000 | [diff] [blame] | 408 | To build a 64-bit executable on an Itanium 2 system using HP's |
| 409 | compiler, use these environment variables: |
Guido van Rossum | 539c662 | 2005-09-14 17:49:54 +0000 | [diff] [blame] | 410 | |
Anthony Baxter | 70e0796 | 2006-04-03 14:16:27 +0000 | [diff] [blame] | 411 | CC=cc |
| 412 | CXX=aCC |
| 413 | BASECFLAGS="+DD64" |
| 414 | LDFLAGS="+DD64 -lxnet" |
Guido van Rossum | 539c662 | 2005-09-14 17:49:54 +0000 | [diff] [blame] | 415 | |
Anthony Baxter | 70e0796 | 2006-04-03 14:16:27 +0000 | [diff] [blame] | 416 | and call configure as: |
Guido van Rossum | 539c662 | 2005-09-14 17:49:54 +0000 | [diff] [blame] | 417 | |
Anthony Baxter | 70e0796 | 2006-04-03 14:16:27 +0000 | [diff] [blame] | 418 | ./configure --without-gcc |
Guido van Rossum | 539c662 | 2005-09-14 17:49:54 +0000 | [diff] [blame] | 419 | |
Anthony Baxter | 70e0796 | 2006-04-03 14:16:27 +0000 | [diff] [blame] | 420 | then *unset* the environment variables again before running |
| 421 | make. (At least one of these flags causes the build to fail |
| 422 | if it remains set.) You still have to edit the Makefile and |
| 423 | remove -O from the OPT line. |
Guido van Rossum | 2094e04 | 1998-05-14 15:47:35 +0000 | [diff] [blame] | 424 | |
Guido van Rossum | 4e1dd7d | 2002-04-23 13:06:07 +0000 | [diff] [blame] | 425 | HP PA-RISC 2.0: A recent bug report (http://www.python.org/sf/546117) |
Anthony Baxter | 70e0796 | 2006-04-03 14:16:27 +0000 | [diff] [blame] | 426 | suggests that the C compiler in this 64-bit system has bugs |
| 427 | in the optimizer that break Python. Compiling without |
| 428 | optimization solves the problems. |
Guido van Rossum | 4e1dd7d | 2002-04-23 13:06:07 +0000 | [diff] [blame] | 429 | |
Anthony Baxter | 70e0796 | 2006-04-03 14:16:27 +0000 | [diff] [blame] | 430 | SCO: The following apply to SCO 3 only; Python builds out of the box |
| 431 | on SCO 5 (or so we've heard). |
Guido van Rossum | a16e275 | 1998-04-13 20:14:05 +0000 | [diff] [blame] | 432 | |
Anthony Baxter | 70e0796 | 2006-04-03 14:16:27 +0000 | [diff] [blame] | 433 | 1) Everything works much better if you add -U__STDC__ to the |
| 434 | defs. This is because all the SCO header files are broken. |
| 435 | Anything that isn't mentioned in the C standard is |
| 436 | conditionally excluded when __STDC__ is defined. |
Guido van Rossum | 91cb9d2 | 1995-04-10 11:47:38 +0000 | [diff] [blame] | 437 | |
Anthony Baxter | 70e0796 | 2006-04-03 14:16:27 +0000 | [diff] [blame] | 438 | 2) Due to the U.S. export restrictions, SCO broke the crypt |
| 439 | stuff out into a separate library, libcrypt_i.a so the LIBS |
| 440 | needed be set to: |
Guido van Rossum | 91cb9d2 | 1995-04-10 11:47:38 +0000 | [diff] [blame] | 441 | |
Anthony Baxter | 70e0796 | 2006-04-03 14:16:27 +0000 | [diff] [blame] | 442 | LIBS=' -lsocket -lcrypt_i' |
Guido van Rossum | 91cb9d2 | 1995-04-10 11:47:38 +0000 | [diff] [blame] | 443 | |
Martin v. Löwis | 387c547 | 2001-09-06 08:16:17 +0000 | [diff] [blame] | 444 | UnixWare: There are known bugs in the math library of the system, as well as |
| 445 | problems in the handling of threads (calling fork in one |
| 446 | thread may interrupt system calls in others). Therefore, test_math and |
| 447 | tests involving threads will fail until those problems are fixed. |
| 448 | |
Anthony Baxter | 70e0796 | 2006-04-03 14:16:27 +0000 | [diff] [blame] | 449 | QNX: Chris Herborth (chrish@qnx.com) writes: |
| 450 | configure works best if you use GNU bash; a port is available on |
| 451 | ftp.qnx.com in /usr/free. I used the following process to build, |
| 452 | test and install Python 1.5.x under QNX: |
Guido van Rossum | b7f454d | 1997-12-02 19:44:31 +0000 | [diff] [blame] | 453 | |
Anthony Baxter | 70e0796 | 2006-04-03 14:16:27 +0000 | [diff] [blame] | 454 | 1) CONFIG_SHELL=/usr/local/bin/bash CC=cc RANLIB=: \ |
| 455 | ./configure --verbose --without-gcc --with-libm="" |
Guido van Rossum | b7f454d | 1997-12-02 19:44:31 +0000 | [diff] [blame] | 456 | |
Anthony Baxter | 70e0796 | 2006-04-03 14:16:27 +0000 | [diff] [blame] | 457 | 2) edit Modules/Setup to activate everything that makes sense for |
| 458 | your system... tested here at QNX with the following modules: |
Guido van Rossum | b7f454d | 1997-12-02 19:44:31 +0000 | [diff] [blame] | 459 | |
Anthony Baxter | 70e0796 | 2006-04-03 14:16:27 +0000 | [diff] [blame] | 460 | array, audioop, binascii, cPickle, cStringIO, cmath, |
| 461 | crypt, curses, errno, fcntl, gdbm, grp, imageop, |
| 462 | _locale, math, md5, new, operator, parser, pcre, |
| 463 | posix, pwd, readline, regex, reop, rgbimg, rotor, |
| 464 | select, signal, socket, soundex, strop, struct, |
| 465 | syslog, termios, time, timing, zlib, audioop, imageop, rgbimg |
Guido van Rossum | b7f454d | 1997-12-02 19:44:31 +0000 | [diff] [blame] | 466 | |
Anthony Baxter | 70e0796 | 2006-04-03 14:16:27 +0000 | [diff] [blame] | 467 | 3) make SHELL=/usr/local/bin/bash |
Guido van Rossum | 16e809e | 2001-04-16 02:07:08 +0000 | [diff] [blame] | 468 | |
Anthony Baxter | 70e0796 | 2006-04-03 14:16:27 +0000 | [diff] [blame] | 469 | or, if you feel the need for speed: |
Guido van Rossum | 16e809e | 2001-04-16 02:07:08 +0000 | [diff] [blame] | 470 | |
Anthony Baxter | 70e0796 | 2006-04-03 14:16:27 +0000 | [diff] [blame] | 471 | make SHELL=/usr/local/bin/bash OPT="-5 -Oil+nrt" |
Guido van Rossum | b7f454d | 1997-12-02 19:44:31 +0000 | [diff] [blame] | 472 | |
Anthony Baxter | 70e0796 | 2006-04-03 14:16:27 +0000 | [diff] [blame] | 473 | 4) make SHELL=/usr/local/bin/bash test |
Guido van Rossum | ec95c7b | 1998-08-04 17:59:56 +0000 | [diff] [blame] | 474 | |
Anthony Baxter | 70e0796 | 2006-04-03 14:16:27 +0000 | [diff] [blame] | 475 | Using GNU readline 2.2 seems to behave strangely, but I |
| 476 | think that's a problem with my readline 2.2 port. :-\ |
Guido van Rossum | ec95c7b | 1998-08-04 17:59:56 +0000 | [diff] [blame] | 477 | |
Anthony Baxter | 70e0796 | 2006-04-03 14:16:27 +0000 | [diff] [blame] | 478 | 5) make SHELL=/usr/local/bin/bash install |
Guido van Rossum | b7f454d | 1997-12-02 19:44:31 +0000 | [diff] [blame] | 479 | |
Anthony Baxter | 70e0796 | 2006-04-03 14:16:27 +0000 | [diff] [blame] | 480 | If you get SIGSEGVs while running Python (I haven't yet, but |
| 481 | I've only run small programs and the test cases), you're |
| 482 | probably running out of stack; the default 32k could be a |
| 483 | little tight. To increase the stack size, edit the Makefile |
| 484 | to read: LDFLAGS = -N 48k |
Guido van Rossum | 40d6358 | 1997-08-14 19:45:30 +0000 | [diff] [blame] | 485 | |
Anthony Baxter | 70e0796 | 2006-04-03 14:16:27 +0000 | [diff] [blame] | 486 | BeOS: See Misc/BeOS-NOTES for notes about compiling/installing |
| 487 | Python on BeOS R3 or later. Note that only the PowerPC |
| 488 | platform is supported for R3; both PowerPC and x86 are |
| 489 | supported for R4. |
Guido van Rossum | ec95c7b | 1998-08-04 17:59:56 +0000 | [diff] [blame] | 490 | |
Michael W. Hudson | 33876f5 | 2002-07-31 09:55:25 +0000 | [diff] [blame] | 491 | Cray T3E: Mark Hadfield (m.hadfield@niwa.co.nz) writes: |
Anthony Baxter | 70e0796 | 2006-04-03 14:16:27 +0000 | [diff] [blame] | 492 | Python can be built satisfactorily on a Cray T3E but based on |
| 493 | my experience with the NIWA T3E (2002-05-22, version 2.2.1) |
| 494 | there are a few bugs and gotchas. For more information see a |
| 495 | thread on comp.lang.python in May 2002 entitled "Building |
| 496 | Python on Cray T3E". |
Michael W. Hudson | 33876f5 | 2002-07-31 09:55:25 +0000 | [diff] [blame] | 497 | |
| 498 | 1) Use Cray's cc and not gcc. The latter was reported not to |
| 499 | work by Konrad Hinsen. It may work now, but it may not. |
| 500 | |
| 501 | 2) To set sys.platform to something sensible, pass the |
| 502 | following environment variable to the configure script: |
| 503 | |
| 504 | MACHDEP=unicosmk |
| 505 | |
Anthony Baxter | 70e0796 | 2006-04-03 14:16:27 +0000 | [diff] [blame] | 506 | 2) Run configure with option "--enable-unicode=ucs4". |
Michael W. Hudson | 33876f5 | 2002-07-31 09:55:25 +0000 | [diff] [blame] | 507 | |
Anthony Baxter | 70e0796 | 2006-04-03 14:16:27 +0000 | [diff] [blame] | 508 | 3) The Cray T3E does not support dynamic linking, so extension |
| 509 | modules have to be built by adding (or uncommenting) lines |
| 510 | in Modules/Setup. The minimum set of modules is |
Michael W. Hudson | 33876f5 | 2002-07-31 09:55:25 +0000 | [diff] [blame] | 511 | |
Anthony Baxter | 70e0796 | 2006-04-03 14:16:27 +0000 | [diff] [blame] | 512 | posix, new, _sre, unicodedata |
Michael W. Hudson | 33876f5 | 2002-07-31 09:55:25 +0000 | [diff] [blame] | 513 | |
Anthony Baxter | 70e0796 | 2006-04-03 14:16:27 +0000 | [diff] [blame] | 514 | On NIWA's vanilla T3E system the following have also been |
