| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
|
| |
development. Note that this set of changes is NOT backward-compatible;
the trunk no longer works with Python 1.5.2, 2.0, or 2.1.
|
| |
|
|
|
|
| |
to some regex problem).
|
|
|
|
| |
a Builder object, or known to generate a Builder object, or a callable.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
http://scons.tigris.org/svn/scons/branches/pending
........
r4648 | managan | 2010-01-22 09:40:52 -0800 (Fri, 22 Jan 2010) | 9 lines
Added dismbiguate calls on a targets side effects in the
make_ready_all and make_ready_current routines of Taskmaster.py.
This fixes Stefan Hepp's problem with latex. Directories to
hold the side effect files were not created when using
variantDir with duplicate=0
Modifying subdir_variantdir_include2.py to run scons with an
explicit target to reveal this problem.
........
r4654 | stevenknight | 2010-01-27 07:41:01 -0800 (Wed, 27 Jan 2010) | 5 lines
Issue 2534: fix Scanners' default ability to return Dir nodes by
making the default node_class the real SCons.Node.FS.Base class,
not SCons.Node.FS.Entry (which is now, and has been for some time,
a subclass).
........
|
| |
|
|
|
|
|
| |
This change should itself be backed out before merge "pending"
back into trunk.
|
|
|
|
|
|
| |
command tries to work with a directory named foo instead of the
file foo.tex. The builder now ignores a directory and continues
searching to find the correct file.
|
|
|
|
| |
Add commented line that will reveal a failure.
|
| |
|
|
|
|
| |
file is listed in both the env.PCH() and env.Program() calls.
|
|
|
|
| |
.py file associations.
|
|
|
|
|
|
|
|
| |
directory of the main file instead of the file it is included from.
Also update the emitter to add the .aux file associated with
\include{filename} commands. This makes sure the required directories
if any are created for variantdir cases.
Half of the patch from Stefan Hepp.
|
| |
|
|
|
|
| |
Add a necessary serialVersionUID.
|
|
|
|
| |
RMIC tests to silence warnings from gcj-4.4.
|
|
|
|
| |
moc and uic utilities by using fputs() instead of printf().
|
|
|
|
| |
Patch from Stefan Hepp.
|
| |
|
|
|
|
|
|
|
| |
by default. Fix use of subprocess in Python 2.4+ and exit statuses
of popen'ed scripts in earlier versions of Python. Support the ability
to execute a directory's tests by naming the directory as a command-
line argument.
|
|
|
|
| |
the tests of those options expect an exit status of 1.
|
|
|
|
| |
printing "Entering Directory" with -C/-u.
|
|
|
|
|
| |
to get the path to the SDK include directory (so our test program can
#include <windows.h>).
|
| |
|
|
|
|
| |
Use test.wrap_stdout() instead of hard-coding the SCons messages.
|
|
|
|
|
|
|
|
|
|
| |
file that contains the documentclass command
Add logic to is_LaTeX routine to search through included files until
\documentclass is found
Also added comments, converted comments at start of 2 routines into doc
strings..
|
|
|
|
| |
Make an educated guess on the generated module name based on the .i file name.
|
| |
|
|
|
|
|
| |
TryRun fails to find the executable when VariantDir is set up from
SConscript/SConstruct.
|
|
|
|
|
|
| |
centos5 slave).
Code by Rob Managan.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
failing tests by setting SCONSFLAGS='-warn=no-python-version' in def
run() in TestSCons.py yesterday, where previously it was done in def
__init__(...).
This fixed most of the broken tests, but broke some of the working tests.
I readded the logic to def __init__
In addition I've added an extra arg to the __init__ for TestSCons.py
ignore_python_version (which defaults to true).
For those tests which need to be able to see the deprecation and/or
test the functionality of the deprecation.
|
| |
|
| |
|
|
|
|
| |
deprecation errors aren't output when --debug=memoizer is on command line, but are when it's passed via SCONSFLAGS.
|
|
|
|
| |
(hopefully this doesn't break other versions)
|
|
|
|
| |
Otherwise skip the test.
|
|
|
|
|
|
|
|
|
|
|
| |
I found that when only one tool (say pdftex) is initialized
that some of the actions for bibtex... are not created.
Moved most common actions and environment settings to one routine
in tex.py that the other text tools (pdftex, pdflatex, latex) call.
Also fixed a typo in the nomenclature action setup. There was
a '$' that should not have been there.
|
|
|
|
| |
to support it
|
| |
|
|
|
|
| |
to execute in up to 0.002 seconds.
|
|
|
|
|
|
|
| |
(specifically of output matches using regular expressions).
Update tests for corresponding inteface changes.
Add use of diff_re() to test/sconsign/script/Configure.py
so we can get accurate information about its buildbot failure.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This was prompted because MikTeX, used on Windows,
does not put the same information on files opened
into the log file. The -recorder option gives a .fls
file that is the same on all platforms.
We still use the .log file contents to find warnings
and errors that mean we need to rerun latex...
Also add message about errors so user does not have
to scroll up through all the latex output to find
if there was an error.
Update all tests to handle the new command line option.
Add one more test on grpahics conversion.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
msvc, mslib, mslink
as they cause test/import.py to fail.
