There were two problems:
- 'glob' was using __name__ as the name of the project and not
project.current, so it would search in project root, and not in child
project that inherited project root rules.
- the 'import' that's overrided in project context used __name__ to
restore the value of current project
Thanks to Emil Dotchevski for the bug report.
[SVN r30507]
* If a generator was given a source it could not handle, it used to return
that source together with generated targets. This was nice for some use
cases, but no very nice for others, and this behaviour could not be turned
off. One use case where it worked bad was:
lib plugin : plugin.cpp helper ;
lib helper : helper.cpp ;
On windows, 'plugin' would link to the 'import library' and pass the DLL
target though. So, when installing 'plugin', we'd also install 'helper.dll',
and it was not possible to do anything about it.
* If we asked generators.construct to produce sources of type CPP,
and the selected generator produced both targets of type CPP, and of
some other type, we'd try again to convert those other targets to CPP.
This simply complicated the logic for no good reason.
* Most generator function had 'multiple' parameter, which function
was long forgotten by anybody.
As a bit of history, I believe some of the above decisions were due to a
certain use case:
CPP <------- WHL
\
WD
/
CPP <------- DLP
Here, a source file is converted to two targets with one command, and each
produced file is converted to CPP. Our generators search would notice that
there are two generators for CPP: the WHL->CPP and DPL->CPP
generators. Neither is better that the other so both are tried, and produce
(CPP, DPL) and (CPP, WHL) pairs of targets. To avoid reporting an ambiguity,
we'd try to convert, DLP to CPP and WHL to CPP, do it successfully, notice
that produced targets are the same and decide that there's no ambiguity.
However, this is rather complex logic for a relatively rare case. It can
be handled by writing another WD->CPP generator that would handle
disambiguation itself.
This commit has one user-visible change. The code:
exe a : a.cpp b ;
obj b : b.cpp helper;
lib helper ;
No longer works -- the 'a' target won't link to 'helper'. However, this is
pretty stange code and worked before almost by accident.
[SVN r29361]
The problem with the previous implementation is that it would iterate
over all elements in a dir even if 'pattern' had no metacharacters. First,
that was slow -- if you handed
/cygdrive/w/My Documents/boost/test/a.cpp
to it, it would crawl all the way to the top, listing each directory and
checking each file. Second, it would fail, because Cygwin is broken and
does not show 'cygdrive' is the list of directory elements of '/'.
Now we check if a pattern has metacharacters, and if not, just do
a simple 'timestamp' call.
The new glob is implemented as new 'GLOB-RECURSIVELY' builtin. I've decided
to use builtin since otherwise, we'd need 'does this name exist' builtin,
and if we need new builtin, why don't implement all globbing in core.
[SVN r29163]
The <tag> rule if often used in Jamfile like this: <tag>@$(__name__).rule.
The __name__ looks like Jamfile<jamfiles'directory>. When directory has
a dot, we we confused and took the part till the first dot as the
name of module.
Now, such values are processed closer to the definition point, where
we know the name of the module and don't need to guess it by looking at
the name.
Thanks to Juergen Hunold for the bug report.
[SVN r28063]
* build/generators.jam
(generator.generated-targets): Comment the desired semantics.
* util/utility.jam
(basename): Take the part till the last dot, not the first.
[SVN r27374]
tools/testing.jam
Support rules for measuring timing of build steps.
build/generators.jam
Generators with empty source lists can now consume any target type.
[SVN r27335]
builtins.{c,h}
Support for the REBUILDS rule
remove unused variable
rules.h
make.c
Support for the REBUILDS rule
make1.c
Support for the REBUILDS rule
Support for recording timing information
Restructured a case statement because it was masking a bug I
introduced.
execcmd.h
execnt.c
execunix.c
Support for recording timing information
Also removed NT-specific stuff from execunix
expand.c
Removed tabs from critical comment
Added tab-width variable setting comment for emacs.
v2/test
rebuilds.py, timedata.py, test_all.py
Tests for REBUILDS and timing.
[SVN r27334]
to direct dependents. The previous behaviour can be confusing and
there are no use cases where the new behaviour will be bad.
It will add more include paths in some cases, but who cares?
[SVN r26528]
Thanks to Toon Knapen for the bug report.
* tools/builtin.jam
(archiving-generator.run): Add <library> properties to sources.
* test/library_chain.py: New test.
[SVN r26447]
* tools/stage.jam
(copy-file, relink-file): Add stage's dependency properies to
the staged targets properties, so that <dependency> is copied too.
* test/stage.py: New test.
[SVN r26420]
enviroment variables in the list of values. No it's possible to add
an element to a variable.
* build/project.jam: Remove obsolete test.
* build/toolset.jam (find-property-subset): No longer local,
so that test can call it.
* test/test.jam: Import 'print' and 'common'.
* util/print.jam: Comment out failing unit test.
[SVN r26413]
* build/targets.jam
(project-target.find): Better error reporting.
* build/project.jam
(load-jamfile): Use 'path.native' when loading Jamfile,
otherwise, target id with absolute directory name won't
work for Windows.
[SVN r26271]
