blob: 8868023812fc60d3004cccb9954ea7e2ddf94898 [file] [log] [blame]
Willy Tarreaub1a34b62010-05-09 22:37:12 +02001 ----------------------
2 HAProxy how-to
3 ----------------------
Willy Tarreau2e858402014-05-28 17:50:53 +02004 version 1.5-dev26
Willy Tarreaub1a34b62010-05-09 22:37:12 +02005 willy tarreau
Willy Tarreau2e858402014-05-28 17:50:53 +02006 2014/05/28
willy tarreau78345332005-12-18 01:33:16 +01007
8
Willy Tarreaub1a34b62010-05-09 22:37:12 +020091) How to build it
10------------------
11
willy tarreau78345332005-12-18 01:33:16 +010012To build haproxy, you will need :
Willy Tarreaub1a34b62010-05-09 22:37:12 +020013 - GNU make. Neither Solaris nor OpenBSD's make work with the GNU Makefile.
Willy Tarreau3543cdb2014-05-10 09:12:46 +020014 If you get many syntax errors when running "make", you may want to retry
15 with "gmake" which is the name commonly used for GNU make on BSD systems.
Willy Tarreau32e65ef2013-04-02 08:14:29 +020016 - GCC between 2.91 and 4.7. Others may work, but not tested.
willy tarreau78345332005-12-18 01:33:16 +010017 - GNU ld
18
19Also, you might want to build with libpcre support, which will provide a very
Willy Tarreaub1a34b62010-05-09 22:37:12 +020020efficient regex implementation and will also fix some badness on Solaris' one.
willy tarreau78345332005-12-18 01:33:16 +010021
22To build haproxy, you have to choose your target OS amongst the following ones
23and assign it to the TARGET variable :
24
Willy Tarreau83b30c12008-05-25 10:32:50 +020025 - linux22 for Linux 2.2
26 - linux24 for Linux 2.4 and above (default)
27 - linux24e for Linux 2.4 with support for a working epoll (> 0.21)
Willy Tarreau83b30c12008-05-25 10:32:50 +020028 - linux26 for Linux 2.6 and above
Willy Tarreaue0c623d2012-06-04 00:42:09 +020029 - linux2628 for Linux 2.6.28 and above (enables splice and tproxy)
Willy Tarreau83b30c12008-05-25 10:32:50 +020030 - solaris for Solaris 8 or 10 (others untested)
Willy Tarreaub1a34b62010-05-09 22:37:12 +020031 - freebsd for FreeBSD 5 to 8.0 (others untested)
Willy Tarreau8624cab2013-04-02 08:17:43 +020032 - osx for Mac OS/X
Willy Tarreau3b8e9792012-11-22 00:43:09 +010033 - openbsd for OpenBSD 3.1 to 5.2 (others untested)
Willy Tarreau50abe302014-04-02 20:44:43 +020034 - aix51 for AIX 5.1
Willy Tarreau7dec9652012-06-06 16:15:03 +020035 - aix52 for AIX 5.2
Yitzhak Sapir32087312009-06-14 18:27:54 +020036 - cygwin for Cygwin
Willy Tarreau83b30c12008-05-25 10:32:50 +020037 - generic for any other OS.
38 - custom to manually adjust every setting
willy tarreau78345332005-12-18 01:33:16 +010039
40You may also choose your CPU to benefit from some optimizations. This is
41particularly important on UltraSparc machines. For this, you can assign
42one of the following choices to the CPU variable :
43
44 - i686 for intel PentiumPro, Pentium 2 and above, AMD Athlon
45 - i586 for intel Pentium, AMD K6, VIA C3.
46 - ultrasparc : Sun UltraSparc I/II/III/IV processor
Willy Tarreaua5899aa2010-11-28 07:41:00 +010047 - native : use the build machine's specific processor optimizations
willy tarreau78345332005-12-18 01:33:16 +010048 - generic : any other processor or no specific optimization. (default)
49
Willy Tarreau83b30c12008-05-25 10:32:50 +020050Alternatively, you may just set the CPU_CFLAGS value to the optimal GCC options
51for your platform.
