summaryrefslogtreecommitdiff
path: root/include_server
Commit message (Collapse)AuthorAgeFilesLines
* include_server python(3) readypython3clutton2016-04-1425-381/+408
|
* Apply patch from Rafael Ávila de Espíndola <rafael.espindola@gmail.com>:fergus.henderson2014-02-273-11/+19
| | | | | | | | | | | | | 1. Add support for -isysroot and -arch. 2. Fix a really nasty corner case: - CompressFiles::Compress uses os.makedirs to create a directory (foo) for one compilation unit - Another compilation unit does the equivalent of '#include "foo/../bar.h"' - The compilation unit uses no other file in foo - We conclude that we don't have to create a dummy foo/forcing_technique_271828
* include_server: Add missing "-x" to getoptmandyke2013-05-081-1/+1
| | | Fixes issue 113
* Fix issue 108 <http://code.google.com/p/distcc/issues/detail?id=108>:Kamal Mostafa2012-04-203-2/+15
| | | | | | | | | | | | | | | include_server --no_force_dirs inhibits forcing_technique The "forcing_technique" trick which handles the (rare) case of files included as e.g. <foo/../file.h> causes a performance degradation: The forcing files accumulate over the life of the include_server, making distcc send more and more (useless) files as the build progresses. While the forcing files are zero-length, the transmission filenames themselves introduce a significant overhead for builds with many files and directories. This patch implements the include_server arg --no_force_dirs, which if set switches off the "forcing_technique" scheme.
* Generalize dcc_{r,x}_argv routinesKamal Mostafa2012-04-201-2/+2
| | | | | | Generalize the dcc_{r,x}_argv routines so they can be used to transfer arbitrary array-of-strings tokens (useful for protocol hacking experiments; no functional change).
* Fix issue 104 <http://code.google.com/p/distcc/issues/detail?id=104>:fergus.henderson2012-04-091-4/+4
| | | | | handle header file names containing commas.
* Fix for distcc issue 99 ↵Kamal Mostafa2012-04-041-1/+2
| | | | <http://code.google.com/p/distcc/issues/detail?id=99>: include_server crash: local variable resolved referenced before assignment.
* Apply --sysroot patch from Simon Baldwin <simonb@google.com>:fergus.henderson2010-08-193-31/+69
| | | | | | | | | | | | | | | | | | | | | 2010-08-18 Simon Baldwin <simonb@google.com> * include_server/compiler_defaults.py (_SystemSearchdirsGCC): Add a 'sysroot' parameter. Add --sysroot, if needed, to the gcc invocation that identifies search directories and default paths. Add debugging output. (SetSystemDirsDefaults): Add 'sysroot' parameter. Add another dictionary level to system_dirs_default for sysroot values. Add debugging output. * include_server/parse_command.py (ParseCommandArgs): Identify and handle arguments of the form '--opt=value'. Pass any --sysroot value to compiler_defaults.SetSystemDirsDefaults, use it in compiler_defaults.system_dirs_default lookups. (CPP_OPTIONS_APPEARING_AS_ASSIGNMENTS): New. * include_server/parse_command_test.py (Mock_SetSystemDirsDefaults): Add sysroot parameter, verify it on function calls. (ParseCommandUnitTest): Add --sysroot to mock gcc invocations.
* Fix issue 35 <http://code.google.com/p/distcc/issues/detail?id=35>.fergus.henderson2009-01-302-6/+5
| | | | | | | | | | | | | | | | | This was a bug where the include server was crashing in certain cases (when the argument to a macro ended in a backslash) due to an escaping problem in the use of Python's re.sub() function. The fix was to replace all occurrences of backslash in the replacement string with double-backslash, which re.sub() will then translate back to a single backslash. (I also changed the code to not bother using re.compile() since we only use the regexp once.) Also, fix a bug where we were not allowing backslashes in filenames. I added regression tests for both of these bugs (and I verified that they were true regression tests). Reviewed by Craig Silverstein.
