| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
|
| |
All regression tests pass. Clients are working live. PPS observed on Macx-1.
|
|
|
|
| |
No code or logic changes. All regression tests pass.
|
|
|
|
|
|
| |
The backwards-compatibility code caused a bug.
All regression tests pass.
|
| |
|
| |
|
|
|
|
|
|
| |
Also, ensure this always happens within the regression-test driver.
All regression tests pass.
|
|
|
|
|
|
| |
Previous rule was cleaning only the last one made.
No code changes.
|
|
|
|
|
|
|
| |
As a nice side effect, fix Savannah bug #43642: static build failure -
is trying to build a shared library.
No code changes.
|
|
|
|
|
| |
The other one works better - it needs to go after -lm for one-pass
linkers.
|
| |
|
|
|
|
| |
No code changes.
|
|
|
|
|
|
| |
The comment in the changed portion explains it.
All regression tests pass.
|
|
|
|
|
|
|
|
|
| |
pylint was not checking python_progs correctly anymore due to name
mismatch. Converted to joining python_progs and a list of other scripts
to explicitly check
also added pep8 audit with error/warnings ignored for stylistic issues
that are not blockers.
|
| |
|
|
|
|
| |
All regression tests pass.
|
| |
|
|
|
|
| |
All regression tests pass.
|
|
|
|
| |
All regression tests pass.
|
| |
|
|
|
|
| |
This means better odds of porting to uclibc, mingw, etc.
|
|
|
|
|
|
|
|
| |
Instead, we're going to try telling the tty layer to deliver input in chunks
larger than chracters, with a timeout. That way the select call in the main
loop will return data ready less often.
All regression tests pass.
|
|
|
|
| |
All regression tests pass.
|
|
|
|
|
|
|
|
|
|
|
| |
Add guards when adding in warnings to check for so that we don't add
duplicate entries to CFLAGS
This also deals with the warning generated with scons qt=on
about -Wmissing-prototypes not a valid C++ option
The error is that qt_env is stripping one copy of -Wmissing-prototype
and -Wstrict-prototyes from CFLAGS so that it can compile with CXX
|
| |
|
| |
|
| |
|
| |
|
|
|
|
|
|
| |
As a side effect, simplify configuration slightly.
All regression tests pass.
|
|
|
|
|
|
|
|
|
|
|
|
| |
Guard includes of termios.h with #ifdef HAVE_TERMIOS_H.
Test for termios.h in scons, and define/undefine HAVE_TERMIOS_H accordingly.
Fixes the below error seen on cygwin, which lacks termios.h:
In file included from gpsutils.c:22:0:
gpsd.h:349:21: fatal error: termios.h: No such file or directory
#include <termios.h>
^
|
|
|
|
|
|
|
|
| |
At least on FreeBSD, one needs sys/types.h. This seems to be a
header present on all systems that has sys/timepps.h, so I don't see a
reason to conditionalize this.
>From Daniel O'Connor via gpsd-dev@.
|
|
|
|
| |
All regression tests pass.
|
| |
|
|
|
|
|
|
| |
Fortunately, there's a much simpler way to do the computation correctly.
All regression tests pass,
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
The recent addition of a prototype for gmtime in gpsd.h breaks the
cygwin compile. Cygwin has gmtime, so the declarations conflict.
Handle this by adding an scons check for gmtime, and make declaration
of the bundled version conditional on absence of the system-supplied
version.
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1
iQEcBAEBAgAGBQJUJvXAAAoJECH/phKUbjXCdgwH/RrMznXe51HHR5eLD3Y/vor8
h514XtvlzB/ghWQcAuiWX8mGy/J7k8OJFb8HmUXPAhSuzeiRooZXLz7uk3UvfeGv
4p5MfND3E7FCZVBu9/o9n5gf27wddXKChHVfAZ1MsnusRmwfjAiPoTr0iVoarWLw
s/8tjvnHJg4A3i08nu4Y+Z9rwyDqkmB2YPQ0Zs+deCsbK4AA4Lx3oVkgDkVCk7zz
vBZPueS6ylAMb1hbEthW3l7St3S2ESLG7nIxwv1lQ0mmAi70pknHeLYbiBLU0QqT
A2ZsOOajespqq9/pmxP6ZJcwy8DtNXcj8LvRcNAOrjiV52MQUwTZGLLwVP5WMxk=
=PoQy
-----END PGP SIGNATURE-----
|
|
|
|
| |
All regression tests pass.
|
| |
|
| |
|
|
|
|
|
| |
For some odd reason trying to faxtor out the symmetrize operation
induces a core dump. To be investigated.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
Move '-Wmissing-prototypes' from CCFLAGS to CFLAGS. Silences build
warning:
cc1plus: warning: command line option '-Wmissing-prototypes' is valid
for C/ObjC but not for C++
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1
iQEcBAEBAgAGBQJUDAa+AAoJECH/phKUbjXCHRYIAM8HGbViHXTYdzhZaFQXBEtk
Fgk4u4OGO+12DLP+RxRbyzdX31fe3IfRANKCtblNX/32edZcJmwQxmQf6vpXx/tM
RWT6vNtTQJkjoQAwCjqp0WYeKePEJyJCZA6KrrPR0Xn+zzMe5okdciZKwTqg89bM
B9u6zrJ/hg2UmjGJr5VicIBa+3Zz0rbor0QqJxRmUn2/WC+5xPtCFG8wJ8uJ3/JP
j0eBhUuCj9PUhnoN/rJwea6QtkIeX0e5A4iQR8XJ6meFI6Qd1ux5KyQ+Tnp/Fl7T
6oWz30XULshBL6p91MwQS7EK6bk1HqmYghyD1LdTnRhYLdV4PfEEAmZ9mzoH7dM=
=Bwsk
-----END PGP SIGNATURE-----
|
|
|
|
| |
...but thy won't of you turn nofloats on.
|
| |
|
|
|
|
|
| |
All regression tests pass with default options (GPS tests are disabled with
minimal on, because socket_export is off).
|
|
|
|
| |
...indirect effects break the build. This will be fixed.
|
|
|
|
|
|
|
|
|
|
|
|
| |
(Default inverts too, so rge build still defaults to stripping.)
This is the only change required to make a build with all options False
be minimal. It's preparation for a 'minimal' option to make life slightly
easier for the Android and embedded guys.
Required that the scons line for Android be retconned.
All regression tests pass.
|
| |
|
|
|
|
|
|
| |
Some toolchains, such as the Android one used with Bionic, have
linkers that only resolve in one direction rather than making
multiple passes. This switch is required to keep them happy.
|
|
|
|
|
|
|
|
| |
The version of gcc 4.9 in Cygwin doesn't understand '-pthread'.
So the attached patch conditionalises supplying that flag to the
compiler using CheckCompilerOption.
No new compile/test/check problems seen under Linux, and it gets the
compile a little further under cygwin.
|
| |
|
|
|
|
|
| |
This makes libQgpsmm actually be a deprecated synonym. However, this
is pointing out that it might be best to just get rid of it.
|
| |
|