52
Willy Tarreauef7341d2009-04-11 19:45:50 +020053You may want to build specific target binaries which do not match your native
54compiler's target. This is particularly true on 64-bit systems when you want
55to build a 32-bit binary. Use the ARCH variable for this purpose. Right now
Willy Tarreaua5899aa2010-11-28 07:41:00 +010056it only knows about a few x86 variants (i386,i486,i586,i686,x86_64), two
57generic ones (32,64) and sets -m32/-m64 as well as -march=<arch> accordingly.
Willy Tarreauef7341d2009-04-11 19:45:50 +020058
willy tarreau78345332005-12-18 01:33:16 +010059If your system supports PCRE (Perl Compatible Regular Expressions), then you
60really should build with libpcre which is between 2 and 10 times faster than
61other libc implementations. Regex are used for header processing (deletion,
62rewriting, allow, deny). The only inconvenient of libpcre is that it is not
63yet widely spread, so if you build for other systems, you might get into
64trouble if they don't have the dynamic library. In this situation, you should
65statically link libpcre into haproxy so that it will not be necessary to
Willy Tarreau83b30c12008-05-25 10:32:50 +020066install it on target systems. Available build options for PCRE are :
willy tarreau78345332005-12-18 01:33:16 +010067
Willy Tarreau83b30c12008-05-25 10:32:50 +020068 - USE_PCRE=1 to use libpcre, in whatever form is available on your system
willy tarreau78345332005-12-18 01:33:16 +010069 (shared or static)
70
Willy Tarreau83b30c12008-05-25 10:32:50 +020071 - USE_STATIC_PCRE=1 to use a static version of libpcre even if the dynamic
72 one is available. This will enhance portability.
73
Willy Tarreau663148c2012-12-12 00:38:22 +010074 - with no option, use your OS libc's standard regex implementation (default).
Willy Tarreau83b30c12008-05-25 10:32:50 +020075 Warning! group references on Solaris seem broken. Use static-pcre whenever
76 possible.
willy tarreau78345332005-12-18 01:33:16 +010077
Willy Tarreau64bc40b2011-03-23 20:00:53 +010078Recent systems can resolve IPv6 host names using getaddrinfo(). This primitive
79is not present in all libcs and does not work in all of them either. Support in
80glibc was broken before 2.3. Some embedded libs may not properly work either,
81thus, support is disabled by default, meaning that some host names which only
82resolve as IPv6 addresses will not resolve and configs might emit an error
83during parsing. If you know that your OS libc has reliable support for
84getaddrinfo(), you can add USE_GETADDRINFO=1 on the make command line to enable
85it. This is the recommended option for most Linux distro packagers since it's
86working fine on all recent mainstream distros. It is automatically enabled on
87Solaris 8 and above, as it's known to work.
88
Willy Tarreau3543cdb2014-05-10 09:12:46 +020089It is possible to add native support for SSL using the GNU makefile, by passing
90"USE_OPENSSL=1" on the make command line. The libssl and libcrypto will
91automatically be linked with haproxy. Some systems also require libz, so if the
92build fails due to missing symbols such as deflateInit(), then try again with
93"ADDLIB=-lz".
Willy Tarreaud4508812012-09-10 09:07:41 +020094
Lukas Tribus3fe9f1e2013-05-19 16:28:17 +020095To link OpenSSL statically against haproxy, build OpenSSL with the no-shared
96keyword and install it to a local directory, so your system is not affected :
97
98 $ export STATICLIBSSL=/tmp/staticlibssl
99 $ ./config --prefix=$STATICLIBSSL no-shared
100 $ make && make install_sw
101
Lukas Tribus130ddf72013-10-01 00:28:03 +0200102When building haproxy, pass that path via SSL_INC and SSL_LIB to make and
103include additional libs with ADDLIB if needed (in this case for example libdl):
Willy Tarreau3543cdb2014-05-10 09:12:46 +0200104
Lukas Tribus130ddf72013-10-01 00:28:03 +0200105 $ make TARGET=linux26 USE_OPENSSL=1 SSL_INC=$STATICLIBSSL/include SSL_LIB=$STATICLIBSSL/lib ADDLIB=-ldl
Lukas Tribus3fe9f1e2013-05-19 16:28:17 +0200106
William Lallemand82fe75c2012-10-23 10:25:10 +0200107It is also possible to include native support for ZLIB to benefit from HTTP
108compression. For this, pass "USE_ZLIB=1" on the "make" command line and ensure
109that zlib is present on the system.