I don't have an OS2 system to resolve the issues on.
Merged revisions 4065,4067-4068,4071-4072,4074,4076,4085,4111,4163,4222 via svnmerge from
http://scons.tigris.org/svn/scons/branches/vs_revamp
................
r4065 | stevenknight | 2009-03-02 16:48:05 -0800 (Mon, 02 Mar 2009) | 42 lines
Copy over changes by hand from branches/core (I was working in a
directory pointing to the wrong branch):
........
r4059 | stevenknight | 2009-03-02 06:54:13 -0800 (Mon, 02 Mar 2009) | 14 lines
Initial prototypes of modules for architecture support (MSCommon/arch.py)
and Visual C/C++ configuration distinct from (and in preference to)
the Visual Studio configuration, incorporating configuration info and
ideas from Jason Kenny's latest parts/tools/MSCommon work.
This has a lot of cut-and-paste right now and does NOT actually work.
In particular, there are no changes to the rest of the infrastructure
to actually call any of this new stuff. I'm checking these in so I can
sync on other systems for continued development.
The ultimate goal is for much of the current MSCommon/vs.py module's
logic to migrate to MSCommon/vc.py, since that's what we're actually
interested in configuring.
........
r4058 | stevenknight | 2009-03-02 06:49:35 -0800 (Mon, 02 Mar 2009) | 2 lines
Fix doc string.
........
r4056 | stevenknight | 2009-03-01 06:10:31 -0800 (Sun, 01 Mar 2009) | 4 lines
Change the sanity_check file for 6.1 per Jason's update.
Add an optional version string argument to detect_sdk().
Import directly from common, not SCons.Tool.MSCommon.common.
........
r4055 | stevenknight | 2009-02-28 05:29:03 -0800 (Sat, 28 Feb 2009) | 5 lines
Minor common.py enhancements:
Refactor is_win64() to cache its return value so we don't have to hit
registry every time we want to know if the OS is 64-bit.
Accept SCONS_MSCOMMON_DEBUG=- to mean print debug messages to stdout.
........
................
r4067 | stevenknight | 2009-03-06 04:59:21 -0800 (Fri, 06 Mar 2009) | 4 lines
Fix entries in batch_file_map, after a double-check with the original code.
Use a host_architecture of None in the map to indicate the "default"
entry if there isn't a specific .bat file for our current architecture.
................
r4068 | stevenknight | 2009-03-06 05:15:39 -0800 (Fri, 06 Mar 2009) | 2 lines
Add include_subdir and lib_subdir data to the SDK definitions.
................
r4071 | stevenknight | 2009-03-10 07:42:53 -0700 (Tue, 10 Mar 2009) | 4 lines
Capture basically working snapshot.
src/engine/SCons/Tool/msvsTests.py and test/IDL/midl.py still fail,
and code needs clean up.
................
r4072 | stevenknight | 2009-03-11 06:38:37 -0700 (Wed, 11 Mar 2009) | 2 lines
Fix path to Common7\Tools. Find vs*.bat files.
................
r4074 | stevenknight | 2009-03-13 17:11:55 -0700 (Fri, 13 Mar 2009) | 3 lines
Import SCons.Tool.MSCommon.vc so it works regardless of other sys.path
manipulation.
................
r4076 | stevenknight | 2009-03-19 08:08:48 -0700 (Thu, 19 Mar 2009) | 7 lines
Refactor SDK support:
* New mssdk_exists() and mssdk_setup_env() methods as the primary
public entry points.
* Call internal sdk.*() utility functions that just return information
(like the SDK structure) not handle the actual setting.
* Give the VisualStudio definitions explicit settings for the SDK version.
................
r4085 | bdbaddog | 2009-04-05 14:30:46 -0700 (Sun, 05 Apr 2009) | 3 lines
Fix "can't find module vc" issue when building on windows.
................
r4111 | stevenknight | 2009-04-19 07:47:47 -0700 (Sun, 19 Apr 2009) | 2 lines
Doc updates for how MSVS_VERSION is used.
................
r4163 | bdbaddog | 2009-04-29 17:32:43 -0700 (Wed, 29 Apr 2009) | 4 lines
Fix for bug 2405
Also refactored some logic from Defaults._defines() into Defaults.processDefines() which is then used by msvs.py
................
r4222 | bdbaddog | 2009-05-31 17:01:22 -0700 (Sun, 31 May 2009) | 8 lines
Added definition of HOST_OS|ARCH TARGET_OS|ARCH (For win32 only right now)
Migrated some logic to determine arch from Tools/MSCommon -> Platform/Win32.py
Fixed lots of minor issues with string values for directories, architectures, and registry keys
Switch to find VS installs by registry (was using VC install location previously)
Added but commented out that HOST_{OS|ARCH} and TARGET_{OS|ARCH} will become reserved var names in Environment()
Need to resolve issues that SCons is issuing messages when SCons itself sets these variables.
................
|
|
|
|
| |
on systems that actually support it
|
| |
|
| |
|
| |
|
|
|
|
| |
of 'cat' if win32.
|
|
|
|
| |
than having a list in the regression test that has to be updated.
|