* Fix a bug where the SystemIncludeDirectories_Case testfergus.henderson2008-12-221-12/+15
| | | | | | | | | | | | | | | | | | | was failing on systems where /tmp was a symlink. This failure indicated a real bug: on such systems, distcc was not correctly handling -I<dir> where <dir> is a subdirectory of a system include directory, e.g. -I/usr/include/foo. The code was calling _RealPrefix(client_root + system_dir) but was implicitly assuming that the answer would always start with client_root. That is, it was implicitly assuming that client_root did not contain any symlinks. I changed the code to use _RealPrefixWithinClientRoot(client_root, system_dir) so that it would find the appropriate prefix of system_dir rather than looking at the prefixes of client_root.
* When computing the compiler's default search path by parsing the outputfergus.henderson2008-12-221-2/+20
| | | | | from "gcc -v", ignore Apple-modified MacOS gcc's "framework" directories.
* Avoid a spurious message 'setup.py: PYTHON_CFLAGS must be defined.'fergus.henderson2008-12-031-6/+1
| | | | | | The top-level Makefile sets CFLAGS="$(CFLAGS) $(PYTHON_CFLAGS)" before invoking setup.py, so there's no need for it to set PYTHON_CFLAGS.
* 1. Fix some compilation errors arising from the use offergus.henderson2008-12-021-18/+6
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | -Wwrite-strings when compiling the python extension module, due to lack of const correctness in the Python 2.2 "Python.h" header file. This problem was introduced by the fix for distcc issue 26 <http://code.google.com/p/distcc/issues/detail?id=26>. The solution was to explicitly disable these warnings with -Wno-write-strings. 2. Centralize the use of gcc-specific compilation options in conditional code in configure.ac that is only executed if we're using gcc. Previously include_server/setup.py was hard-coding gcc-specific options, regardless of whether we're using gcc. 3. Don't use -Wuninitialized if CFLAGS doesn't contain "-O*", because -Wuninitialized only works if optimization is enabled. This avoids a gcc warning (and hence an error with -Werror) about -Wuninitialized not having any effect when distcc is configured with "CFLAGS=-g ./configure". 4. Add a new configure option "--disable-Werror". For the 3.0 release of distcc, some people porting distcc resorted to patching distcc to remove the -Werror option. -Werror is useful, so I want to keep it enabled by default, but I'd prefer that people are able to port distcc easily, hence the configure option. This addresses distcc issue 20 <http://code.google.com/p/distcc/issues/detail?id=20>. Reviewers: Craig Silverstein
* Applied matsuu's patch for issue 27 "failed to make distcc"fergus.henderson2008-10-271-3/+4
| | | | | | | <http://code.google.com/p/distcc/issues/detail?id=27>, which was a build problem on "Gentoo, x86-64, gcc-4.1.2, glibc-2.6.1, python-2.5.2".
* Delete an unused test data file; it became dead after myfergus.henderson2008-09-271-1/+0
| | | | | previous change which deleted the test that used it.
* Delete an unused test case: the test was hard-coded to always pass anyway,fergus.henderson2008-09-271-20/+0
| | | | | so it wasn't testing anything useful.
* Provides a simple shell API to the include server:fergus.henderson2008-09-111-2/+4
| | | | | | | add a new "--scan-includes" option to distcc which shows which headers distcc would send to the remote machine in pump mode, without actually compiling anything.
* Fix a TODO in compiler_defaults.py that was responsible for a bugfergus.henderson2008-08-021-2/+13
| | | | | | | | | | | that broke pump mode when one of the default system include directories was a subdirectory of another, as turns out to be the case for GNU C++. The CPlusPlus_SystemIncludeDirectories_Case, which was failing before this change, now passes. (TODO: re-enable it.) Reviewers: Craig Silverstein
* Increase MAX_COMPONENTS_IN_SERVER_ROOT from 10 to 20.fergus.henderson2008-08-011-1/+4
| | | | | | | It turns out that our own test infrastructure (test/testdistcc.py) sets TMPDIR before invoking distccd, so this needs to be reasonably high, otherwise 'make distcheck' will fail.