| 515 | included successfully: |
Michael W. Hudson | 33876f5 | 2002-07-31 09:55:25 +0000 | [diff] [blame] | 516 | |
Anthony Baxter | 70e0796 | 2006-04-03 14:16:27 +0000 | [diff] [blame] | 517 | _codecs, _locale, _socket, _symtable, _testcapi, _weakref |
| 518 | array, binascii, cmath, cPickle, crypt, cStringIO, dbm |
| 519 | errno, fcntl, grp, math, md5, operator, parser, pcre, pwd |
| 520 | regex, rotor, select, struct, strop, syslog, termios |
| 521 | time, timing, xreadlines |
Michael W. Hudson | 33876f5 | 2002-07-31 09:55:25 +0000 | [diff] [blame] | 522 | |
Anthony Baxter | 70e0796 | 2006-04-03 14:16:27 +0000 | [diff] [blame] | 523 | 4) Once the python executable and library have been built, make |
| 524 | will execute setup.py, which will attempt to build remaining |
| 525 | extensions and link them dynamically. Each of these attempts |
| 526 | will fail but should not halt the make process. This is |
| 527 | normal. |
Michael W. Hudson | 33876f5 | 2002-07-31 09:55:25 +0000 | [diff] [blame] | 528 | |
Anthony Baxter | 70e0796 | 2006-04-03 14:16:27 +0000 | [diff] [blame] | 529 | 5) Running "make test" uses a lot of resources and causes |
| 530 | problems on our system. You might want to try running tests |
| 531 | singly or in small groups. |
Guido van Rossum | 1bf0bf4 | 1997-08-20 23:50:51 +0000 | [diff] [blame] | 532 | |
Anthony Baxter | 70e0796 | 2006-04-03 14:16:27 +0000 | [diff] [blame] | 533 | SGI: SGI's standard "make" utility (/bin/make or /usr/bin/make) |
| 534 | does not check whether a command actually changed the file it |
| 535 | is supposed to build. This means that whenever you say "make" |
| 536 | it will redo the link step. The remedy is to use SGI's much |
| 537 | smarter "smake" utility (/usr/sbin/smake), or GNU make. If |
| 538 | you set the first line of the Makefile to #!/usr/sbin/smake |
| 539 | smake will be invoked by make (likewise for GNU make). |
Guido van Rossum | 0078aaf | 1997-08-21 03:05:11 +0000 | [diff] [blame] | 540 | |
Anthony Baxter | 70e0796 | 2006-04-03 14:16:27 +0000 | [diff] [blame] | 541 | WARNING: There are bugs in the optimizer of some versions of |
| 542 | SGI's compilers that can cause bus errors or other strange |
| 543 | behavior, especially on numerical operations. To avoid this, |
| 544 | try building with "make OPT=". |
Fred Drake | 0b5fb2b | 2000-09-29 17:45:05 +0000 | [diff] [blame] | 545 | |
Guido van Rossum | 6477380 | 1997-11-26 17:07:02 +0000 | [diff] [blame] | 546 | OS/2: If you are running Warp3 or Warp4 and have IBM's VisualAge C/C++ |
| 547 | compiler installed, just change into the pc\os2vacpp directory |
| 548 | and type NMAKE. Threading and sockets are supported by default |
| 549 | in the resulting binaries of PYTHON15.DLL and PYTHON.EXE. |
| 550 | |
Fred Drake | 36ed560 | 2000-10-06 01:58:48 +0000 | [diff] [blame] | 551 | Monterey (64-bit AIX): The current Monterey C compiler (Visual Age) |
| 552 | uses the OBJECT_MODE={32|64} environment variable to set the |
| 553 | compilation mode to either 32-bit or 64-bit (32-bit mode is |
| 554 | the default). Presumably you want 64-bit compilation mode for |
| 555 | this 64-bit OS. As a result you must first set OBJECT_MODE=64 |
| 556 | in your environment before configuring (./configure) or |
| 557 | building (make) Python on Monterey. |
| 558 | |
| 559 | Reliant UNIX: The thread support does not compile on Reliant UNIX, and |
| 560 | there is a (minor) problem in the configure script for that |
| 561 | platform as well. This should be resolved in time for a |
| 562 | future release. |
Trent Mick | 635f6fb | 2000-08-23 21:33:05 +0000 | [diff] [blame] | 563 | |
Barry Warsaw | 6a8557d | 2002-10-14 18:04:39 +0000 | [diff] [blame] | 564 | MacOSX: The tests will crash on both 10.1 and 10.2 with SEGV in |
Barry Warsaw | 6e75364 | 2002-10-14 18:15:35 +0000 | [diff] [blame] | 565 | test_re and test_sre due to the small default stack size. If |
| 566 | you set the stack size to 2048 before doing a "make test" the |
| 567 | failure can be avoided. If you're using the tcsh (the default |
| 568 | on OSX), or csh shells use "limit stacksize 2048" and for the |
| 569 | bash shell, use "ulimit -s 2048". |
Barry Warsaw | 0655745 | 2001-12-14 20:47:12 +0000 | [diff] [blame] | 570 | |
Jack Jansen | 6a77cf1 | 2001-12-06 21:49:02 +0000 | [diff] [blame] | 571 | On naked Darwin you may want to add the configure option |
Jack Jansen | 4c398fd | 2001-10-08 13:21:15 +0000 | [diff] [blame] | 572 | "--disable-toolbox-glue" to disable the glue code for the Carbon |
Jack Jansen | 6a77cf1 | 2001-12-06 21:49:02 +0000 | [diff] [blame] | 573 | interface modules. The modules themselves are currently only built |
| 574 | if you add the --enable-framework option, see below. |
Barry Warsaw | 0655745 | 2001-12-14 20:47:12 +0000 | [diff] [blame] | 575 | |
| 576 | On a clean OSX /usr/local does not exist. Do a |
Jack Jansen | 61fec30 | 2002-01-04 15:59:57 +0000 | [diff] [blame] | 577 | "sudo mkdir -m 775 /usr/local" |
Jack Jansen | 84a0164 | 2002-08-04 21:19:55 +0000 | [diff] [blame] | 578 | before you do a make install. It is probably not a good idea to |
| 579 | do "sudo make install" which installs everything as superuser, |
| 580 | as this may later cause problems when installing distutils-based |
| 581 | additions. |
Michael W. Hudson | 71dcc3e | 2005-02-22 15:33:26 +0000 | [diff] [blame] | 582 | |
Jack Jansen | 84a0164 | 2002-08-04 21:19:55 +0000 | [diff] [blame] | 583 | Some people have reported problems building Python after using "fink" |
Anthony Baxter | 70e0796 | 2006-04-03 14:16:27 +0000 | [diff] [blame] | 584 | to install additional unix software. Disabling fink (remove all |
| 585 | references to /sw from your .profile or .login) should solve this. |
Barry Warsaw | 0655745 | 2001-12-14 20:47:12 +0000 | [diff] [blame] | 586 | |
| 587 | You may want to try the configure option "--enable-framework" |
| 588 | which installs Python as a framework. The location can be set |
| 589 | as argument to the --enable-framework option (default |
Jack Jansen | 84a0164 | 2002-08-04 21:19:55 +0000 | [diff] [blame] | 590 | /Library/Frameworks). A framework install is probably needed if you |
| 591 | want to use any Aqua-based GUI toolkit (whether Tkinter, wxPython, |
| 592 | Carbon, Cocoa or anything else). |
Michael W. Hudson | 71dcc3e | 2005-02-22 15:33:26 +0000 | [diff] [blame] | 593 | |
Ronald Oussoren | 988117f | 2006-04-29 11:31:35 +0000 | [diff] [blame] | 594 | You may also want to try the configure option "--enable-universalsdk" |
| 595 | which builds Python as a universal binary with support for the |
| 596 | i386 and PPC architetures. This requires Xcode 2.1 or later to build. |
| 597 | |
| 598 | See Mac/OSX/README for more information on framework and |
| 599 | universal builds. |
Fred Drake | dabed75 | 2001-02-01 19:41:13 +0000 | [diff] [blame] | 600 | |
Michael W. Hudson | 1c4523f | 2001-12-19 19:49:58 +0000 | [diff] [blame] | 601 | Cygwin: With recent (relative to the time of writing, 2001-12-19) |
| 602 | Cygwin installations, there are problems with the interaction |
| 603 | of dynamic linking and fork(). This manifests itself in build |
| 604 | failures during the execution of setup.py. |
Tim Peters | 42fc7cc | 2001-04-04 18:35:19 +0000 | [diff] [blame] | 605 | |
Michael W. Hudson | 1c4523f | 2001-12-19 19:49:58 +0000 | [diff] [blame] | 606 | There are two workarounds that both enable Python (albeit |
| 607 | without threading support) to build and pass all tests on |
| 608 | NT/2000 (and most likely XP as well, though reports of testing |
| 609 | on XP would be appreciated). |
Guido van Rossum | 16e809e | 2001-04-16 02:07:08 +0000 | [diff] [blame] | 610 | |
Michael W. Hudson | 1c4523f | 2001-12-19 19:49:58 +0000 | [diff] [blame] | 611 | The workarounds: |
| 612 | |
| 613 | (a) the band-aid fix is to link the _socket module statically |
| 614 | rather than dynamically (which is the default). |
| 615 | |
| 616 | To do this, run "./configure --with-threads=no" including any |
| 617 | other options you need (--prefix, etc.). Then in Modules/Setup |
Walter Dörwald | 7dc65d6 | 2003-02-02 23:37:05 +0000 | [diff] [blame] | 618 | uncomment the lines: |
Tim Peters | 42fc7cc | 2001-04-04 18:35:19 +0000 | [diff] [blame] | 619 | |
Michael W. Hudson | be5ce18 | 2001-12-12 11:46:52 +0000 | [diff] [blame] | 620 | #SSL=/usr/local/ssl |
| 621 | #_socket socketmodule.c \ |
Anthony Baxter | 70e0796 | 2006-04-03 14:16:27 +0000 | [diff] [blame] | 622 | # -DUSE_SSL -I$(SSL)/include -I$(SSL)/include/openssl \ |
| 623 | # -L$(SSL)/lib -lssl -lcrypto |
Guido van Rossum | 16e809e | 2001-04-16 02:07:08 +0000 | [diff] [blame] | 624 | |
Michael W. Hudson | 1c4523f | 2001-12-19 19:49:58 +0000 | [diff] [blame] | 625 | and remove "local/" from the SSL variable. Finally, just run |
