William S Fulton [Fri, 18 Oct 2024 07:38:20 +0000]
python: update for SWIG 4.3
SWIG 4.3 comes with an API change to the SWIG_Python_AppendOutput()
function, so switch to the recommended SWIG_AppendOutput() instead.
Because the new function is a macro in fact also switch to the {..}
delimiters as the code needs to be preprocessed. The change is
compatible with the previous SWIG releases 4.2 and 4.1 at least.
Fixes: https://bugzilla.redhat.com/show_bug.cgi?id=2319133
Related: https://github.com/swig/swig/pull/2907
Yegor Yefremov [Fri, 21 Mar 2025 07:51:51 +0000]
CMakeLists.txt: remove obsolete policies
Both CMP0011 and CMP0057 are covered by setting the CMake minimum
version to 3.5.
Yegor Yefremov [Tue, 12 Dec 2023 13:20:36 +0000]
CMake: bump the minimal required version to 3.5
Older CMake versions are treated as deprecated.
Dan Horák [Mon, 19 Dec 2022 16:40:43 +0000]
python: move from distutils to sysconfig
The distutils module was deprecated in Python 3.10, and will be removed
in 3.12 [1], thus switch to the sysconfig module instead.
[1] https://peps.python.org/pep-0632/
Yegor Yefremov [Thu, 12 Oct 2023 07:11:13 +0000]
CMake: use ${PC_LIBUSB_LIBRARIES} instead of a library name
Patch taken from the Debian package:
https://sources.debian.org/src/libftdi1/1.5-6/debian/patches/02-kfreebsd.diff/
Yegor Yefremov [Thu, 12 Oct 2023 07:11:12 +0000]
CMake: fix multiarch support
Let CMake determine the exact library destination path.
Patch taken from the Debian package:
https://sources.debian.org/src/libftdi1/1.5-6/debian/patches/01-cmake-multiarch.diff/
Yegor Yefremov [Mon, 11 Sep 2023 08:04:04 +0000]
CMake: make the project compatible with building as a subproject
Use CMAKE_CURRENT_*_DIR in the main CMakeLists.txt and PROJECT_*_DIR
in the subdirectories.
Yegor Yefremov [Mon, 11 Sep 2023 08:04:03 +0000]
ftdipp/CMakeLists.txt: remove VIM modline settings
Indentation settings will be handled via the EditorConfig project.
Yegor Yefremov [Mon, 11 Sep 2023 08:04:02 +0000]
CMake: rework subdirectory handling
Move all mandatory subdirs to the top and also make if clauses more
readable.
Yegor Yefremov [Mon, 11 Sep 2023 08:04:01 +0000]
CMake: rework findlibusb module
Rename the module to FindLibUSB.cmake and use it as the only way
of finding libusb.
Bonus: perform CMake linting of the module.
Yegor Yefremov [Mon, 11 Sep 2023 08:04:00 +0000]
CMake: report CMake version
Yegor Yefremov [Mon, 11 Sep 2023 08:03:59 +0000]
CMake: use dedicated recipe for documentation generation
Use PROJECT_SOURCE_DIR and PROJECT_BINARY_DIR to refer to the top
source directory.
Yegor Yefremov [Fri, 1 Sep 2023 09:33:02 +0000]
CMake: disable deprecated behavior
Set policies CMP0011, CMP0057, CMP0099 to the new behavior. Since
version 3.27.4 if the policy is not set to the new behavior, CMake
will generate a warning.
Yegor Yefremov [Thu, 31 Aug 2023 07:34:10 +0000]
Doxygen: resolve warnings
Remove obsolete values like MSCGEN_PATH, TCL_SUBST, etc. from the
configuration file. All these options used the default values, hence
it is safe to remove them.
Paper type a4wide doesn't exist, so use a4.
Yegor Yefremov [Thu, 31 Aug 2023 07:23:30 +0000]
CMake: rework documentation creation section
Add comments and also make the if ( DOCUMENTATION ) section more
readable.