110
willy tarreau78345332005-12-18 01:33:16 +0100111By default, the DEBUG variable is set to '-g' to enable debug symbols. It is
112not wise to disable it on uncommon systems, because it's often the only way to
113get a complete core when you need one. Otherwise, you can set DEBUG to '-s' to
114strip the binary.
115
116For example, I use this to build for Solaris 8 :
117
Willy Tarreau83b30c12008-05-25 10:32:50 +0200118 $ make TARGET=solaris CPU=ultrasparc USE_STATIC_PCRE=1
willy tarreau78345332005-12-18 01:33:16 +0100119
Willy Tarreau83b30c12008-05-25 10:32:50 +0200120And I build it this way on OpenBSD or FreeBSD :
willy tarreaud38e72d2006-03-19 20:56:52 +0100121
Willy Tarreau3543cdb2014-05-10 09:12:46 +0200122 $ gmake TARGET=freebsd USE_PCRE=1 USE_OPENSSL=1 USE_ZLIB=1
willy tarreaud38e72d2006-03-19 20:56:52 +0100123
Willy Tarreau663148c2012-12-12 00:38:22 +0100124And on a classic Linux with SSL and ZLIB support (eg: Red Hat 5.x) :
125
126 $ make TARGET=linux26 CPU=native USE_PCRE=1 USE_OPENSSL=1 USE_ZLIB=1
127
128And on a recent Linux >= 2.6.28 with SSL and ZLIB support :
Willy Tarreaud4508812012-09-10 09:07:41 +0200129
William Lallemand82fe75c2012-10-23 10:25:10 +0200130 $ make TARGET=linux2628 CPU=native USE_PCRE=1 USE_OPENSSL=1 USE_ZLIB=1
Willy Tarreaud4508812012-09-10 09:07:41 +0200131
William Lallemand82fe75c2012-10-23 10:25:10 +0200132In order to build a 32-bit binary on an x86_64 Linux system with SSL support
133without support for compression but when OpenSSL requires ZLIB anyway :
Willy Tarreauef7341d2009-04-11 19:45:50 +0200134
Willy Tarreaud4508812012-09-10 09:07:41 +0200135 $ make TARGET=linux26 ARCH=i386 USE_OPENSSL=1 ADDLIB=-lz
Willy Tarreauef7341d2009-04-11 19:45:50 +0200136
Willy Tarreaub1efede2014-05-09 00:44:48 +0200137The SSL stack supports session cache synchronization between all running
138processes. This involves some atomic operations and synchronization operations
139which come in multiple flavors depending on the system and architecture :
140
141 Atomic operations :
142 - internal assembler versions for x86/x86_64 architectures
143
144 - gcc builtins for other architectures. Some architectures might not
145 be fully supported or might require a more recent version of gcc.
146 If your architecture is not supported, you willy have to either use
147 pthread if supported, or to disable the shared cache.
148
149 - pthread (posix threads). Pthreads are very common but inter-process
150 support is not that common, and some older operating systems did not
151 report an error when enabling multi-process mode, so they used to
152 silently fail, possibly causing crashes. Linux's implementation is
153 fine. OpenBSD doesn't support them and doesn't build. FreeBSD 9 builds
154 and reports an error at runtime, while certain older versions might
155 silently fail. Pthreads are enabled using USE_PTHREAD_PSHARED=1.