| 626 | "make"! |
Tim Peters | 42fc7cc | 2001-04-04 18:35:19 +0000 | [diff] [blame] | 627 | |
Michael W. Hudson | b655464 | 2001-12-19 22:09:09 +0000 | [diff] [blame] | 628 | (b) The "proper" fix is to rebase the Cygwin DLLs to prevent |
Michael W. Hudson | 1c4523f | 2001-12-19 19:49:58 +0000 | [diff] [blame] | 629 | base address conflicts. Details on how to do this can be |
| 630 | found in the following mail: |
| 631 | |
| 632 | http://sources.redhat.com/ml/cygwin/2001-12/msg00894.html |
| 633 | |
| 634 | It is hoped that a version of this solution will be |
Michael W. Hudson | b655464 | 2001-12-19 22:09:09 +0000 | [diff] [blame] | 635 | incorporated into the Cygwin distribution fairly soon. |
Michael W. Hudson | 1c4523f | 2001-12-19 19:49:58 +0000 | [diff] [blame] | 636 | |
| 637 | Two additional problems: |
| 638 | |
| 639 | (1) Threading support should still be disabled due to a known |
| 640 | bug in Cygwin pthreads that causes test_threadedtempfile to |
| 641 | hang. |
| 642 | |
Neal Norwitz | 4ebde09 | 2001-12-19 20:44:13 +0000 | [diff] [blame] | 643 | (2) The _curses module does not build. This is a known |
Michael W. Hudson | 1c4523f | 2001-12-19 19:49:58 +0000 | [diff] [blame] | 644 | Cygwin ncurses problem that should be resolved the next time |
| 645 | that this package is released. |
Guido van Rossum | 16e809e | 2001-04-16 02:07:08 +0000 | [diff] [blame] | 646 | |
Michael W. Hudson | be5ce18 | 2001-12-12 11:46:52 +0000 | [diff] [blame] | 647 | On older versions of Cygwin, test_poll may hang and test_strftime |
| 648 | may fail. |
Tim Peters | 42fc7cc | 2001-04-04 18:35:19 +0000 | [diff] [blame] | 649 | |
Michael W. Hudson | 1c4523f | 2001-12-19 19:49:58 +0000 | [diff] [blame] | 650 | The situation on 9X/Me is not accurately known at present. |
| 651 | Some time ago, there were reports that the following |
| 652 | regression tests failed: |
Tim Peters | 42fc7cc | 2001-04-04 18:35:19 +0000 | [diff] [blame] | 653 | |
| 654 | test_pwd |
| 655 | test_select (hang) |
| 656 | test_socket |
| 657 | |
Michael W. Hudson | 1c4523f | 2001-12-19 19:49:58 +0000 | [diff] [blame] | 658 | Due to the test_select hang on 9X/Me, one should run the |
| 659 | regression test using the following: |
Tim Peters | 42fc7cc | 2001-04-04 18:35:19 +0000 | [diff] [blame] | 660 | |
Michael W. Hudson | be5ce18 | 2001-12-12 11:46:52 +0000 | [diff] [blame] | 661 | make TESTOPTS='-l -x test_select' test |
Tim Peters | 42fc7cc | 2001-04-04 18:35:19 +0000 | [diff] [blame] | 662 | |
Michael W. Hudson | 1c4523f | 2001-12-19 19:49:58 +0000 | [diff] [blame] | 663 | News regarding these platforms with more recent Cygwin |
| 664 | versions would be appreciated! |
Guido van Rossum | 6d9cc80 | 1996-08-01 17:31:22 +0000 | [diff] [blame] | 665 | |
Skip Montanaro | 56d8301 | 2003-01-04 04:05:51 +0000 | [diff] [blame] | 666 | AtheOS: From Octavian Cerna <tavy at ylabs.com>: |
| 667 | |
Anthony Baxter | 70e0796 | 2006-04-03 14:16:27 +0000 | [diff] [blame] | 668 | Before building: |
Skip Montanaro | 56d8301 | 2003-01-04 04:05:51 +0000 | [diff] [blame] | 669 | |
Anthony Baxter | 70e0796 | 2006-04-03 14:16:27 +0000 | [diff] [blame] | 670 | Make sure you have shared versions of the libraries you |
| 671 | want to use with Python. You will have to compile them |
| 672 | yourself, or download precompiled packages. |
Skip Montanaro | 56d8301 | 2003-01-04 04:05:51 +0000 | [diff] [blame] | 673 | |
Anthony Baxter | 70e0796 | 2006-04-03 14:16:27 +0000 | [diff] [blame] | 674 | Recommended libraries: |
Skip Montanaro | 56d8301 | 2003-01-04 04:05:51 +0000 | [diff] [blame] | 675 | |
Anthony Baxter | 70e0796 | 2006-04-03 14:16:27 +0000 | [diff] [blame] | 676 | ncurses-4.2 |
| 677 | readline-4.2a |
| 678 | zlib-1.1.4 |
Skip Montanaro | 56d8301 | 2003-01-04 04:05:51 +0000 | [diff] [blame] | 679 | |
Anthony Baxter | 70e0796 | 2006-04-03 14:16:27 +0000 | [diff] [blame] | 680 | Build: |
Skip Montanaro | 56d8301 | 2003-01-04 04:05:51 +0000 | [diff] [blame] | 681 | |
Anthony Baxter | 70e0796 | 2006-04-03 14:16:27 +0000 | [diff] [blame] | 682 | $ ./configure --prefix=/usr/python |
| 683 | $ make |
Skip Montanaro | 56d8301 | 2003-01-04 04:05:51 +0000 | [diff] [blame] | 684 | |
Anthony Baxter | 70e0796 | 2006-04-03 14:16:27 +0000 | [diff] [blame] | 685 | Python is always built as a shared library, otherwise |
| 686 | dynamic loading would not work. |
Skip Montanaro | 56d8301 | 2003-01-04 04:05:51 +0000 | [diff] [blame] | 687 | |
Anthony Baxter | 70e0796 | 2006-04-03 14:16:27 +0000 | [diff] [blame] | 688 | Testing: |
Skip Montanaro | 56d8301 | 2003-01-04 04:05:51 +0000 | [diff] [blame] | 689 | |
Anthony Baxter | 70e0796 | 2006-04-03 14:16:27 +0000 | [diff] [blame] | 690 | $ make test |
Skip Montanaro | 56d8301 | 2003-01-04 04:05:51 +0000 | [diff] [blame] | 691 | |
Anthony Baxter | 70e0796 | 2006-04-03 14:16:27 +0000 | [diff] [blame] | 692 | Install: |
Skip Montanaro | 56d8301 | 2003-01-04 04:05:51 +0000 | [diff] [blame] | 693 | |
Anthony Baxter | 70e0796 | 2006-04-03 14:16:27 +0000 | [diff] [blame] | 694 | # make install |
| 695 | # pkgmanager -a /usr/python |
Skip Montanaro | 56d8301 | 2003-01-04 04:05:51 +0000 | [diff] [blame] | 696 | |
| 697 | |
Anthony Baxter | 70e0796 | 2006-04-03 14:16:27 +0000 | [diff] [blame] | 698 | AtheOS issues: |
Skip Montanaro | 56d8301 | 2003-01-04 04:05:51 +0000 | [diff] [blame] | 699 | |
Anthony Baxter | 70e0796 | 2006-04-03 14:16:27 +0000 | [diff] [blame] | 700 | - large file support: due to a stdio bug in glibc/libio, |
| 701 | access to large files may not work correctly. fseeko() |
| 702 | tries to seek to a negative offset. ftello() returns a |
| 703 | negative offset, it looks like a 32->64bit |
| 704 | sign-extension issue. The lowlevel functions (open, |
| 705 | lseek, etc) are OK. |
| 706 | - sockets: AF_UNIX is defined in the C library and in |
| 707 | Python, but not implemented in the system. |
| 708 | - select: poll is available in the C library, but does not |
| 709 | work (It does not return POLLNVAL for bad fds and |
| 710 | hangs). |
| 711 | - posix: statvfs and fstatvfs always return ENOSYS. |
| 712 | - disabled modules: |
| 713 | - mmap: not yet implemented in AtheOS |
| 714 | - nis: broken (on an unconfigured system |
| 715 | yp_get_default_domain() returns junk instead of |
| 716 | error) |
| 717 | - dl: dynamic loading doesn't work via dlopen() |
| 718 | - resource: getrimit and setrlimit are not yet |
| 719 | implemented |
Skip Montanaro | 56d8301 | 2003-01-04 04:05:51 +0000 | [diff] [blame] | 720 | |
Anthony Baxter | 70e0796 | 2006-04-03 14:16:27 +0000 | [diff] [blame] | 721 | - if you are getting segmentation faults, you probably are |
| 722 | low on memory. AtheOS doesn't handle very well an |
| 723 | out-of-memory condition and simply SEGVs the process. |
Skip Montanaro | 56d8301 | 2003-01-04 04:05:51 +0000 | [diff] [blame] | 724 | |
Anthony Baxter | 70e0796 | 2006-04-03 14:16:27 +0000 | [diff] [blame] | 725 | Tested on: |
Skip Montanaro | 56d8301 | 2003-01-04 04:05:51 +0000 | [diff] [blame] | 726 | |
Anthony Baxter | 70e0796 | 2006-04-03 14:16:27 +0000 | [diff] [blame] | 727 | AtheOS-0.3.7 |
| 728 | gcc-2.95 |
| 729 | binutils-2.10 |
| 730 | make-3.78 |
Skip Montanaro | 56d8301 | 2003-01-04 04:05:51 +0000 | [diff] [blame] | 731 | |
| 732 | |
Skip Montanaro | 57454e5 | 2002-06-14 20:30:31 +0000 | [diff] [blame] | 733 | Configuring the bsddb and dbm modules |
| 734 | ------------------------------------- |
| 735 | |
Skip Montanaro | 56d8301 | 2003-01-04 04:05:51 +0000 | [diff] [blame] | 736 | Beginning with Python version 2.3, the PyBsddb package |
| 737 | <http://pybsddb.sf.net/> was adopted into Python as the bsddb package, |
| 738 | exposing a set of package-level functions which provide |
Gregory P. Smith | 3adc4aa | 2006-04-13 19:19:01 +0000 | [diff] [blame] | 739 | backwards-compatible behavior. Only versions 3.3 through 4.4 of |
Skip Montanaro | 56d8301 | 2003-01-04 04:05:51 +0000 | [diff] [blame] | 740 | Sleepycat's libraries provide the necessary API, so older versions |
| 741 | aren't supported through this interface. The old bsddb module has |
| 742 | been retained as bsddb185, though it is not built by default. Users |
| 743 | wishing to use it will have to tweak Modules/Setup to build it. The |
| 744 | dbm module will still be built against the Sleepycat libraries if |
Gregory P. Smith | 3adc4aa | 2006-04-13 19:19:01 +0000 | [diff] [blame] | 745 | other preferred alternatives (ndbm, gdbm) are not found. |
Guido van Rossum | 62d45c0 | 2002-12-02 15:49:13 +0000 | [diff] [blame] | 746 | |
Anthony Baxter | 70e0796 | 2006-04-03 14:16:27 +0000 | [diff] [blame] | 747 | Building the sqlite3 module |
| 748 | --------------------------- |
| 749 | |
| 750 | To build the sqlite3 module, you'll need the sqlite3 or libsqlite3 |