* This is based on klarlund's original version of this patch (klarlund/d3b3):fergus.henderson2008-07-298-54/+262
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | gvn --project https://distcc.googlecode.com/svn/trunk review klarlund/d3b3 Fix Issue 7: Compiling with -I/usr/include/FOO or ... in pump mode Problem: Default system include directories are the directories on the default include search path, such as /usr/include, that is determined by the compiler. Distcc will not send default system include directories to the distcc compilation servers. Nevertheless, distcc on the server blindly rewrites -I options such as -I/usr/include/foo to -I/tmp/distccNNN/usr/include/foo. This doesn't work, since the files in /usr/include/foo are not sent to the distccd server. Solution: The present solution keeps the rewriting on the server, because we would like to not manage starting the compiler, parsing its output, and storing the default system paths on the server Instead, we use the existing mechanism for defining relative symbolic links under the server root. We escape from the root by using a sufficient number of '../'s. All this is tremendously complicated by: -- The possibility that other links encountered may point into the system default dirs in which case the usual mirroring of the reachable places should not take place, because the routing of such links will go through the link created for the system directory. -- The fact that determination of default-system-dirness is lazy: a deeply-seated link in a default system dir may become obsolete if it is later determined that a higher directory than the parent directory of the link is in fact also a default system dir. In that case, a new symlink, sitting in a place higher than the previous one will need to be created. Tests: make check benchmarks: samba (still 3X faster than with distcc), linux 2.6 kernel, and a couple of others TODO: better testing TODO: > In compiler_defaults.py, _MakeLinkFromMirrorToRealLocation: > Maybe comment each of the 'if' cases with an example of how this case > might be triggered. eg the real_prefix == rooted_system_dir is > triggered when we see /usr/include/foo, and the later see > /usr/include.
* Revert r393.klarlund2008-06-306-161/+70
| | | | | | | | | This change is mostly reverted -- some variable renamings are not changed back. A subsequent CL proposes a less arbitrary way to solve the problem of -I<some_dir_in_default_place_where_compiler_looks>, which involves sending no system default header files at all. Tests: make pump-check and make include-server-check
* Fix bug where the include server would crash if the PATH environmentfergus.henderson2008-06-181-1/+5
| | | | | variable wasn't set.
* The missing include_server(1) man page and additional material about the basicklarlund2008-06-091-13/+14
| | | | | | | | | | | | | | | | assumptions of distcc-pump added to man/distcc.1 as well. The BUGS section of distcc.1 has been updated. A spelling error correction is made to man/distcc.1. The pump.in script is changed so that it does not introduce other certain defaults than those mentioned in the manual (which are those of the of include server). Finally, I changed the order of the options in the include server source code to be alphabetical in accordance with include_server(1) man page. Reviewers: fergus, csilvers.
* Fix option real_path_warning_re not working.klarlund2008-06-023-42/+37
| | | | | | | | | | | | | | | | | | | I renamed the option to path_observation_re. Now by prefixing INCLUDE_SERVER_ARGS='-d1 --path_observation_re="/usr/.*"' to say make include-server-maintainer-check, one gets a message each time the include server finds a path whose realpath name matches the regular expression. That results in messages like: WARNING include server: For translation unit 'src/testtmp.c', lookup of file 'bits/stdio_lim.h' resolved to '/usr/include/bits/stdio_lim.h' whose realpath is '/usr/include/bits/stdio_lim.h'. To make the interpretation of quoted arguments within INCLUDE_SERVER_ARGS correct, I added 'eval' to the command that cranks up the include server. I remembered to put extra quoted quotes inside the parameters that should be considered a token after the double interpretation that eval implies.