156
157 Synchronization operations :
158 - internal spinlock : this mode is OS-independant, light but will not
159 scale well to many processes. However, accesses to the session cache
160 are rare enough that this mode could certainly always be used. This
161 is the default mode.
162
163 - Futexes, which are Linux-specific highly scalable light weight mutexes
164 implemented in user-space with some limited assistance from the kernel.
165 This is the default on Linux 2.6 and above and is enabled by passing
166 USE_FUTEX=1
167
168 - pthread (posix threads). See above.
169
170If none of these mechanisms is supported by your platform, you may need to
171build with USE_PRIVATE_CACHE=1 to totally disable SSL cache sharing. Then
172it is better not to run SSL on multiple processes.
173
willy tarreau78345332005-12-18 01:33:16 +0100174If you need to pass other defines, includes, libraries, etc... then please
175check the Makefile to see which ones will be available in your case, and
Willy Tarreau3543cdb2014-05-10 09:12:46 +0200176use the USE_* variables in the Makefile.
willy tarreau78345332005-12-18 01:33:16 +0100177
Willy Tarreau97ec9692010-01-28 20:52:05 +0100178AIX 5.3 is known to work with the generic target. However, for the binary to
179also run on 5.2 or earlier, you need to build with DEFINE="-D_MSGQSUPPORT",
180otherwise __fd_select() will be used while not being present in the libc.
Willy Tarreau7dec9652012-06-06 16:15:03 +0200181If you get build errors because of strange symbols or section mismatches,
182simply remove -g from DEBUG_CFLAGS.
Willy Tarreau97ec9692010-01-28 20:52:05 +0100183
Willy Tarreau32e65ef2013-04-02 08:14:29 +0200184You can easily define your own target with the GNU Makefile. Unknown targets
185are processed with no default option except USE_POLL=default. So you can very
186well use that property to define your own set of options. USE_POLL can even be
187disabled by setting USE_POLL="". For example :
188
189 $ gmake TARGET=tiny USE_POLL="" TARGET_CFLAGS=-fomit-frame-pointer
190
Willy Tarreaub1a34b62010-05-09 22:37:12 +0200191
1922) How to install it
193--------------------
194
195To install haproxy, you can either copy the single resulting binary to the
196place you want, or run :
197
198 $ sudo make install
199
200If you're packaging it for another system, you can specify its root directory
201in the usual DESTDIR variable.
202
203
2043) How to set it up
205-------------------
206
207There is some documentation in the doc/ directory :
208
209 - architecture.txt : this is the architecture manual. It is quite old and
210 does not tell about the nice new features, but it's still a good starting
211 point when you know what you want but don't know how to do it.
212
213 - configuration.txt : this is the configuration manual. It recalls a few
214 essential HTTP basic concepts, and details all the configuration file
215 syntax (keywords, units). It also describes the log and stats format. It
216 is normally always up to date. If you see that something is missing from
Willy Tarreau74774c02014-04-23 00:57:08 +0200217 it, please report it as this is a bug. Please note that this file is
218 huge and that it's generally more convenient to review Cyril Bonté's
219 HTML translation online here :
220
221 http://cbonte.github.io/haproxy-dconv/configuration-1.5.html
Willy Tarreaub1a34b62010-05-09 22:37:12 +0200222
223 - haproxy-en.txt / haproxy-fr.txt : these are the old outdated docs. You
224 should never need them. If you do, then please report what you didn't
225 find in the other ones.
226
227 - gpl.txt / lgpl.txt : the copy of the licenses covering the software. See
228 the 'LICENSE' file at the top for more information.
229
230 - the rest is mainly for developers.
231
232There are also a number of nice configuration examples in the "examples"
233directory as well as on several sites and articles on the net which are linked
234to from the haproxy web site.