| 751 | packages installed, including the header files. Many modern operating |
| 752 | systems distribute the headers in a separate package to the library - |
| 753 | often it will be the same name as the main package, but with a -dev or |
| 754 | -devel suffix. |
| 755 | |
| 756 | The version of pysqlite2 that's including in Python needs sqlite3 3.0.8 |
| 757 | or later. setup.py attempts to check that it can find a correct version. |
Skip Montanaro | 57454e5 | 2002-06-14 20:30:31 +0000 | [diff] [blame] | 758 | |
Guido van Rossum | dc8a3cb | 1998-05-12 15:29:18 +0000 | [diff] [blame] | 759 | Configuring threads |
| 760 | ------------------- |
| 761 | |
Guido van Rossum | 7ac1153 | 2000-08-29 21:36:40 +0000 | [diff] [blame] | 762 | As of Python 2.0, threads are enabled by default. If you wish to |
| 763 | compile without threads, or if your thread support is broken, pass the |
| 764 | --with-threads=no switch to configure. Unfortunately, on some |
| 765 | platforms, additional compiler and/or linker options are required for |
| 766 | threads to work properly. Below is a table of those options, |
Guido van Rossum | 4f40c2d | 2000-09-05 16:46:22 +0000 | [diff] [blame] | 767 | collected by Bill Janssen. We would love to automate this process |
Guido van Rossum | 7ac1153 | 2000-08-29 21:36:40 +0000 | [diff] [blame] | 768 | more, but the information below is not enough to write a patch for the |
| 769 | configure.in file, so manual intervention is required. If you patch |
| 770 | the configure.in file and are confident that the patch works, please |
Guido van Rossum | 4f40c2d | 2000-09-05 16:46:22 +0000 | [diff] [blame] | 771 | send in the patch. (Don't bother patching the configure script itself |
Walter Dörwald | 7dc65d6 | 2003-02-02 23:37:05 +0000 | [diff] [blame] | 772 | -- it is regenerated each time the configure.in file changes.) |
Guido van Rossum | dc8a3cb | 1998-05-12 15:29:18 +0000 | [diff] [blame] | 773 | |
| 774 | Compiler switches for threads |
| 775 | ............................. |
| 776 | |
Martin v. Löwis | 1c24f9c | 2001-03-06 12:07:27 +0000 | [diff] [blame] | 777 | The definition of _REENTRANT should be configured automatically, if |
| 778 | that does not work on your system, or if _REENTRANT is defined |
| 779 | incorrectly, please report that as a bug. |
Guido van Rossum | dc8a3cb | 1998-05-12 15:29:18 +0000 | [diff] [blame] | 780 | |
Martin v. Löwis | 1c24f9c | 2001-03-06 12:07:27 +0000 | [diff] [blame] | 781 | OS/Compiler/threads Switches for use with threads |
| 782 | (POSIX is draft 10, DCE is draft 4) compile & link |
| 783 | |
| 784 | SunOS 5.{1-5}/{gcc,SunPro cc}/solaris -mt |
| 785 | SunOS 5.5/{gcc,SunPro cc}/POSIX (nothing) |
Guido van Rossum | 16e809e | 2001-04-16 02:07:08 +0000 | [diff] [blame] | 786 | DEC OSF/1 3.x/cc/DCE -threads |
Anthony Baxter | 70e0796 | 2006-04-03 14:16:27 +0000 | [diff] [blame] | 787 | (butenhof@zko.dec.com) |
Guido van Rossum | 16e809e | 2001-04-16 02:07:08 +0000 | [diff] [blame] | 788 | Digital UNIX 4.x/cc/DCE -threads |
Anthony Baxter | 70e0796 | 2006-04-03 14:16:27 +0000 | [diff] [blame] | 789 | (butenhof@zko.dec.com) |
Guido van Rossum | 16e809e | 2001-04-16 02:07:08 +0000 | [diff] [blame] | 790 | Digital UNIX 4.x/cc/POSIX -pthread |
Anthony Baxter | 70e0796 | 2006-04-03 14:16:27 +0000 | [diff] [blame] | 791 | (butenhof@zko.dec.com) |
Guido van Rossum | 16e809e | 2001-04-16 02:07:08 +0000 | [diff] [blame] | 792 | AIX 4.1.4/cc_r/d7 (nothing) |
Anthony Baxter | 70e0796 | 2006-04-03 14:16:27 +0000 | [diff] [blame] | 793 | (buhrt@iquest.net) |
Guido van Rossum | 16e809e | 2001-04-16 02:07:08 +0000 | [diff] [blame] | 794 | AIX 4.1.4/cc_r4/DCE (nothing) |
Anthony Baxter | 70e0796 | 2006-04-03 14:16:27 +0000 | [diff] [blame] | 795 | (buhrt@iquest.net) |
Guido van Rossum | 16e809e | 2001-04-16 02:07:08 +0000 | [diff] [blame] | 796 | IRIX 6.2/cc/POSIX (nothing) |
Anthony Baxter | 70e0796 | 2006-04-03 14:16:27 +0000 | [diff] [blame] | 797 | (robertl@cwi.nl) |
Guido van Rossum | dc8a3cb | 1998-05-12 15:29:18 +0000 | [diff] [blame] | 798 | |
| 799 | |
| 800 | Linker (ld) libraries and flags for threads |
| 801 | ........................................... |
| 802 | |
| 803 | OS/threads Libraries/switches for use with threads |
| 804 | |
| 805 | SunOS 5.{1-5}/solaris -lthread |
| 806 | SunOS 5.5/POSIX -lpthread |
Guido van Rossum | 16e809e | 2001-04-16 02:07:08 +0000 | [diff] [blame] | 807 | DEC OSF/1 3.x/DCE -lpthreads -lmach -lc_r -lc |
Anthony Baxter | 70e0796 | 2006-04-03 14:16:27 +0000 | [diff] [blame] | 808 | (butenhof@zko.dec.com) |
Guido van Rossum | dc8a3cb | 1998-05-12 15:29:18 +0000 | [diff] [blame] | 809 | Digital UNIX 4.x/DCE -lpthreads -lpthread -lmach -lexc -lc |
Anthony Baxter | 70e0796 | 2006-04-03 14:16:27 +0000 | [diff] [blame] | 810 | (butenhof@zko.dec.com) |
Guido van Rossum | 16e809e | 2001-04-16 02:07:08 +0000 | [diff] [blame] | 811 | Digital UNIX 4.x/POSIX -lpthread -lmach -lexc -lc |
Anthony Baxter | 70e0796 | 2006-04-03 14:16:27 +0000 | [diff] [blame] | 812 | (butenhof@zko.dec.com) |
Guido van Rossum | 16e809e | 2001-04-16 02:07:08 +0000 | [diff] [blame] | 813 | AIX 4.1.4/{draft7,DCE} (nothing) |
Anthony Baxter | 70e0796 | 2006-04-03 14:16:27 +0000 | [diff] [blame] | 814 | (buhrt@iquest.net) |
Guido van Rossum | 16e809e | 2001-04-16 02:07:08 +0000 | [diff] [blame] | 815 | IRIX 6.2/POSIX -lpthread |
Anthony Baxter | 70e0796 | 2006-04-03 14:16:27 +0000 | [diff] [blame] | 816 | (jph@emilia.engr.sgi.com) |
Guido van Rossum | dc8a3cb | 1998-05-12 15:29:18 +0000 | [diff] [blame] | 817 | |
| 818 | |
Martin v. Löwis | 1142de3 | 2002-03-29 16:28:31 +0000 | [diff] [blame] | 819 | Building a shared libpython |
| 820 | --------------------------- |
| 821 | |
| 822 | Starting with Python 2.3, the majority of the interpreter can be built |
Michael W. Hudson | 71dcc3e | 2005-02-22 15:33:26 +0000 | [diff] [blame] | 823 | into a shared library, which can then be used by the interpreter |
Martin v. Löwis | 1142de3 | 2002-03-29 16:28:31 +0000 | [diff] [blame] | 824 | executable, and by applications embedding Python. To enable this feature, |
Michael W. Hudson | 71dcc3e | 2005-02-22 15:33:26 +0000 | [diff] [blame] | 825 | configure with --enable-shared. |
Martin v. Löwis | 6506967 | 2002-08-03 21:38:27 +0000 | [diff] [blame] | 826 | |
Fred Drake | 55512dc | 2003-07-21 16:01:39 +0000 | [diff] [blame] | 827 | If you enable this feature, the same object files will be used to create |
| 828 | a static library. In particular, the static library will contain object |
| 829 | files using position-independent code (PIC) on platforms where PIC flags |
| 830 | are needed for the shared library. |
Martin v. Löwis | 1142de3 | 2002-03-29 16:28:31 +0000 | [diff] [blame] | 831 | |
| 832 | |
Guido van Rossum | 6d9cc80 | 1996-08-01 17:31:22 +0000 | [diff] [blame] | 833 | Configuring additional built-in modules |
Guido van Rossum | 91cb9d2 | 1995-04-10 11:47:38 +0000 | [diff] [blame] | 834 | --------------------------------------- |
Guido van Rossum | 19e0c26 | 1995-01-17 16:36:34 +0000 | [diff] [blame] | 835 | |
Andrew M. Kuchling | ebe3045 | 2001-02-27 03:24:24 +0000 | [diff] [blame] | 836 | Starting with Python 2.1, the setup.py script at the top of the source |
| 837 | distribution attempts to detect which modules can be built and |
| 838 | automatically compiles them. Autodetection doesn't always work, so |
Guido van Rossum | 59b0b60 | 2002-02-08 22:22:35 +0000 | [diff] [blame] | 839 | you can still customize the configuration by editing the Modules/Setup |
| 840 | file; but this should be considered a last resort. The rest of this |
| 841 | section only applies if you decide to edit the Modules/Setup file. |
| 842 | You also need this to enable static linking of certain modules (which |
| 843 | is needed to enable profiling on some systems). |
| 844 | |
Andrew M. Kuchling | ebe3045 | 2001-02-27 03:24:24 +0000 | [diff] [blame] | 845 | This file is initially copied from Setup.dist by the configure script; |
| 846 | if it does not exist yet, create it by copying Modules/Setup.dist |
| 847 | yourself (configure will never overwrite it). Never edit Setup.dist |
| 848 | -- always edit Setup or Setup.local (see below). Read the comments in |
| 849 | the file for information on what kind of edits are allowed. When you |
| 850 | have edited Setup in the Modules directory, the interpreter will |
Guido van Rossum | ef67ded | 2001-10-05 17:04:42 +0000 | [diff] [blame] | 851 | automatically be rebuilt the next time you run make (in the toplevel |
| 852 | directory). |
Guido van Rossum | 627b2d7 | 1993-12-24 10:39:16 +0000 | [diff] [blame] | 853 | |
Andrew M. Kuchling | ebe3045 | 2001-02-27 03:24:24 +0000 | [diff] [blame] | 854 | Many useful modules can be built on any Unix system, but some optional |