* Remove spurious warning messages about absolute filepaths.klarlund2008-05-282-9/+20
| | | | | | | | | | | | | | | | | | | | | | During compilation of the Linux kernel messages like: absolute filepath blabla.h was IGNORED (correctness of build may be affected) are issued. They are wrong. The problem is that -include or -imacro or even source file names with absolute filepaths trigger a complaint from the caching mechanism usually used for relative file names. We correct this problem. We also substitute fp_map for includepath_map, which is the now preferred term. REVIEWER: csilvers@google.com TESTS: make pump-maintainer-check make include-server-maintainer-check
* Make the Linux 2.6 kernel build with distcc-pump mode.klarlund2008-05-276-71/+164
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Add a new piece of command line parse information: send_systemdirs. This boolean is normally false. It is true if there is a -isystem option with a default system directory (one known to the compiler). With send_systemdirs true, the compiler headers of system headers are sent to the servers and mounted under the server root. The isystem option is as usual rewritten to be relative to the root. Without this flag setting, distcc quickly decides that pump mode is not viable because remote compilations fail. Also, fixed the comment: # TODO(klarlund): Make mechanism for handling -U, -undef options, along with # default symbols. (-D options are handled.) --------------- TESTS: make pump-maintainer-check make include-server-maintainer-check make 'linux kernel' in a directory made by the benchmark script (as shown below) DISTCC_EMAILLOG_WHOM_TO_BLAME=klarlund@google.com DISTCC_ENABLE_DISCREPANCY_EMAIL=1 INCLUDE_SERVER_ARGS='-d1 -t --email --unsafe_absolute_includes --send_systemdirs --stat_reset_triggers=include/linux/*' DISTCC_HOSTS="--randomize `/home/klarlund/svn-distcc/distcc/lsdistcc -P3 -k100 -d`" PATH=/home/klarlund/svn-distcc/distcc:$PATH ./masquerade pump make -j120 REVIVIEWER: csilvers@google.com TRIVIA: the about 1000 files in the kernel build involve on average 400 headers. The build time on a quad-cpu machine drops from about 1m40 s to 1m20s. With pump-mode, the machine is underutilized judging from the total sys + user time, which is around only 2m20s. NOTE: this is a successor to 'g6n8', which simply introduced a command line option to force sending of system files. A bug in gvn did not allow me to resuse that change list.
* Normalize the copyright text, to the latest version of the GPLv2 textCraig Silverstein2008-05-2323-92/+115
| | | | | | | | (typically the only change is in the FSF street address). Add Google copyright line in some places it was missing. Add GPLv2 notice to avahi patches. OKed by fergus
* Fix a typo in a comment.fergus.henderson2008-05-231-1/+1
|
* Fix unsoundess of directory replication.klarlund2008-05-232-4/+35
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Testing of distcc-pump mode on the Linux kernel showed that the computed include closure was sometimes wrong: header files were missing. Further testing showed that sometimes apparently not even the compilation unit made it across to the server or rather, the c-file in its real location would come across, but the path designating it would not. Take as an example: compilation unit is designated by the path ../bar/baz.c and the current directory is /foo but no files in /foo are in actuality referenced during include processing. Then the compilation server will get /PREFIX/bar/baz.h -- the image of the real file -- but /PREFIX/foo, the image of the current directory, will not even be created! So, this CL corrects an oversight in the way that the include processor explores files: although it does take into account the