235
236
2374) How to report a bug
238----------------------
239
240It is possible that from time to time you'll find a bug. A bug is a case where
241what you see is not what is documented. Otherwise it can be a misdesign. If you
242find that something is stupidly design, please discuss it on the list (see the
243"how to contribute" section below). If you feel like you're proceeding right
244and haproxy doesn't obey, then first ask yourself if it is possible that nobody
245before you has even encountered this issue. If it's unlikely, the you probably
246have an issue in your setup. Just in case of doubt, please consult the mailing
247list archives :
248
Willy Tarreaub1a34b62010-05-09 22:37:12 +0200249 http://marc.info/?l=haproxy
250
251Otherwise, please try to gather the maximum amount of information to help
252reproduce the issue and send that to the mailing list :
253
254 haproxy@formilux.org
255
256Please include your configuration and logs. You can mask your IP addresses and
257passwords, we don't need them. But it's essential that you post your config if
258you want people to guess what is happening.
259
260Also, keep in mind that haproxy is designed to NEVER CRASH. If you see it die
261without any reason, then it definitely is a critical bug that must be reported
262and urgently fixed. It has happened a couple of times in the past, essentially
263on development versions running on new architectures. If you think your setup
264is fairly common, then it is possible that the issue is totally unrelated.
265Anyway, if that happens, feel free to contact me directly, as I will give you
266instructions on how to collect a usable core file, and will probably ask for
267other captures that you'll not want to share with the list.
268
269
2705) How to contribute
271--------------------
272
273It is possible that you'll want to add a specific feature to satisfy your needs
274or one of your customers'. Contributions are welcome, however I'm often very
275picky about changes. I will generally reject patches that change massive parts
276of the code, or that touch the core parts without any good reason if those
277changes have not been discussed first.
278
279The proper place to discuss your changes is the HAProxy Mailing List. There are
280enough skilled readers to catch hazardous mistakes and to suggest improvements.
Willy Tarreau9a639a12011-09-10 22:48:36 +0200281I trust a number of them enough to merge a patch if they say it's OK, so using
282the list is the fastest way to get your code reviewed and merged. You can
283subscribe to it by sending an empty e-mail at the following address :
Willy Tarreaub1a34b62010-05-09 22:37:12 +0200284
285 haproxy+subscribe@formilux.org
286
Willy Tarreau9a639a12011-09-10 22:48:36 +0200287If you have an idea about something to implement, *please* discuss it on the
288list first. It has already happened several times that two persons did the same
289thing simultaneously. This is a waste of time for both of them. It's also very
290common to see some changes rejected because they're done in a way that will
291conflict with future evolutions, or that does not leave a good feeling. It's
292always unpleasant for the person who did the work, and it is unpleasant for me
293too because I value people's time and efforts. That would not happen if these
294were discussed first. There is no problem posting work in progress to the list,
295it happens quite often in fact. Also, don't waste your time with the doc when
296submitting patches for review, only add the doc with the patch you consider
297ready to merge.
298
Willy Tarreau2ddccb72013-05-01 10:07:21 +0200299Another important point concerns code portability. Haproxy requires gcc as the
300C compiler, and may or may not work with other compilers. However it's known
301to build using gcc 2.95 or any later version. As such, it is important to keep
302in mind that certain facilities offered by recent versions must not be used in
303the code :
304
305 - declarations mixed in the code (requires gcc >= 3.x)
306 - GCC builtins without checking for their availability based on version and
307 architecture ;
308 - assembly code without any alternate portable form for other platforms
309 - use of stdbool.h, "bool", "false", "true" : simply use "int", "0", "1"
310 - in general, anything which requires C99 (such as declaring variables in
311 "for" statements)
312
313Since most of these restrictions are just a matter of coding style, it is
314normally not a problem to comply.
315
Willy Tarreaub1a34b62010-05-09 22:37:12 +0200316If your work is very confidential and you can't publicly discuss it, you can
317also mail me directly about it, but your mail may be waiting several days in
318the queue before you get a response.
319
320If you'd like a feature to be added but you think you don't have the skills to
321implement it yourself, you should follow these steps :
322
323 1. discuss the feature on the mailing list. It is possible that someone
324 else has already implemented it, or that someone will tell you how to
325 proceed without it, or even why not to do it. It is also possible that
326 in fact it's quite easy to implement and people will guide you through
327 the process. That way you'll finally have YOUR patch merged, providing
328 the feature YOU need.