| 855 | modules can't be reliably autodetected. Often the quickest way to |
| 856 | determine whether a particular module works or not is to see if it |
| 857 | will build: enable it in Setup, then if you get compilation or link |
Fred Drake | 6caae14 | 2000-10-25 17:51:02 +0000 | [diff] [blame] | 858 | errors, disable it -- you're either missing support or need to adjust |
| 859 | the compilation and linking parameters for that module. |
Guido van Rossum | 6d9cc80 | 1996-08-01 17:31:22 +0000 | [diff] [blame] | 860 | |
| 861 | On SGI IRIX, there are modules that interface to many SGI specific |
Andrew M. Kuchling | ebe3045 | 2001-02-27 03:24:24 +0000 | [diff] [blame] | 862 | system libraries, e.g. the GL library and the audio hardware. These |
| 863 | modules will not be built by the setup.py script. |
Guido van Rossum | 91cb9d2 | 1995-04-10 11:47:38 +0000 | [diff] [blame] | 864 | |
Guido van Rossum | 4462e93 | 1997-01-22 21:00:32 +0000 | [diff] [blame] | 865 | In addition to the file Setup, you can also edit the file Setup.local. |
| 866 | (the makesetup script processes both). You may find it more |
| 867 | convenient to edit Setup.local and leave Setup alone. Then, when |
| 868 | installing a new Python version, you can copy your old Setup.local |
| 869 | file. |
| 870 | |
Guido van Rossum | 91cb9d2 | 1995-04-10 11:47:38 +0000 | [diff] [blame] | 871 | |
| 872 | Setting the optimization/debugging options |
| 873 | ------------------------------------------ |
| 874 | |
Guido van Rossum | 6d9cc80 | 1996-08-01 17:31:22 +0000 | [diff] [blame] | 875 | If you want or need to change the optimization/debugging options for |
| 876 | the C compiler, assign to the OPT variable on the toplevel make |
| 877 | command; e.g. "make OPT=-g" will build a debugging version of Python |
| 878 | on most platforms. The default is OPT=-O; a value for OPT in the |
| 879 | environment when the configure script is run overrides this default |
| 880 | (likewise for CC; and the initial value for LIBS is used as the base |
| 881 | set of libraries to link with). |
Guido van Rossum | 91cb9d2 | 1995-04-10 11:47:38 +0000 | [diff] [blame] | 882 | |
Fred Drake | 6caae14 | 2000-10-25 17:51:02 +0000 | [diff] [blame] | 883 | When compiling with GCC, the default value of OPT will also include |
| 884 | the -Wall and -Wstrict-prototypes options. |
| 885 | |
| 886 | Additional debugging code to help debug memory management problems can |
| 887 | be enabled by using the --with-pydebug option to the configure script. |
| 888 | |
Brett Cannon | 08cd598 | 2005-04-24 22:26:38 +0000 | [diff] [blame] | 889 | For flags that change binary compatibility, use the EXTRA_CFLAGS |
| 890 | variable. |
| 891 | |
Guido van Rossum | 91cb9d2 | 1995-04-10 11:47:38 +0000 | [diff] [blame] | 892 | |
Guido van Rossum | a68ca35 | 2001-10-04 17:00:07 +0000 | [diff] [blame] | 893 | Profiling |
| 894 | --------- |
| 895 | |
| 896 | If you want C profiling turned on, the easiest way is to run configure |
| 897 | with the CC environment variable to the necessary compiler |
| 898 | invocation. For example, on Linux, this works for profiling using |
| 899 | gprof(1): |
| 900 | |
| 901 | CC="gcc -pg" ./configure |
| 902 | |
| 903 | Note that on Linux, gprof apparently does not work for shared |
| 904 | libraries. The Makefile/Setup mechanism can be used to compile and |
Walter Dörwald | 7dc65d6 | 2003-02-02 23:37:05 +0000 | [diff] [blame] | 905 | link most extension modules statically. |
Guido van Rossum | a68ca35 | 2001-10-04 17:00:07 +0000 | [diff] [blame] | 906 | |
| 907 | |
Guido van Rossum | 91cb9d2 | 1995-04-10 11:47:38 +0000 | [diff] [blame] | 908 | Testing |
| 909 | ------- |
Guido van Rossum | 433c8ad | 1994-08-01 12:07:07 +0000 | [diff] [blame] | 910 | |
Guido van Rossum | 4f40c2d | 2000-09-05 16:46:22 +0000 | [diff] [blame] | 911 | To test the interpreter, type "make test" in the top-level directory. |
Guido van Rossum | 6ae5d3d | 1997-05-14 21:39:05 +0000 | [diff] [blame] | 912 | This runs the test set twice (once with no compiled files, once with |
| 913 | the compiled files left by the previous test run). The test set |
| 914 | produces some output. You can generally ignore the messages about |
Guido van Rossum | 59b0b60 | 2002-02-08 22:22:35 +0000 | [diff] [blame] | 915 | skipped tests due to optional features which can't be imported. |
Guido van Rossum | 4f40c2d | 2000-09-05 16:46:22 +0000 | [diff] [blame] | 916 | If a message is printed about a failed test or a traceback or core |
| 917 | dump is produced, something is wrong. On some Linux systems (those |
Guido van Rossum | 24df684 | 1997-12-30 04:32:30 +0000 | [diff] [blame] | 918 | that are not yet using glibc 6), test_strftime fails due to a |
Guido van Rossum | 4f40c2d | 2000-09-05 16:46:22 +0000 | [diff] [blame] | 919 | non-standard implementation of strftime() in the C library. Please |
| 920 | ignore this, or upgrade to glibc version 6. |
Guido van Rossum | 433c8ad | 1994-08-01 12:07:07 +0000 | [diff] [blame] | 921 | |
Guido van Rossum | 91cb9d2 | 1995-04-10 11:47:38 +0000 | [diff] [blame] | 922 | IMPORTANT: If the tests fail and you decide to mail a bug report, |
| 923 | *don't* include the output of "make test". It is useless. Run the |
Guido van Rossum | 4f40c2d | 2000-09-05 16:46:22 +0000 | [diff] [blame] | 924 | failing test manually, as follows: |
Guido van Rossum | 91cb9d2 | 1995-04-10 11:47:38 +0000 | [diff] [blame] | 925 | |
Anthony Baxter | 70e0796 | 2006-04-03 14:16:27 +0000 | [diff] [blame] | 926 | ./python ./Lib/test/test_whatever.py |
Guido van Rossum | 91cb9d2 | 1995-04-10 11:47:38 +0000 | [diff] [blame] | 927 | |
Andrew M. Kuchling | 83d042d | 2002-03-21 23:52:20 +0000 | [diff] [blame] | 928 | (substituting the top of the source tree for '.' if you built in a |
Guido van Rossum | 6ae5d3d | 1997-05-14 21:39:05 +0000 | [diff] [blame] | 929 | different directory). This runs the test in verbose mode. |
Guido van Rossum | 91cb9d2 | 1995-04-10 11:47:38 +0000 | [diff] [blame] | 930 | |
| 931 | |
| 932 | Installing |
| 933 | ---------- |
| 934 | |
Guido van Rossum | 6d9cc80 | 1996-08-01 17:31:22 +0000 | [diff] [blame] | 935 | To install the Python binary, library modules, shared library modules |
| 936 | (see below), include files, configuration files, and the manual page, |
Guido van Rossum | 6477380 | 1997-11-26 17:07:02 +0000 | [diff] [blame] | 937 | just type |
| 938 | |
Anthony Baxter | 70e0796 | 2006-04-03 14:16:27 +0000 | [diff] [blame] | 939 | make install |
Guido van Rossum | 6477380 | 1997-11-26 17:07:02 +0000 | [diff] [blame] | 940 | |
Guido van Rossum | 4f40c2d | 2000-09-05 16:46:22 +0000 | [diff] [blame] | 941 | This will install all platform-independent files in subdirectories of |
| 942 | the directory given with the --prefix option to configure or to the |
| 943 | `prefix' Make variable (default /usr/local). All binary and other |
| 944 | platform-specific files will be installed in subdirectories if the |
| 945 | directory given by --exec-prefix or the `exec_prefix' Make variable |
| 946 | (defaults to the --prefix directory) is given. |
Guido van Rossum | 6477380 | 1997-11-26 17:07:02 +0000 | [diff] [blame] | 947 | |
Martin v. Löwis | 3b8ee08 | 2003-05-11 20:25:35 +0000 | [diff] [blame] | 948 | If DESTDIR is set, it will be taken as the root directory of the |
| 949 | installation, and files will be installed into $(DESTDIR)$(prefix), |
| 950 | $(DESTDIR)$(exec_prefix), etc. |
| 951 | |
Guido van Rossum | 6477380 | 1997-11-26 17:07:02 +0000 | [diff] [blame] | 952 | All subdirectories created will have Python's version number in their |
| 953 | name, e.g. the library modules are installed in |
Guido van Rossum | 67345eb | 2001-01-18 14:51:12 +0000 | [diff] [blame] | 954 | "/usr/local/lib/python<version>/" by default, where <version> is the |
| 955 | <major>.<minor> release number (e.g. "2.1"). The Python binary is |
| 956 | installed as "python<version>" and a hard link named "python" is |
| 957 | created. The only file not installed with a version number in its |
| 958 | name is the manual page, installed as "/usr/local/man/man1/python.1" |
| 959 | by default. |
Guido van Rossum | 433c8ad | 1994-08-01 12:07:07 +0000 | [diff] [blame] | 960 | |
Neil Schemenauer | 0e91dc7 | 2001-02-16 04:18:08 +0000 | [diff] [blame] | 961 | If you have a previous installation of Python that you don't |
Guido van Rossum | 6477380 | 1997-11-26 17:07:02 +0000 | [diff] [blame] | 962 | want to replace yet, use |
| 963 | |
Anthony Baxter | 70e0796 | 2006-04-03 14:16:27 +0000 | [diff] [blame] | 964 | make altinstall |
Guido van Rossum | 6477380 | 1997-11-26 17:07:02 +0000 | [diff] [blame] | 965 | |
| 966 | This installs the same set of files as "make install" except it |