Yegor Yefremov [Thu, 31 Aug 2023 07:23:29 +0000]
CMake: require 2.8.12 for cmake_example
Make CMake minimal version requirement consistent for the entire
project.
Yegor Yefremov [Thu, 31 Aug 2023 07:07:51 +0000]
examples/eeprom.c: use AUTO_DETACH_REATACH_SIO_MODULE
This way, the device's state remains unchanged after the
eeprom invocation.
Bonus: remove trailing white spaces.
Yegor Yefremov [Wed, 30 Aug 2023 09:57:01 +0000]
CMake: fix indentation and normalize braces
Matthias Klein [Fri, 18 Nov 2022 08:59:40 +0000]
ftdi_readstream: fix timeout setting
Divide the whole seconds into tv_sec, and the remaining microseconds into tv_usec.
Signed-off-by: Matthias Klein <matthias@extraklein.de>
Chris Johnson [Wed, 3 May 2023 13:52:01 +0000]
Introduce ftdi_set_module_detach_mode()
Added a public setter to allow the usage of the
AUTO_DETACH_REATACH_SIO_MODULE feature added in
5bf1c1e3ff1616fda20c26cc3e3df7b807744c99.
Yegor Yefremov [Mon, 31 Jul 2023 10:23:41 +0000]
EEPROM: make user_data_addr value readable
Before this change, the user_data_addr could be only written
but not read.
Thomas Jarosch [Tue, 29 Aug 2023 13:47:41 +0000]
FT2232C: show mode for B channel
Idea by Yegor Yefremov. Thanks!
Yegor Yefremov [Tue, 29 Aug 2023 08:50:28 +0000]
FT4232: show mode for both C and D channels
Yegor Yefremov [Tue, 29 Aug 2023 08:57:00 +0000]
Align is_not_pnp property
As seen when written with FTD2XX.
Yegor Yefremov [Wed, 9 Aug 2023 18:20:43 +0000]
examples: eeprom: add the missing newlines
Yegor Yefremov [Wed, 9 Aug 2023 12:38:26 +0000]
EEPROM: mention FT230X in the dynamic content calculation
Yegor Yefremov [Wed, 9 Aug 2023 12:28:15 +0000]
EEPROM: fix checksum for FT230X series
ftdi_eeprom_build() computes the checksum via reading the "Factory
Configuration Data" region. ftdi_eeprom_decode() computes the checksum
only on the EEPROM buffer. Hence, both checksums differ.
Add the info from "Factory Configuration Data" to the EEPROM buffer.
Yegor Yefremov [Sat, 5 Aug 2023 06:32:58 +0000]
Add .editorconfig for C and CMake files
Yegor Yefremov [Thu, 3 Aug 2023 13:48:58 +0000]
Fix typos and harmonize spaces
Yegor Yefremov [Tue, 1 Aug 2023 09:53:35 +0000]
CMake: bump the minimal required version to 2.8.12
The older CMake versions are deprecated.
Remove support for CMake code compatible with CMake older than 2.8.8.
Also move cmake_minimum_required() command before the project()
command invocation.
Fabian Vogt [Tue, 29 Aug 2023 12:57:27 +0000]
Fix race during build of python bindings
The CMake documentation says about add_custom_command that the output must
not be used as dependency in more than one target, but this was the case
here: Both the doc_i target and the SWIG target depended on ftdi1_doc.i.
Make the SWIG target depend on doc_i instead.
This fixes that ftdi1_doc.i was built twice with parallel builds, leading
to random errors.