chasing of symbolic links, no provisions are in place to properly model '..' for climbing up directories. Specifically, to climb up a directory, it must exist! Usually, the directory in fact would exist on the server, because it would be replicated thanks to the header files it contains. But there is no guarentee. We solve this problem by forcing the creation of dummy .c files the first time a path is explored. Because paths are being investigated backwards (from the end), forcing results in extra files only at the "deep" levels. Since files are usually clustered together in directories, the addition of a dummy file per deep directory should be inconsequential for performance. TBD: integration level tests that this really work. I will either include that later or in a separate CL. Tested: make pump-maintainer-check build the Linux 2.6 kernel in true pump-mode (with extra patch, to be described separately that sends compiler system files to the server) [all files but a few towards the end suffer from absolute includes are preprocessed and compiled on the compile servers!] The command I used is: DISTCC_EMAILLOG_WHOM_TO_BLAME=klarlund@google.com DISTCC_ENABLE_DISCREPANCY_EMAIL=1 INCLUDE_SERVER_ARGS='-d1 -t --email --stat_reset_triggers=include/linux/compile.h' DISTCC_HOSTS="--randomize `/home/klarlund/svn-distcc/distcc/lsdistcc -P3 -k100 -d`" PATH=/home/klarlund/svn-distcc/distcc:$PATH ./masquerade pump make -j120 I had to delete /usr/bin/distcc for this work, because of the PATH acrobatics that the kernel makefiles apparently exercise. Otherwise, /usr/bin/distcc would sometimes get called. Reviewed: fergus@google.com, csilvers@google.com
* Fix a test case that was failing on MacOS X.fergus.henderson2008-05-221-1/+1
| | | | | | | The test was implicitly assuming that /tmp == realpath(/tmp). But in MacOS X, /tmp is a symbolic link to /private/tmp. Reviewers: Craig Silverstein, Nils Klarlund
* Add '--unsafe_absolute_includes' option so that more can be compiled inklarlund2008-05-214-15/+97
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | pump-mode. An occurrence of say #include "/usr/include/acl.h" will normally force the include processor to abort (even when this include is platform-specific and #ifdef'ed away), because there are no guarantees that the file /usr/include/acl.h exists on the host. With this option, these includes will be ignored. A message like: WARNING include server: absolute filepath '/usr/include/acl.h' was IGNORED (correctness of build may be affected) will be printed. The normal message raised when this option is off, the default, has been modified. It now mentions the option. An include server manual page is to follow, which explains the consequences in detail of using this option. Testing: this makes the samba-2.2.7 benchmark build without hiccups using distcc pump. Without the change many compilations fail because the include server bails out and because another unrelated bug in the include server makes it later dish up with include closures that are too small. Revievers: csilvers@google.com, fergus@google.com
* Re-enable Objective C and Objective C++ support.fergus.henderson2008-05-204-31/+26
| | | | | | | | | | | | | | | | | | This required the following changes: - Compute the default search paths for each language lazily, when the language is first encountered, rather than computing it for 'c' and 'c++'. - Make the tests of Objective C and Objective C++ conditional on the relevant compiler being installed. Tested by "make check" on two systems: - One with Objective C installed, but not Objective C++. - One with neither installed. I didn't test on a system with Objective C++ installed, because I didn't have one easily available. Reviewers: Craig Silverstein, Nils Klarlund.