329
330 2. if you really can't code it yourself after discussing it, then you may
331 consider contacting someone to do the job for you. Some people on the
332 list might be OK with trying to do it. Otherwise, you can check the list
333 of contributors at the URL below, some of the regular contributors may
334 be able to do the work, probably not for free but their time is as much
335 valuable as yours after all, you can't eat the cake and have it too.
336
337The list of past and regular contributors is available below. It lists not only
338significant code contributions (features, fixes), but also time or money
339donations :
340
Willy Tarreau6346f0a2014-05-10 11:04:39 +0200341 http://www.haproxy.org/contrib.html
Willy Tarreaub1a34b62010-05-09 22:37:12 +0200342
343Note to contributors: it's very handy when patches comes with a properly
Willy Tarreau9a639a12011-09-10 22:48:36 +0200344formated subject. There are 3 criteria of particular importance in any patch :
Willy Tarreaub1a34b62010-05-09 22:37:12 +0200345
Willy Tarreau9a639a12011-09-10 22:48:36 +0200346 - its nature (is it a fix for a bug, a new feature, an optimization, ...)
347 - its importance, which generally reflects the risk of merging/not merging it
348 - what area it applies to (eg: http, stats, startup, config, doc, ...)
Willy Tarreaub1a34b62010-05-09 22:37:12 +0200349
Willy Tarreau9a639a12011-09-10 22:48:36 +0200350It's important to make these 3 criteria easy to spot in the patch's subject,
351because it's the first (and sometimes the only) thing which is read when
352reviewing patches to find which ones need to be backported to older versions.
Willy Tarreaub1a34b62010-05-09 22:37:12 +0200353
Willy Tarreau9a639a12011-09-10 22:48:36 +0200354Specifically, bugs must be clearly easy to spot so that they're never missed.
355Any patch fixing a bug must have the "BUG" tag in its subject. Most common
356patch types include :
357
358 - BUG fix for a bug. The severity of the bug should also be indicated
359 when known. Similarly, if a backport is needed to older versions,
360 it should be indicated on the last line of the commit message. If
361 the bug has been identified as a regression brought by a specific
362 patch or version, this indication will be appreciated too. New
363 maintenance releases are generally emitted when a few of these
364 patches are merged.
365
366 - CLEANUP code cleanup, silence of warnings, etc... theorically no impact.
Willy Tarreaub1a34b62010-05-09 22:37:12 +0200367 These patches will rarely be seen in stable branches, though they
Willy Tarreau9a639a12011-09-10 22:48:36 +0200368 may appear when they remove some annoyance or when they make
369 backporting easier. By nature, a cleanup is always minor.
370
371 - REORG code reorganization. Some blocks may be moved to other places,
372 some important checks might be swapped, etc... These changes
373 always present a risk of regression. For this reason, they should
374 never be mixed with any bug fix nor functional change. Code is
375 only moved as-is. Indicating the risk of breakage is highly
376 recommended.
377
378 - BUILD updates or fixes for build issues. Changes to makefiles also fall
379 into this category. The risk of breakage should be indicated if
380 known. It is also appreciated to indicate what platforms and/or
381 configurations were tested after the change.
382
383 - OPTIM some code was optimised. Sometimes if the regression risk is very
384 low and the gains significant, such patches may be merged in the
385 stable branch. Depending on the amount of code changed or replaced
386 and the level of trust the author has in the change, the risk of
387 regression should be indicated.
388
389 - RELEASE release of a new version (development or stable).
Willy Tarreaub1a34b62010-05-09 22:37:12 +0200390
Willy Tarreau9a639a12011-09-10 22:48:36 +0200391 - LICENSE licensing updates (may impact distro packagers).
