Compile-tested with commit 8ea3f0e (tag: 3.4.10)
from https://github.com/osmocom/libusrp
Fixes:
usrp_impl.cc:33:10: fatal error: usrp_standard.h: No such file or directory
usrp_impl.cc:49:63: error: expected '(' before 'malloc'
usrp_impl.cc:49:100: error: expected ')' before ';' token
usrp_impl.cc:72:39: error: invalid 'static_cast' from type 'rig_state*' to type 'usrp_priv_data*'
usrp_impl.cc:86:39: error: invalid 'static_cast' from type 'rig_state*' to type 'usrp_priv_data*'
usrp_impl.cc:104:39: error: invalid 'static_cast' from type 'rig_state*' to type 'usrp_priv_data*'
usrp_impl.cc:129:45: error: invalid 'static_cast' from type 'rig_state*' to type 'usrp_priv_data*'
usrp_impl.cc:151:45: error: invalid 'static_cast' from type 'rig_state*' to type 'usrp_priv_data*'
usrp_impl.cc:169:45: error: invalid 'static_cast' from type 'rig_state*' to type 'usrp_priv_data*'
usrp_impl.cc:114:37: warning: invalid suffix on literal; C++11 requires a space between literal and string macro [-Wliteral-suffix]
usrp_impl.cc:139:38: warning: invalid suffix on literal; C++11 requires a space between literal and string macro [-Wliteral-suffix]
/usr/bin/ld: ../src/.libs/libhamlib.so: undefined reference to `usrp_caps'
If pkg-config has found tcl.pc, then tclConfig.sh is not executed
and $TCL_INCLUDE_SPEC is empty so the test for tcl.h fails.
This patch moves the test inside the "else" case where the variable
is always defined.
The is a bug exposed by the fact that --with-xml-support redefines
the macro PKG_CHECK_MODULES.
Fix verified moving aside /usr/lib/x86_64-linux-gnu/pkgconfig/tcl.pc
Fixes:
configure: error: Unable to find Tcl headers
Library directory ../dummy/.libs doesn't exist anymore and ../c++/.libs
(a.k.a. .libs) is already used by the libtool wrapper script that runs
the testcpp program.
This makes it possible to run "make -C src/" or "make -C tests/ rigctl"
or "make -C bindings/ check" (and so on) in a clean tree, but it doesn't
rebuild those targets if libhamlib.la is changed; for this run make from
the top directory as usual, to rebuild all SUBDIRS if needed.
* Enable `.fm_filters` in `IC705_priv_caps`
* `icom_get_mode_without_data()`: activate FM filter selection code if `RIG_IS_IC705`
* `icom2rig_mode()`: activate FM filter fixed width code if `RIG_IS_IC705`
* TODO: cases in WFM should be solved independently
* `icom2rig_mode()`: handle FM and WFM separately and correctly at least for IC-705, no changes for IC-7300 and IC-9700
* `icom_get_mode_without_data()`: add WFM to the code assuming that values from `icom2rig_mode()` is correct
* icom.c: A partial rollback for a395b91be6
* The workaround to use `icom_set_mode_without_data()` is not necessary
* The later experiments showed CI-V command 0x26 worked OK too for WFM
* Add WFM freq to ic705_caps.filters
* Fix icom_set_mode_x26() FM behavior
`icom_set_mode_x26()` did not pass the correct command value
for FM or PKTFM modes when width is set to
`RIG_PASSBAND_NORMAL` (i.e., 0 (zero)).
With this source code change,
the command value `buf[2]` is forcefully set to 1
when `RIG_PASSBAND_NORMAL` or `RIG_PASSBAND_NOCHANGE` are
passed to the parameter `width`.
This fix solves the bug for IC-705 with rigctl when
entering the command `M FM 0` after `M WFM 0` *did not*
change the mode properly to (narrow) FM.
Per GitHub issue #1704, frequencies higher than 1 GHz passed from AC Log
have an embedded comma. Even though sscanf() offers the "'" (single
quote) character as a means of ignoring thousands separator, apparently
it depends on the environment variable LC_NUMERIC being set correctly
and that may not be supported on all platforms.
This patch just parses through the string while skipping any comma that
may appear and then uses strtold() to convert to a numeric variable. It
is supected that AC Log always uses a comma as a thousands separator.
Daniele Forsi IU5HKX