* Tune gc for 25% speed-up of include processor.klarlund2008-05-172-0/+15
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Set first parameter of gc.set_threshold to 10000 instead of default 700. See comments in program text. Notes ----- I used a little script to find the include server timings as a function of this parameter. This script executes: "DISTCC_PUMP_INSTALLATION=/tmp/distcc3.0 INCLUDE_SERVER_ARGS='-t --gc_threshold=700' \ /usr/bin/time -p make-dbg -j120 -r -experimental-ld -g0 gws:gws" with various values of gc_threshold. The measurement were done on a four processor system (dual Opterons), loaded with an nxserver process. Files were served through Fuse. gc include server user + sys time setting 700 155.1s 154.0s 164.7s 186.1s 154.3s Mean: 162.8 10000 123.8s 125.4s 120.6s 120.4s 121.0s Mean: 122.2 100000 123.4s 123.3s 122.2s 123.3s 116.5s Mean: 121.7 700 166.5s 156.4s 150.8s 153.4s 155.4s Mean: 156.5 10000 120.7s 142.1s 118.3s 117.7s 119.2s Mean: 123.6 100000 140.3s 117.1s 149.7s 116.7s 141.6s Mean: 123.6 The corresponding elapsed times: 371.0s 363.3s 361.9s 370.8s 369.0s Mean: 367.2 341.8s 346.1s 336.3s 338.3s 340.5s Mean: 340.6 345.9s 338.5s 342.9s 374.4s 340.0s Mean: 348.3 377.4s 379.9s 387.1s 377.4s 373.2s Mean: 379.0 332.1s 337.7s 336.5s 332.2s 314.7s Mean: 330.6 341.1s 334.6s 357.8s 338.5s 352.4s Mean: 344.9 Note that the saving in include server time is around 35s, but the savings in elapsed time is around 44s. These numbers are uncertain. Even so, they strongly indicate that the include server is a signifcant bottleneck. BTW, for comparison with plain distcc: 821.4s 822.0s 842.4s 821.5s 852.0s Mean: 831.9 878.7s 857.2s 860.3s 880.1s 856.0s Mean: 866.5 Thus, pump-mode 2.5X is faster than plain distcc (with the set_threshold = 10000 of this CL). Revievers: fergus@google.com, csilvers@google.com
* Fix freak name collision on OSs that do not distinguish upper and lowercase inklarlund2008-05-161-2/+2
| | | | | | | | | | | | | | | | file names. The symbol 'TESTING' is potentially unevaluated according to the overapproximation semantics of the include server. So if this symbol is used in a file name and if 'TESTING' really means 'testing' according to the OS and if in fact 'testing' does exist as file, then there is trouble. Revievewers: csilvers@google.com Testing: make include-server-maintainer-check (on Linux)
* Add a comment. From klarlund@google.com:Craig Silverstein2008-05-151-4/+7
| | | | | | | | | | | | | | "Assume that: (1) 'gcc' is a link to 'distcc' earlier on PATH than the real gcc. (2) gcc is invoked; in fact that invokes distcc, which will rewrite the PATH in src/climasq.c before invoking 'gcc' again. That will make calls by distcc such as 'gcc -E' for preprocessing work OK, but the include processor will in fact be calling the 'distcc' as 'gcc' again! But because the input argument is an I/O redirection, distcc will then correctly call 'gcc'. In args.c it fails to recognize an input file."
* When running gcc -v, we were clearing the environment tooCraig Silverstein2008-05-151-3/+12
| | | | | | | | | | | | aggressively: in particular, we were clearing $PATH, which is needed to find gcc when it's in an unusual location. Now we clear everything but $PATH. Tested by running 'make include-server-maintainer-check' on solaris x86, which has gcc outside the standard path, and on linux (ubuntu), which does not. Reviewed by fergus@google.com
* Show more data when an assert fails, to help in debugging.Craig Silverstein2008-05-141-6/+5
| | | | Reviewed by klarlund@google.com and fergus@google.com
* Make distcc 3.0 compatible with Python 2.5.klarlund2008-05-141-4/+3
| | | | | | | | | | | | | Tests: I made PATH=$PATH:/usr/python/bin (where python2.5 resides on my machine). I ran autogen and reconfigured. Then I did: make include-server-maintainer-check make pump-maintainer-check and I made sure that -Is now are set according to Python2.5.