Willy Tarreaub1a34b62010-05-09 22:37:12 +0200392
Willy Tarreaub1a34b62010-05-09 22:37:12 +0200393
Willy Tarreau9a639a12011-09-10 22:48:36 +0200394When the patch cannot be categorized, it's best not to put any tag. This is
395commonly the case for new features, which development versions are mostly made
396of.
397
398Additionally, the importance of the patch should be indicated when known. A
399single upper-case word is preferred, among :
400
401 - MINOR minor change, very low risk of impact. It is often the case for
402 code additions that don't touch live code. For a bug, it generally
403 indicates an annoyance, nothing more.
404
405 - MEDIUM medium risk, may cause unexpected regressions of low importance or
406 which may quickly be discovered. For a bug, it generally indicates
407 something odd which requires changing the configuration in an
408 undesired way to work around the issue.
409
410 - MAJOR major risk of hidden regression. This happens when I rearrange
Willy Tarreaub1a34b62010-05-09 22:37:12 +0200411 large parts of code, when I play with timeouts, with variable
412 initializations, etc... We should only exceptionally find such
Willy Tarreau9a639a12011-09-10 22:48:36 +0200413 patches in stable branches. For a bug, it indicates severe
414 reliability issues for which workarounds are identified with or
415 without performance impacts.
Willy Tarreaub1a34b62010-05-09 22:37:12 +0200416
Willy Tarreau9a639a12011-09-10 22:48:36 +0200417 - CRITICAL medium-term reliability or security is at risk and workarounds,
418 if they exist, might not always be acceptable. An upgrade is
419 absolutely required. A maintenance release may be emitted even if
420 only one of these bugs are fixed. Note that this tag is only used
421 with bugs. Such patches must indicate what is the first version
422 affected, and if known, the commit ID which introduced the issue.
423
424If this criterion doesn't apply, it's best not to put it. For instance, most
425doc updates and most examples or test files are just added or updated without
426any need to qualify a level of importance.
427
428The area the patch applies to is quite important, because some areas are known
429to be similar in older versions, suggesting a backport might be desirable, and
430conversely, some areas are known to be specific to one version. When the tag is
431used alone, uppercase is preferred for readability, otherwise lowercase is fine
432too. The following tags are suggested but not limitative :
433
434 - doc documentation updates or fixes. No code is affected, no need to
435 upgrade. These patches can also be sent right after a new feature,
436 to document it.
Willy Tarreaub1a34b62010-05-09 22:37:12 +0200437
Willy Tarreau9a639a12011-09-10 22:48:36 +0200438 - examples example files. Be careful, sometimes these files are packaged.
Willy Tarreaub1a34b62010-05-09 22:37:12 +0200439
Willy Tarreau9a639a12011-09-10 22:48:36 +0200440 - tests regression test files. No code is affected, no need to upgrade.
Willy Tarreaub1a34b62010-05-09 22:37:12 +0200441
Willy Tarreau9a639a12011-09-10 22:48:36 +0200442 - init initialization code, arguments parsing, etc...
Willy Tarreaub1a34b62010-05-09 22:37:12 +0200443
Willy Tarreau9a639a12011-09-10 22:48:36 +0200444 - config configuration parser, mostly used when adding new config keywords
Willy Tarreaub1a34b62010-05-09 22:37:12 +0200445
Willy Tarreau9a639a12011-09-10 22:48:36 +0200446 - http the HTTP engine
Willy Tarreaub1a34b62010-05-09 22:37:12 +0200447
Willy Tarreau9a639a12011-09-10 22:48:36 +0200448 - stats the stats reporting engine as well as the stats socket CLI
Willy Tarreaub1a34b62010-05-09 22:37:12 +0200449
Willy Tarreau9a639a12011-09-10 22:48:36 +0200450 - checks the health checks engine (eg: when adding new checks)
Willy Tarreaub1a34b62010-05-09 22:37:12 +0200451
Willy Tarreau9a639a12011-09-10 22:48:36 +0200452 - acl the ACL processing core or some ACLs from other areas
453
454 - peers the peer synchronization engine
455
456 - listeners everything related to incoming connection settings
457
458 - frontend everything related to incoming connection processing
459
460 - backend everything related to LB algorithms and server farm
461
462 - session session processing and flags (very sensible, be careful)
463
464 - server server connection management, queueing
465
466 - proxy proxy maintenance (start/stop)
467
468 - log log management
469
470 - poll any of the pollers
471
472 - halog the halog sub-component in the contrib directory
473
474 - contrib any addition to the contrib directory
475
476Other names may be invented when more precise indications are meaningful, for
477instance : "cookie" which indicates cookie processing in the HTTP core. Last,
478indicating the name of the affected file is also a good way to quickly spot
479changes. Many commits were already tagged with "stream_sock" or "cfgparse" for
480instance.