Signed-off-by: Fabian Vogt <fvogt@suse.de>
Fabrice Fontaine [Fri, 13 May 2022 16:11:50 +0000]
CMakeLists.txt: fix static build with libusb and -latomic
Use pkg-config to retrieve libusb dependencies such as -latomic and
avoid the following static build failure:
/nvmedata/autobuild/instance-29/output-1/per-package/libftdi1/host/bin/../lib/gcc/sparc-buildroot-linux-uclibc/10.3.0/../../../../sparc-buildroot-linux-uclibc/bin/ld: /nvmedata/autobuild/instance-29/output-1/per-package/libftdi1/host/sparc-buildroot-linux-uclibc/sysroot/lib/libusb-1.0.a(core.o): in function `libusb_unref_device':
/nvmedata/autobuild/instance-29/output-1/build/libusb-1.0.25/libusb/core.c:1186: undefined reference to `__atomic_fetch_add_4'
Fixes:
- http://autobuild.buildroot.org/results/
1ca7cd85ae60ad4797a6d8a83b2fb51d7eab96d9
Signed-off-by: Fabrice Fontaine <fontaine.fabrice@gmail.com>
Thomas Jarosch [Tue, 4 May 2021 16:51:15 +0000]
Fix indentation
Thomas Jarosch [Sun, 2 May 2021 07:32:02 +0000]
Add Tomasz Wasilczyk to AUTHORS
Tomasz Wasilczyk [Wed, 21 Apr 2021 19:43:34 +0000]
Don't set EEPROM serial when it's not used.
This makes a difference in USB enumeration -- if serial is missing,
its string is not present in string descriptor list.
Tomasz Wasilczyk [Wed, 21 Apr 2021 19:42:08 +0000]
Remove legacy EEPROM fields from FT4232H.
These are not present when flashed with FT_Prog tool.
Tomasz Wasilczyk [Wed, 21 Apr 2021 19:39:52 +0000]
Don't fail reading missing strings.
This happens i.e. when serial number is not set -- in such case the
string is missing from string descriptor list.
Tomasz Wasilczyk [Thu, 11 Mar 2021 23:47:06 +0000]
ftdi_eeprom_initdefaults const arguments
This function don't need to modify its arguments.
Tomasz Wasilczyk [Thu, 11 Mar 2021 23:43:12 +0000]
Don't spam stderr when flashing FT232H eeprom
Thomas Jarosch [Mon, 24 Aug 2020 17:45:41 +0000]
examples/bitbang_cbus.c: Check return value of fgets()
Thomas Jarosch [Mon, 24 Aug 2020 17:27:22 +0000]
Fix building unit tests without FTDIPP
Needed to run the baudrate unit tests.
Probably another fallout from:
****************************
commit
0209a3633dc877a577af07d883eb5059e22f6a91
cmake: do not check for g++ when FTDIPP is disabled
****************************
Uwe Bonnes [Fri, 31 Jul 2020 09:00:50 +0000]
Fix GCC fall-through warnings.
When compiling ftdi.c in the context of blackmagic debug, more enabled
warnings and -Werror made me to add appended fixes.
Fixing the signedness of baudrate would probably be more logically by
making all baudrate and clk unsigned, but would need more code changes.
I decided to make clk signed as a fix with minimal
impact. The FTDI chips are far away from 2 GHz clock rate...
Xiaofan Chen [Fri, 31 Jul 2020 02:25:45 +0000]
Ignore purge_test example build for MinGW
termios.h is not available on MinGW.
Fabrice Fontaine [Fri, 17 Jul 2020 19:25:03 +0000]
CMakeLists.txt: fix paths when FTDIPP is set
Use the same project name (i.e. libftdi1 and not libftdipp1) when FTDIPP
is enabled as suggested by Aurelien Jarno in
http://developer.intra2net.com/mailarchive/html/libftdi/2020/msg00044.html
Without this change, the libftdi1.pc config file defines the include
path as /usr/local/include/libftdipp1 while the ftdi.h file is actually
installed in /usr/local/include/libftdi1
This is an issue for example for libsigrok which will fail on:
In file included from src/hardware/asix-sigma/protocol.c:27:
src/hardware/asix-sigma/protocol.h:28:10: fatal error: ftdi.h: No such file or directory
28 | #include <ftdi.h>
| ^~~~~~~~
Fixes:
- http://autobuild.buildroot.org/results/
1427f44e36752c337791597fab47a1889552a2fe
Signed-off-by: Fabrice Fontaine <fontaine.fabrice@gmail.com>
Thomas Jarosch [Tue, 7 Jul 2020 19:32:55 +0000]
Release libftdi 1.5 final
Thomas Jarosch [Tue, 9 Jun 2020 21:19:34 +0000]
Merge branch 'prepare-release'
Thomas Jarosch [Tue, 9 Jun 2020 20:39:33 +0000]
Set version number to 1.5rc1
Thomas Jarosch [Tue, 9 Jun 2020 20:28:50 +0000]
Prepare changelog for 1.5rc1
Thomas Jarosch [Tue, 9 Jun 2020 20:54:37 +0000]
Increase copyright year to 2020
Thomas Jarosch [Tue, 9 Jun 2020 20:36:57 +0000]
Update AUTHORS file
Thomas Jarosch [Tue, 9 Jun 2020 20:43:59 +0000]
Update release checklist
Thomas Jarosch [Tue, 9 Jun 2020 21:11:04 +0000]
Remove outdated rpm .spec file template
The distro source packages are much more useful than this.