Guido van Rossum | 67345eb | 2001-01-18 14:51:12 +0000 | [diff] [blame] | 967 | doesn't create the hard link to "python<version>" named "python" and |
| 968 | it doesn't install the manual page at all. |
Guido van Rossum | 6d9cc80 | 1996-08-01 17:31:22 +0000 | [diff] [blame] | 969 | |
| 970 | The only thing you may have to install manually is the Python mode for |
Guido van Rossum | 4f40c2d | 2000-09-05 16:46:22 +0000 | [diff] [blame] | 971 | Emacs found in Misc/python-mode.el. (But then again, more recent |
| 972 | versions of Emacs may already have it.) Follow the instructions that |
| 973 | came with Emacs for installation of site-specific files. |
Guido van Rossum | 433c8ad | 1994-08-01 12:07:07 +0000 | [diff] [blame] | 974 | |
Jack Jansen | a39ef86 | 2001-08-19 21:17:03 +0000 | [diff] [blame] | 975 | On Mac OS X, if you have configured Python with --enable-framework, you |
| 976 | should use "make frameworkinstall" to do the installation. Note that this |
| 977 | installs the Python executable in a place that is not normally on your |
| 978 | PATH, you may want to set up a symlink in /usr/local/bin. |
| 979 | |
Guido van Rossum | 91cb9d2 | 1995-04-10 11:47:38 +0000 | [diff] [blame] | 980 | |
| 981 | Configuration options and variables |
| 982 | ----------------------------------- |
Guido van Rossum | 433c8ad | 1994-08-01 12:07:07 +0000 | [diff] [blame] | 983 | |
Guido van Rossum | 6d9cc80 | 1996-08-01 17:31:22 +0000 | [diff] [blame] | 984 | Some special cases are handled by passing options to the configure |
| 985 | script. |
Guido van Rossum | 433c8ad | 1994-08-01 12:07:07 +0000 | [diff] [blame] | 986 | |
Guido van Rossum | 6d9cc80 | 1996-08-01 17:31:22 +0000 | [diff] [blame] | 987 | WARNING: if you rerun the configure script with different options, you |
| 988 | must run "make clean" before rebuilding. Exceptions to this rule: |
| 989 | after changing --prefix or --exec-prefix, all you need to do is remove |
Guido van Rossum | b06df27 | 1997-08-05 21:50:20 +0000 | [diff] [blame] | 990 | Modules/getpath.o. |
Guido van Rossum | 433c8ad | 1994-08-01 12:07:07 +0000 | [diff] [blame] | 991 | |
Guido van Rossum | 91cb9d2 | 1995-04-10 11:47:38 +0000 | [diff] [blame] | 992 | --with(out)-gcc: The configure script uses gcc (the GNU C compiler) if |
Anthony Baxter | 70e0796 | 2006-04-03 14:16:27 +0000 | [diff] [blame] | 993 | it finds it. If you don't want this, or if this compiler is |
| 994 | installed but broken on your platform, pass the option |
| 995 | --without-gcc. You can also pass "CC=cc" (or whatever the |
| 996 | name of the proper C compiler is) in the environment, but the |
| 997 | advantage of using --without-gcc is that this option is |
| 998 | remembered by the config.status script for its --recheck |
| 999 | option. |
Guido van Rossum | 76be6ed | 1995-01-02 18:33:54 +0000 | [diff] [blame] | 1000 | |
Guido van Rossum | 91cb9d2 | 1995-04-10 11:47:38 +0000 | [diff] [blame] | 1001 | --prefix, --exec-prefix: If you want to install the binaries and the |
Anthony Baxter | 70e0796 | 2006-04-03 14:16:27 +0000 | [diff] [blame] | 1002 | Python library somewhere else than in /usr/local/{bin,lib}, |
| 1003 | you can pass the option --prefix=DIRECTORY; the interpreter |
| 1004 | binary will be installed as DIRECTORY/bin/python and the |
| 1005 | library files as DIRECTORY/lib/python/*. If you pass |
| 1006 | --exec-prefix=DIRECTORY (as well) this overrides the |
| 1007 | installation prefix for architecture-dependent files (like the |
| 1008 | interpreter binary). Note that --prefix=DIRECTORY also |
| 1009 | affects the default module search path (sys.path), when |
| 1010 | Modules/config.c is compiled. Passing make the option |
| 1011 | prefix=DIRECTORY (and/or exec_prefix=DIRECTORY) overrides the |
| 1012 | prefix set at configuration time; this may be more convenient |
| 1013 | than re-running the configure script if you change your mind |
| 1014 | about the install prefix. |
Guido van Rossum | 76be6ed | 1995-01-02 18:33:54 +0000 | [diff] [blame] | 1015 | |
Guido van Rossum | 59b0b60 | 2002-02-08 22:22:35 +0000 | [diff] [blame] | 1016 | --with-readline: This option is no longer supported. GNU |
Anthony Baxter | 70e0796 | 2006-04-03 14:16:27 +0000 | [diff] [blame] | 1017 | readline is automatically enabled by setup.py when present. |
Guido van Rossum | 91cb9d2 | 1995-04-10 11:47:38 +0000 | [diff] [blame] | 1018 | |
Guido van Rossum | 7ac1153 | 2000-08-29 21:36:40 +0000 | [diff] [blame] | 1019 | --with-threads: On most Unix systems, you can now use multiple |
Anthony Baxter | 70e0796 | 2006-04-03 14:16:27 +0000 | [diff] [blame] | 1020 | threads, and support for this is enabled by default. To |
| 1021 | disable this, pass --with-threads=no. If the library required |
| 1022 | for threads lives in a peculiar place, you can use |
| 1023 | --with-thread=DIRECTORY. IMPORTANT: run "make clean" after |
| 1024 | changing (either enabling or disabling) this option, or you |
| 1025 | will get link errors! Note: for DEC Unix use |
| 1026 | --with-dec-threads instead. |
Guido van Rossum | 91cb9d2 | 1995-04-10 11:47:38 +0000 | [diff] [blame] | 1027 | |
| 1028 | --with-sgi-dl: On SGI IRIX 4, dynamic loading of extension modules is |
Anthony Baxter | 70e0796 | 2006-04-03 14:16:27 +0000 | [diff] [blame] | 1029 | supported by the "dl" library by Jack Jansen, which is |
| 1030 | ftp'able from ftp://ftp.cwi.nl/pub/dynload/dl-1.6.tar.Z. |
| 1031 | This is enabled (after you've ftp'ed and compiled the dl |
| 1032 | library) by passing --with-sgi-dl=DIRECTORY where DIRECTORY |
| 1033 | is the absolute pathname of the dl library. (Don't bother on |
| 1034 | IRIX 5, it already has dynamic linking using SunOS style |
| 1035 | shared libraries.) THIS OPTION IS UNSUPPORTED. |
Guido van Rossum | 91cb9d2 | 1995-04-10 11:47:38 +0000 | [diff] [blame] | 1036 | |
Guido van Rossum | 4f40c2d | 2000-09-05 16:46:22 +0000 | [diff] [blame] | 1037 | --with-dl-dld: Dynamic loading of modules is rumored to be supported |
Anthony Baxter | 70e0796 | 2006-04-03 14:16:27 +0000 | [diff] [blame] | 1038 | on some other systems: VAX (Ultrix), Sun3 (SunOS 3.4), Sequent |
| 1039 | Symmetry (Dynix), and Atari ST. This is done using a |
| 1040 | combination of the GNU dynamic loading package |
| 1041 | (ftp://ftp.cwi.nl/pub/dynload/dl-dld-1.1.tar.Z) and an |
| 1042 | emulation of the SGI dl library mentioned above (the emulation |
| 1043 | can be found at |
| 1044 | ftp://ftp.cwi.nl/pub/dynload/dld-3.2.3.tar.Z). To |
| 1045 | enable this, ftp and compile both libraries, then call |
| 1046 | configure, passing it the option |
| 1047 | --with-dl-dld=DL_DIRECTORY,DLD_DIRECTORY where DL_DIRECTORY is |
| 1048 | the absolute pathname of the dl emulation library and |
| 1049 | DLD_DIRECTORY is the absolute pathname of the GNU dld library. |
| 1050 | (Don't bother on SunOS 4 or 5, they already have dynamic |
| 1051 | linking using shared libraries.) THIS OPTION IS UNSUPPORTED. |
Guido van Rossum | 91cb9d2 | 1995-04-10 11:47:38 +0000 | [diff] [blame] | 1052 | |
| 1053 | --with-libm, --with-libc: It is possible to specify alternative |
Anthony Baxter | 70e0796 | 2006-04-03 14:16:27 +0000 | [diff] [blame] | 1054 | versions for the Math library (default -lm) and the C library |
| 1055 | (default the empty string) using the options |
| 1056 | --with-libm=STRING and --with-libc=STRING, respectively. For |
| 1057 | example, if your system requires that you pass -lc_s to the C |
| 1058 | compiler to use the shared C library, you can pass |
| 1059 | --with-libc=-lc_s. These libraries are passed after all other |
| 1060 | libraries, the C library last. |
Guido van Rossum | 16e809e | 2001-04-16 02:07:08 +0000 | [diff] [blame] | 1061 | |
Guido van Rossum | 4f40c2d | 2000-09-05 16:46:22 +0000 | [diff] [blame] | 1062 | --with-libs='libs': Add 'libs' to the LIBS that the python interpreter |
Anthony Baxter | 70e0796 | 2006-04-03 14:16:27 +0000 | [diff] [blame] | 1063 | is linked against. |
Guido van Rossum | d02ba45 | 1996-07-31 17:36:01 +0000 | [diff] [blame] | 1064 | |
Martin v. Löwis | 0f48d98 | 2006-04-14 14:34:26 +0000 | [diff] [blame] | 1065 | --with-cxx-main=<compiler>: If you plan to use C++ extension modules, |
| 1066 | then -- on some platforms -- you need to compile python's main() |
| 1067 | function with the C++ compiler. With this option, make will use |
| 1068 | <compiler> to compile main() *and* to link the python executable. |
| 1069 | It is likely that the resulting executable depends on the C++ |
| 1070 | runtime library of <compiler>. (The default is --without-cxx-main.) |
| 1071 | |
| 1072 | There are platforms that do not require you to build Python |
| 1073 | with a C++ compiler in order to use C++ extension modules. |
| 1074 | E.g., x86 Linux with ELF shared binaries and GCC 3.x, 4.x is such |
| 1075 | a platform. We recommend that you configure Python |