* Finish refactoring and add tests. Move _CleanOutClientRoot and _CleanOutOthersklarlund2008-05-143-82/+86
| | | | | | | | | | | | to the ClientRootKeeper package in basics. Replace _RemoveDirectoryTree with shutil function. Add tests to see that directories are created and deleted. Add a couple of title headers to basics.py. Remove a couple of now irrelevant comments. Tests: make pump-maintainer-check make include-server-maintainer-check Reviewer: csilvers@google.com
* Include snprintf.c in the source-list for the python extension. ThisCraig Silverstein2008-05-141-0/+1
| | | | | | | | | is needed on systems that don't define asprintf/etc on their own. Tested by running 'make check' on a linux ubuntu machine and on a solaris machine. Reviewed by fergus@google.com
* I found a machine whose python was compiled with _FORTIFY_SOURCE,Craig Silverstein2008-05-131-7/+18
| | | | | | | | | | | | | | | | | | which -- on linux FC5 at least -- gives warnings (which we turn into errors via -Werror) on distcc code for libc functions where we ignore the return value. Unfortunately, setup.py always uses the same flags that python was compiled with on the local machine, so _FORTIFY_SOURCE is always set when compiling distcc extensions via setup.py, and that compilation fails. Without getting into a philosophical debate about the merits of the warn-unused warnings, we'd rather the code compile, so I turn off _FORTIFY_SOURCE when it exists. Tested by running 'make check' on a goobuntu machine (no _FORTIFY_SOURCE) and an FC5 machine (with _FORTIFY_SOURCE). Reviewed by fergus@google.com
* Clean-up the handling of directories for storing compressed files.klarlund2008-05-1310-244/+290
| | | | | | | | | | | | | | | | | | The code was lacked encapsulation and proper object-orientation. (I wrote it -- it had started out much simpler.) For example, key parameteres were communicated through file name patterns expressed in file names on disk. These patterns were then inferred from reading file names. Now the patterns are properly abstracted away in a class. There should be no semantic changes. Testing: make include-server-maintainer-check make pump-maintainer-check DISTCC_CLIENT_TMP='/tmp' make pump-maintainer-check The latter test exercices the padding gymnastics.
* Be more informative on an assert.Craig Silverstein2008-05-131-1/+1
| | | | Reviewed by fergus@google.com and klarlund@google.com
* Handle gcc's "-MF" option in the include server.fergus.henderson2008-05-121-2/+2
| | | | | | | | | | This option only affects the name of the ".d" file, not the include closure, so the include server doesn't need to do anything for this option; the actual handling of this option in distcc is done by the distcc client, which already handles it correctly. This change just stops the include server from throwing an exception whenever '-MF' is encountered. Reviewers: Nils Klarlund
* Fix call of 'realpath' to work on FreeBSD. Also fix memory leak.klarlund2008-05-121-4/+27
| | | | | | Reviewer: csilvers@google.com, fergus@google.com. Tests: make pump-maintainer-check make include-server-maintainer-check
* Fix include server test so that it passes on system w/o /dev/shm.klarlund2008-05-121-4/+5
| | | | | | Reviewers: csilvers@google.com, fergus@google.com Tests: make include-server-maintainer-check
* Rewrite setup.py and associated logic to fix multiple build problems.klarlund2008-05-091-64/+156
| | | | | | | | | | | | | | | | | | | | The recent distcheck revealed that paths were set incorrectly for the setup.py script. Also, as I dug deeper, it turned out that built files were placed in the src directory (thanks to os.path.join I found in the distutil's source code, that trips up the location if relative paths contains too many '..'s.). [The distccheck revealed this!] I didn't like either that setup was run in the source directory and not from the build directory (that's asking for trouble). I did't like either that the search path, the list of includes, in setup.py was handcoded and not simply that of CPPFLAGS. Finally, I was confused by _builddir being another build directory under $(buildir). I renamed the former _include_server to better explain its role. So, I rewrote most of setup.py (correcting some of the documentation strings along the way). Also, I partly removed the 'jumping through the hoops' gymnastics from the Makefile. TESTING: 'make distcheck' REVIEWERS: csilvers, fergus
* Make this file mostly conformant with gpylint. There should be no semanticklarlund2008-05-081-112/+134
| | | | | | | consequences. TESTS: make distcheck REVIEWER: fergus@google.com
* Insert "Usage: " at the start of the usage message for the include server.fergus.henderson2008-05-071-1/+3
|
* Clean up Python code.klarlund2008-05-011-72/+67
| | | | | | -- Do TODOs. -- Heed most pylint warnings.
View Lorry Controller Status