481
482It is desired that AT LEAST one of the 3 criteria tags is reported in the patch
483subject. Ideally, we would have the 3 most often. The two first criteria should
484be present before a first colon (':'). If both are present, then they should be
485delimited with a slash ('/'). The 3rd criterion (area) should appear next, also
486followed by a colon. Thus, all of the following messages are valid :
Willy Tarreaub1a34b62010-05-09 22:37:12 +0200487
488Examples of messages :
Willy Tarreau9a639a12011-09-10 22:48:36 +0200489 - DOC: document options forwardfor to logasap
490 - DOC/MAJOR: reorganize the whole document and change indenting
491 - BUG: stats: connection reset counters must be plain ascii, not HTML
492 - BUG/MINOR: stats: connection reset counters must be plain ascii, not HTML
493 - MEDIUM: checks: support multi-packet health check responses
494 - RELEASE: Released version 1.4.2
495 - BUILD: stats: stdint is not present on solaris
496 - OPTIM/MINOR: halog: make fgets parse more bytes by blocks
497 - REORG/MEDIUM: move syscall redefinition to specific places
498
499Please do not use square brackets anymore around the tags, because they give me
500more work when merging patches. By default I'm asking Git to keep them but this
501causes trouble when patches are prefixed with the [PATCH] tag because in order
502not to store it, I have to hand-edit the patches. So as of now, I will ask Git
503to remove whatever is located between square brackets, which implies that any
504subject formatted the old way will have its tag stripped out.
505
506In fact, one of the only square bracket tags that still makes sense is '[RFC]'
507at the beginning of the subject, when you're asking for someone to review your
508change before getting it merged. If the patch is OK to be merged, then I can
509merge it as-is and the '[RFC]' tag will automatically be removed. If you don't
510want it to be merged at all, you can simply state it in the message, or use an
511alternate '[WIP]' tag ("work in progress").
512
513The tags are not rigid, follow your intuition first, anyway I reserve the right
514to change them when merging the patch. It may happen that a same patch has a
515different tag in two distinct branches. The reason is that a bug in one branch
516may just be a cleanup in the other one because the code cannot be triggered.
517
Willy Tarreaub1a34b62010-05-09 22:37:12 +0200518
519For a more efficient interaction between the mainline code and your code, I can
520only strongly encourage you to try the Git version control system :
521
522 http://git-scm.com/
523
524It's very fast, lightweight and lets you undo/redo your work as often as you
525want, without making your mistakes visible to the rest of the world. It will
526definitely help you contribute quality code and take other people's feedback
527in consideration. In order to clone the HAProxy Git repository :
528
Willy Tarreau6346f0a2014-05-10 11:04:39 +0200529 $ git clone http://git.haproxy.org/git/haproxy-1.4.git (stable 1.4)
530 $ git clone http://git.haproxy.org/git/haproxy.git/ (development)
Willy Tarreau663148c2012-12-12 00:38:22 +0100531
Willy Tarreaub1a34b62010-05-09 22:37:12 +0200532If you decide to use Git for your developments, then your commit messages will
533have the subject line in the format described above, then the whole description
534of your work (mainly why you did it) will be in the body. You can directly send
535your commits to the mailing list, the format is convenient to read and process.
536
willy tarreau78345332005-12-18 01:33:16 +0100537-- end