| 1076 | --without-cxx-main on those platforms because a mismatch |
| 1077 | between the C++ compiler version used to build Python and to |
| 1078 | build a C++ extension module is likely to cause a crash at |
| 1079 | runtime. |
| 1080 | |
| 1081 | The Python installation also stores the variable CXX that |
| 1082 | determines, e.g., the C++ compiler distutils calls by default |
| 1083 | to build C++ extensions. If you set CXX on the configure command |
| 1084 | line to any string of non-zero length, then configure won't |
| 1085 | change CXX. If you do not preset CXX but pass |
| 1086 | --with-cxx-main=<compiler>, then configure sets CXX=<compiler>. |
| 1087 | In all other cases, configure looks for a C++ compiler by |
| 1088 | some common names (c++, g++, gcc, CC, cxx, cc++, cl) and sets |
| 1089 | CXX to the first compiler it finds. If it does not find any |
| 1090 | C++ compiler, then it sets CXX="". |
| 1091 | |
| 1092 | Similarly, if you want to change the command used to link the |
| 1093 | python executable, then set LINKCC on the configure command line. |
Martin v. Löwis | 2df6694 | 2000-12-13 14:14:32 +0000 | [diff] [blame] | 1094 | |
| 1095 | |
Fred Drake | 6caae14 | 2000-10-25 17:51:02 +0000 | [diff] [blame] | 1096 | --with-pydebug: Enable additional debugging code to help track down |
Anthony Baxter | 70e0796 | 2006-04-03 14:16:27 +0000 | [diff] [blame] | 1097 | memory management problems. This allows printing a list of all |
| 1098 | live objects when the interpreter terminates. |
Michael W. Hudson | 71dcc3e | 2005-02-22 15:33:26 +0000 | [diff] [blame] | 1099 | |
Jack Jansen | 7b8c754 | 2002-04-14 20:12:41 +0000 | [diff] [blame] | 1100 | --with(out)-universal-newlines: enable reading of text files with |
Anthony Baxter | 70e0796 | 2006-04-03 14:16:27 +0000 | [diff] [blame] | 1101 | foreign newline convention (default: enabled). In other words, |
| 1102 | any of \r, \n or \r\n is acceptable as end-of-line character. |
| 1103 | If enabled import and execfile will automatically accept any newline |
| 1104 | in files. Python code can open a file with open(file, 'U') to |
| 1105 | read it in universal newline mode. THIS OPTION IS UNSUPPORTED. |
Fred Drake | 6caae14 | 2000-10-25 17:51:02 +0000 | [diff] [blame] | 1106 | |
Martin v. Löwis | f30d60e | 2004-06-08 08:17:44 +0000 | [diff] [blame] | 1107 | --with-tsc: Profile using the Pentium timestamping counter (TSC). |
Guido van Rossum | 76be6ed | 1995-01-02 18:33:54 +0000 | [diff] [blame] | 1108 | |
Martin v. Löwis | 9176fc1 | 2006-04-11 11:12:43 +0000 | [diff] [blame] | 1109 | --with-system-ffi: Build the _ctypes extension module using an ffi |
| 1110 | library installed on the system. |
| 1111 | |
Michael W. Hudson | 71dcc3e | 2005-02-22 15:33:26 +0000 | [diff] [blame] | 1112 | |
Guido van Rossum | 433c8ad | 1994-08-01 12:07:07 +0000 | [diff] [blame] | 1113 | Building for multiple architectures (using the VPATH feature) |
| 1114 | ------------------------------------------------------------- |
| 1115 | |
| 1116 | If your file system is shared between multiple architectures, it |
| 1117 | usually is not necessary to make copies of the sources for each |
| 1118 | architecture you want to support. If the make program supports the |
| 1119 | VPATH feature, you can create an empty build directory for each |
| 1120 | architecture, and in each directory run the configure script (on the |
| 1121 | appropriate machine with the appropriate options). This creates the |
| 1122 | necessary subdirectories and the Makefiles therein. The Makefiles |
Guido van Rossum | 4f40c2d | 2000-09-05 16:46:22 +0000 | [diff] [blame] | 1123 | contain a line VPATH=... which points to a directory containing the |
Guido van Rossum | d0fe845 | 1996-08-26 03:02:37 +0000 | [diff] [blame] | 1124 | actual sources. (On SGI systems, use "smake -J1" instead of "make" if |
| 1125 | you use VPATH -- don't try gnumake.) |
Guido van Rossum | 433c8ad | 1994-08-01 12:07:07 +0000 | [diff] [blame] | 1126 | |
| 1127 | For example, the following is all you need to build a minimal Python |
| 1128 | in /usr/tmp/python (assuming ~guido/src/python is the toplevel |
| 1129 | directory and you want to build in /usr/tmp/python): |
| 1130 | |
Anthony Baxter | 70e0796 | 2006-04-03 14:16:27 +0000 | [diff] [blame] | 1131 | $ mkdir /usr/tmp/python |
| 1132 | $ cd /usr/tmp/python |
| 1133 | $ ~guido/src/python/configure |
| 1134 | [...] |
| 1135 | $ make |
| 1136 | [...] |
| 1137 | $ |
Guido van Rossum | 433c8ad | 1994-08-01 12:07:07 +0000 | [diff] [blame] | 1138 | |
Neil Schemenauer | 73f8ab2 | 2001-01-26 22:18:55 +0000 | [diff] [blame] | 1139 | Note that configure copies the original Setup file to the build |
Guido van Rossum | 433c8ad | 1994-08-01 12:07:07 +0000 | [diff] [blame] | 1140 | directory if it finds no Setup file there. This means that you can |
| 1141 | edit the Setup file for each architecture independently. For this |
| 1142 | reason, subsequent changes to the original Setup file are not tracked |
| 1143 | automatically, as they might overwrite local changes. To force a copy |
| 1144 | of a changed original Setup file, delete the target Setup file. (The |
| 1145 | makesetup script supports multiple input files, so if you want to be |
| 1146 | fancy you can change the rules to create an empty Setup.local if it |
| 1147 | doesn't exist and run it with arguments $(srcdir)/Setup Setup.local; |
| 1148 | however this assumes that you only need to add modules.) |
| 1149 | |
| 1150 | |
Guido van Rossum | 433c8ad | 1994-08-01 12:07:07 +0000 | [diff] [blame] | 1151 | Building on non-UNIX systems |
| 1152 | ---------------------------- |
| 1153 | |
Martin v. Löwis | 2eba0d6 | 2004-12-24 08:28:28 +0000 | [diff] [blame] | 1154 | For Windows (2000/NT/ME/98/95), assuming you have MS VC++ 7.1, the |
Tim Peters | 0344424 | 2000-09-19 00:38:35 +0000 | [diff] [blame] | 1155 | project files are in PCbuild, the workspace is pcbuild.dsw. See |
| 1156 | PCbuild\readme.txt for detailed instructions. |
Guido van Rossum | 8d7d4ed | 1996-07-30 21:41:07 +0000 | [diff] [blame] | 1157 | |
Martin v. Löwis | 2eba0d6 | 2004-12-24 08:28:28 +0000 | [diff] [blame] | 1158 | For other non-Unix Windows compilers, in particular MS VC++ 6.0 and |
Guido van Rossum | 31ae207 | 1999-04-12 14:47:30 +0000 | [diff] [blame] | 1159 | for OS/2, enter the directory "PC" and read the file "readme.txt". |
Guido van Rossum | 6d9cc80 | 1996-08-01 17:31:22 +0000 | [diff] [blame] | 1160 | |
| 1161 | For the Mac, a separate source distribution will be made available, |
| 1162 | for use with the CodeWarrior compiler. If you are interested in Mac |
| 1163 | development, join the PythonMac Special Interest Group |
| 1164 | (http://www.python.org/sigs/pythonmac-sig/, or send email to |
| 1165 | pythonmac-sig-request@python.org). |
| 1166 | |
| 1167 | Of course, there are also binary distributions available for these |
Guido van Rossum | 67345eb | 2001-01-18 14:51:12 +0000 | [diff] [blame] | 1168 | platforms -- see http://www.python.org/. |
Guido van Rossum | 6d9cc80 | 1996-08-01 17:31:22 +0000 | [diff] [blame] | 1169 | |
| 1170 | To port Python to a new non-UNIX system, you will have to fake the |
| 1171 | effect of running the configure script manually (for Mac and PC, this |
| 1172 | has already been done for you). A good start is to copy the file |
Michael W. Hudson | be6e0f4 | 2002-06-23 16:29:36 +0000 | [diff] [blame] | 1173 | pyconfig.h.in to pyconfig.h and edit the latter to reflect the actual |
Guido van Rossum | 433c8ad | 1994-08-01 12:07:07 +0000 | [diff] [blame] | 1174 | configuration of your system. Most symbols must simply be defined as |
| 1175 | 1 only if the corresponding feature is present and can be left alone |
Michael W. Hudson | be6e0f4 | 2002-06-23 16:29:36 +0000 | [diff] [blame] | 1176 | otherwise; however the *_t type symbols must be defined as some |
| 1177 | variant of int if they need to be defined at all. |
Guido van Rossum | 433c8ad | 1994-08-01 12:07:07 +0000 | [diff] [blame] | 1178 | |
Barry Warsaw | 0655745 | 2001-12-14 20:47:12 +0000 | [diff] [blame] | 1179 | For all platforms, it's important that the build arrange to define the |
| 1180 | preprocessor symbol NDEBUG on the compiler command line in a release |
| 1181 | build of Python (else assert() calls remain in the code, hurting |
| 1182 | release-build performance). The Unix, Windows and Mac builds already |
| 1183 | do this. |
Guido van Rossum | 433c8ad | 1994-08-01 12:07:07 +0000 | [diff] [blame] | 1184 | |
Guido van Rossum | 91cb9d2 | 1995-04-10 11:47:38 +0000 | [diff] [blame] | 1185 | |
| 1186 | Miscellaneous issues |
| 1187 | ==================== |
| 1188 | |
Guido van Rossum | 91cb9d2 | 1995-04-10 11:47:38 +0000 | [diff] [blame] | 1189 | Emacs mode |
| 1190 | ---------- |
| 1191 | |
