| Commit message (Collapse) | Author | Age | Files | Lines |
... | |
|/
|
|
|
|
|
|
| |
"^[^%\n]*" at the beginning of the reg ex would match all but the last
include so that they were lost. There is no point checking for "%" since
comments are stripped anyways, so just remove that part.
Also add test case for multiple includes per line.
|
| |
|
|
|
|
| |
propagate and cause SCons to exit
|
| |
|
|
|
|
| |
false
|
| |
|
| |
|
|\
| |
| | |
Don't hide exceptions when command does not exist
|
| | |
|
|/
|
|
|
|
|
|
| |
Minor: replace an instance of "for i in range(len(foo))" with the now
preferred "for i in foo". This change didn't naturally fit with any of
the other instances of this usage, so is submitted by itself.
Signed-off-by: Mats Wichmann <mats@linux.com>
|
| |
|
| |
|
|
|
|
| |
Also improve exception handling message in test.subdir()
|
| |
|
|\
| |
| | |
Update docs for parse_flags keywords
|
| |
| |
| |
| |
| |
| |
| |
| |
| | |
The somewhat poorly named parse_flags keyword args actually work like
the MergeFlags method - don't just split like the ParseFlags method,
but do the merging as well. Tweak the docs and add a reference to
env.MergeFlags.
Signed-off-by: Mats Wichmann <mats@linux.com>
|
|/ |
|
|
|
|
|
|
|
|
|
|
|
| |
Regexes that contained unescaped backslashes and were not listed
in raw string form caused one more test failure when Python 3.8
was experimentally turned on in the Travis CI build.
Also one utility script had the same, not affecting tests - found through
inspection.
Signed-off-by: Mats Wichmann <mats@linux.com>
|
|
|
|
|
|
| |
Needed a doc regen to pick up the change.
Signed-off-by: Mats Wichmann <mats@linux.com>
|
|
|
|
|
|
|
|
|
|
|
| |
Recognize two additional GNU compiler header directory search options:
-iquote and -idirafter. Each takes a following arg, which scons now
recognizes and adds together with the option to CCFLAGS. Note that
(similar to -isystem which was added in git commit f8614aa2), this does
not tell scons anything special, it only recognizes the flag + argument
so it can be passed on to the compiler.
Signed-off-by: Mats Wichmann <mats@linux.com>
|
|
|
|
|
|
|
| |
The routine used in the builder if lxml is being used had
a doubled write command likely to produce unexpected results.
Signed-off-by: Mats Wichmann <mats@linux.com>
|
|
|
|
|
|
|
|
|
| |
One file close ended up in the wrong method, meaning it was closing
a file which needed to stay open for further reading, while another
method of the same name in a different class was missing one.
(this didn't trigger any test fails, hmmm, just visual inspection)
Signed-off-by: Mats Wichmann <mats@linux.com>
|
|\
| |
| | |
[WIP] Avoid cachedir races
|
| |
| |
| |
| |
| |
| |
| |
| | |
sider complained about spelling in a comment after I
updated the comment so it got to look at those lines
as part of the change. Fix those.
Signed-off-by: Mats Wichmann <mats@linux.com>
|
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Simplify the Py2 check for an existing-but-old cachedir
by using any and a generator expression.
Move an import to module scope.
Signed-off-by: Mats Wichmann <mats@linux.com>
|
| |
| |
| |
| | |
Signed-off-by: Mats Wichmann <mats@linux.com>
|
| |
| |
| |
| |
| |
| |
| |
| |
| | |
1. Add a Py3-only version of the cachedir config read,
using exclusive open to avoid races.
2. In the Py2-only version, add the hack from issue #3351
to dodge one source of races.
Signed-off-by: Mats Wichmann <mats@linux.com>
|
|\ \
| |/
|/| |
[wip] Py38warns4 tests
|
| |
| |
| |
| | |
Signed-off-by: Mats Wichmann <mats@linux.com>
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
File closes in msvs tool
Context manager in rpm tool
Dummy compiler, assembler scripts can be called in unexpected
ways (namely by gcc tool init), which passes --version flag.
don't take exception on the getopt call.
Try again to undo my breakage of _subproc, which was failing on Win+PY27
Signed-off-by: Mats Wichmann <mats@linux.com>
|
| |
| |
| |
| | |
Signed-off-by: Mats Wichmann <mats@linux.com>
|
| |
| |
| |
| | |
Signed-off-by: Mats Wichmann <mats@linux.com>
|
| |
| |
| |
| |
| |
| |
| | |
On a linux host (missing some things that may be on the Travis CI
setup), Py3.8a3 now shows 19 fails, 1048 pass, with 84 Warning: messages.
Signed-off-by: Mats Wichmann <mats@linux.com>
|
| |
| |
| |
| | |
Signed-off-by: Mats Wichmann <mats@linux.com>
|
|\ \
| | |
| | | |
Some more lint-derived cleanups
|
| |/
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Consistently use "not is" and "not in", many instances used
the form "not x is y" instead, which pylint objected to.
A couple of bare except clauses got a qualifier.
Files otherwise touched had trailing whitespace cleaned up as well.
These are all things that sider would complain about if a change
happened nearby, so this is pre-emptive.
Signed-off-by: Mats Wichmann <mats@linux.com>
|
|\ \
| | |
| | | |
Enhanced debug explain
|
| |\ \ |
|
| | | |
| | | |
| | | |
| | | | |
ampping None to allow for explain to handel old and new dependency lists of different lengths
|
| |\ \ \
| | | |/
| | |/| |
|
| | | |
| | | |
| | | |
| | | | |
components to dependency list. Updated test. TODO: better formatting
|
|\ \ \ \
| |_|_|/
|/| | | |
Fix Issue #3135 - Type Bound procedures in Fortran submodules
|
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
Add subroutines that are declared pure and elemental to the test
of the Emitter. Note that in test_1.f90, the interface is repeated
in the submodule, whereas in test_2.f90 the interface is taken from
the module. Also note that the regex does not check whether
"module pure" or "module elemental" is actually followed by "subroutine" or
"function". This would be a syntax error and should trigger a compile time
error.
|
| | | |
| | | |
| | | |
| | | | |
processing interface module declarations
|
| | | | |
|
| | | | |
|
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
This fixes issue #3135 regarding the issue with using type bound
procedures in Fortran submodules. The fix consists of changing the
regex used in the scanner and the emitter to ignore lines starting
with:
module subroutine
and
module function
as these are used to define type bound procedures instead of modules
named 'subroutine' or 'function'. The regex is case insensitive.
|
|\ \ \ \
| | |_|/
| |/| |
| | | | |
fix_mongo_bug_33111_multiple_targets_via_emitter_spurious_rebuilds
|
| | | |
| | | |
| | | |
| | | | |
Signed-off-by: Mats Wichmann <mats@linux.com>
|
| | | |
| | | |
| | | |
| | | | |
Signed-off-by: Mats Wichmann <mats@linux.com>
|