Guido van Rossum | c0be2f5 | 1997-10-08 05:05:28 +0000 | [diff] [blame] | 1192 | There's an excellent Emacs editing mode for Python code; see the file |
| 1193 | Misc/python-mode.el. Originally written by the famous Tim Peters, it |
Guido van Rossum | 4f40c2d | 2000-09-05 16:46:22 +0000 | [diff] [blame] | 1194 | is now maintained by the equally famous Barry Warsaw (it's no |
Guido van Rossum | 67345eb | 2001-01-18 14:51:12 +0000 | [diff] [blame] | 1195 | coincidence that they now both work on the same team). The latest |
Guido van Rossum | c07d5fa | 2000-09-01 22:50:02 +0000 | [diff] [blame] | 1196 | version, along with various other contributed Python-related Emacs |
Guido van Rossum | 67345eb | 2001-01-18 14:51:12 +0000 | [diff] [blame] | 1197 | goodies, is online at http://www.python.org/emacs/python-mode. And |
Guido van Rossum | c07d5fa | 2000-09-01 22:50:02 +0000 | [diff] [blame] | 1198 | if you are planning to edit the Python C code, please pick up the |
Guido van Rossum | 67345eb | 2001-01-18 14:51:12 +0000 | [diff] [blame] | 1199 | latest version of CC Mode http://www.python.org/emacs/cc-mode; it |
Guido van Rossum | c07d5fa | 2000-09-01 22:50:02 +0000 | [diff] [blame] | 1200 | contains a "python" style used throughout most of the Python C source |
| 1201 | files. (Newer versions of Emacs or XEmacs may already come with the |
| 1202 | latest version of python-mode.) |
Guido van Rossum | 91cb9d2 | 1995-04-10 11:47:38 +0000 | [diff] [blame] | 1203 | |
| 1204 | |
Guido van Rossum | 59b0b60 | 2002-02-08 22:22:35 +0000 | [diff] [blame] | 1205 | Tkinter |
| 1206 | ------- |
Guido van Rossum | 91cb9d2 | 1995-04-10 11:47:38 +0000 | [diff] [blame] | 1207 | |
Guido van Rossum | 59b0b60 | 2002-02-08 22:22:35 +0000 | [diff] [blame] | 1208 | The setup.py script automatically configures this when it detects a |
| 1209 | usable Tcl/Tk installation. This requires Tcl/Tk version 8.0 or |
Fred Drake | 8179a9d | 2000-09-28 23:28:30 +0000 | [diff] [blame] | 1210 | higher. |
Guido van Rossum | 6d9cc80 | 1996-08-01 17:31:22 +0000 | [diff] [blame] | 1211 | |
Guido van Rossum | 973e4dc | 2000-07-01 00:34:39 +0000 | [diff] [blame] | 1212 | For more Tkinter information, see the Tkinter Resource page: |
| 1213 | http://www.python.org/topics/tkinter/ |
Guido van Rossum | c0be2f5 | 1997-10-08 05:05:28 +0000 | [diff] [blame] | 1214 | |
Michael W. Hudson | 71dcc3e | 2005-02-22 15:33:26 +0000 | [diff] [blame] | 1215 | There are demos in the Demo/tkinter directory. |
Guido van Rossum | 91cb9d2 | 1995-04-10 11:47:38 +0000 | [diff] [blame] | 1216 | |
| 1217 | Note that there's a Python module called "Tkinter" (capital T) which |
Guido van Rossum | c07d5fa | 2000-09-01 22:50:02 +0000 | [diff] [blame] | 1218 | lives in Lib/lib-tk/Tkinter.py, and a C module called "_tkinter" |
Guido van Rossum | 6d9cc80 | 1996-08-01 17:31:22 +0000 | [diff] [blame] | 1219 | (lower case t and leading underscore) which lives in |
Guido van Rossum | 4f40c2d | 2000-09-05 16:46:22 +0000 | [diff] [blame] | 1220 | Modules/_tkinter.c. Demos and normal Tk applications import only the |
Guido van Rossum | 59b0b60 | 2002-02-08 22:22:35 +0000 | [diff] [blame] | 1221 | Python Tkinter module -- only the latter imports the C _tkinter |
| 1222 | module. In order to find the C _tkinter module, it must be compiled |
| 1223 | and linked into the Python interpreter -- the setup.py script does |
| 1224 | this. In order to find the Python Tkinter module, sys.path must be |
| 1225 | set correctly -- normal installation takes care of this. |
Guido van Rossum | 84c8c7f | 1995-08-28 02:44:24 +0000 | [diff] [blame] | 1226 | |
Guido van Rossum | 91cb9d2 | 1995-04-10 11:47:38 +0000 | [diff] [blame] | 1227 | |
Guido van Rossum | 433c8ad | 1994-08-01 12:07:07 +0000 | [diff] [blame] | 1228 | Distribution structure |
| 1229 | ---------------------- |
| 1230 | |
Guido van Rossum | 4f40c2d | 2000-09-05 16:46:22 +0000 | [diff] [blame] | 1231 | Most subdirectories have their own README files. Most files have |
Guido van Rossum | 433c8ad | 1994-08-01 12:07:07 +0000 | [diff] [blame] | 1232 | comments. |
| 1233 | |
Anthony Baxter | 70e0796 | 2006-04-03 14:16:27 +0000 | [diff] [blame] | 1234 | BeOS/ Files specific to the BeOS port |
Guido van Rossum | 6d9cc80 | 1996-08-01 17:31:22 +0000 | [diff] [blame] | 1235 | Demo/ Demonstration scripts, modules and programs |
Anthony Baxter | 70e0796 | 2006-04-03 14:16:27 +0000 | [diff] [blame] | 1236 | Doc/ Documentation sources (LaTeX) |
Guido van Rossum | 6d9cc80 | 1996-08-01 17:31:22 +0000 | [diff] [blame] | 1237 | Grammar/ Input for the parser generator |
| 1238 | Include/ Public header files |
Anthony Baxter | 70e0796 | 2006-04-03 14:16:27 +0000 | [diff] [blame] | 1239 | LICENSE Licensing information |
Guido van Rossum | 6d9cc80 | 1996-08-01 17:31:22 +0000 | [diff] [blame] | 1240 | Lib/ Python library modules |
Michael W. Hudson | be6e0f4 | 2002-06-23 16:29:36 +0000 | [diff] [blame] | 1241 | Mac/ Macintosh specific resources |
Neil Schemenauer | 3f52da5 | 2001-01-24 17:49:59 +0000 | [diff] [blame] | 1242 | Makefile.pre.in Source from which config.status creates the Makefile.pre |
Guido van Rossum | c0be2f5 | 1997-10-08 05:05:28 +0000 | [diff] [blame] | 1243 | Misc/ Miscellaneous useful files |
Guido van Rossum | 6d9cc80 | 1996-08-01 17:31:22 +0000 | [diff] [blame] | 1244 | Modules/ Implementation of most built-in modules |
| 1245 | Objects/ Implementation of most built-in object types |
Guido van Rossum | c07d5fa | 2000-09-01 22:50:02 +0000 | [diff] [blame] | 1246 | PC/ Files specific to PC ports (DOS, Windows, OS/2) |
Anthony Baxter | 70e0796 | 2006-04-03 14:16:27 +0000 | [diff] [blame] | 1247 | PCbuild/ Build directory for Microsoft Visual C++ |
Guido van Rossum | 6d9cc80 | 1996-08-01 17:31:22 +0000 | [diff] [blame] | 1248 | Parser/ The parser and tokenizer and their input handling |
Guido van Rossum | 4f40c2d | 2000-09-05 16:46:22 +0000 | [diff] [blame] | 1249 | Python/ The byte-compiler and interpreter |
Guido van Rossum | 6d9cc80 | 1996-08-01 17:31:22 +0000 | [diff] [blame] | 1250 | README The file you're reading now |
| 1251 | Tools/ Some useful programs written in Python |
Michael W. Hudson | be6e0f4 | 2002-06-23 16:29:36 +0000 | [diff] [blame] | 1252 | pyconfig.h.in Source from which pyconfig.h is created (GNU autoheader output) |
Guido van Rossum | 6d9cc80 | 1996-08-01 17:31:22 +0000 | [diff] [blame] | 1253 | configure Configuration shell script (GNU autoconf output) |
Guido van Rossum | c07d5fa | 2000-09-01 22:50:02 +0000 | [diff] [blame] | 1254 | configure.in Configuration specification (input for GNU autoconf) |
Guido van Rossum | 6d9cc80 | 1996-08-01 17:31:22 +0000 | [diff] [blame] | 1255 | install-sh Shell script used to install files |
Anthony Baxter | 70e0796 | 2006-04-03 14:16:27 +0000 | [diff] [blame] | 1256 | setup.py Python script used to build extension modules |
Guido van Rossum | 91cb9d2 | 1995-04-10 11:47:38 +0000 | [diff] [blame] | 1257 | |
| 1258 | The following files will (may) be created in the toplevel directory by |
| 1259 | the configuration and build processes: |
| 1260 | |
Guido van Rossum | 6d9cc80 | 1996-08-01 17:31:22 +0000 | [diff] [blame] | 1261 | Makefile Build rules |
Neil Schemenauer | 3f52da5 | 2001-01-24 17:49:59 +0000 | [diff] [blame] | 1262 | Makefile.pre Build rules before running Modules/makesetup |
Michael W. Hudson | be6e0f4 | 2002-06-23 16:29:36 +0000 | [diff] [blame] | 1263 | buildno Keeps track of the build number |
Guido van Rossum | c07d5fa | 2000-09-01 22:50:02 +0000 | [diff] [blame] | 1264 | config.cache Cache of configuration variables |
Michael W. Hudson | be6e0f4 | 2002-06-23 16:29:36 +0000 | [diff] [blame] | 1265 | pyconfig.h Configuration header |
Guido van Rossum | c0be2f5 | 1997-10-08 05:05:28 +0000 | [diff] [blame] | 1266 | config.log Log from last configure run |
Guido van Rossum | 4f40c2d | 2000-09-05 16:46:22 +0000 | [diff] [blame] | 1267 | config.status Status from last run of the configure script |
Anthony Baxter | 70e0796 | 2006-04-03 14:16:27 +0000 | [diff] [blame] | 1268 | getbuildinfo.o Object file from Modules/getbuildinfo.c |
| 1269 | libpython<version>.a The library archive |
Guido van Rossum | 6d9cc80 | 1996-08-01 17:31:22 +0000 | [diff] [blame] | 1270 | python The executable interpreter |
| 1271 | tags, TAGS Tags files for vi and Emacs |
Guido van Rossum | 433c8ad | 1994-08-01 12:07:07 +0000 | [diff] [blame] | 1272 | |
| 1273 | |
Guido van Rossum | c07d5fa | 2000-09-01 22:50:02 +0000 | [diff] [blame] | 1274 | That's all, folks! |
| 1275 | ------------------ |
Guido van Rossum | 433c8ad | 1994-08-01 12:07:07 +0000 | [diff] [blame] | 1276 | |
| 1277 | |
Guido van Rossum | 67345eb | 2001-01-18 14:51:12 +0000 | [diff] [blame] | 1278 | --Guido van Rossum (home page: http://www.python.org/~guido/) |