Grepped source tree for ".spec" and "rpm",
hopefully I caught all occurrences.
Diego Elio Pettenò [Tue, 2 Jun 2020 09:06:33 +0000]
Update Autotools Mythbuster link to avoid multiple redirects.
While I'm hoping to maintain the chain of redirects for the foreseeable
future, it makes sense to point at the final URL now that Autotools
Mythuster has had its domain for a long while.
Signed-off-by: Diego Elio Pettenò <flameeyes@flameeyes.com>
Thomas Jarosch [Thu, 14 May 2020 21:00:24 +0000]
ftdi_eeprom_build(): Fix inverted handling of VCP driver field for TYPE_R chips
From doc/EEPROM-structure:
On TYPE_R 00.3 set means D2XX, on other devices VCP
-> invert the field for TYPE_R so "eeprom->channel_a_driver"
set to 1 always means VCP driver for all chips.
Thomas Jarosch [Thu, 14 May 2020 20:38:37 +0000]
eeprom handling: Unify handling of boolean bit flags
Boolean bit flags should be zero or one in memory,
not the actual eeprom value FTDI assigned to them.
Example of the code change in ftdi_eeprom_build():
ftdi_eeprom_build():
- if (eeprom->suspend_dbus7 == SUSPEND_DBUS7_BIT)
+ if (eeprom->suspend_dbus7)
The changes should not break any backward compatibility
in ftdi_build_eeprom().
The function ftdi_decode_eeprom() will return slightly different values
for boolean fields, but callers should test using "if (flag_enabled)"
instead of "if (flag_enabled == specific_value)" anyway.
Test done: Decoded eeprom of a FT232 serial cable stayed the same.
$ diff -u dump-old.txt dump-new.txt
-> no change
Roman Lapin [Sat, 9 May 2020 20:13:58 +0000]
ftdi_eeprom: Add device release number support
In process of firmware repairing of some FTDI-chip based device I have
found that I can't get binary identical EEPROM dumps from properly
working device and ftdi_eeprom --build_eeprom command because of
ftdi_eeprom does not support changing of device release number (byte
with addr 0x06).
[TJ: Keep old behavior if release_number is not set in the config file]
Yegor Yefremov [Mon, 13 Apr 2020 21:09:27 +0000]
Ignore patch files
Thomas Jarosch [Mon, 13 Apr 2020 21:07:14 +0000]
ftdi_eeprom: Minor cosmetic fix on usage screen
Holger Mößinger [Wed, 5 Feb 2020 12:11:04 +0000]
ftdi_eeprom: Add more detailed example.conf file
Holger Mößinger [Fri, 7 Feb 2020 13:38:46 +0000]
eeprom: Fix handling of high_current_drive parameter
Additionally fix it up for type R chips.
Holger Mößinger [Wed, 5 Feb 2020 13:57:41 +0000]
eeprom: Add channel_a_driver support for type xxR chips
Holger Mößinger [Wed, 5 Feb 2020 12:10:25 +0000]
Make ftdi_eeprom more verbose by adding verbose option
[TJ: Replaced tabs with spaces]
Yegor Yefremov [Thu, 6 Feb 2020 13:37:09 +0000]
Ignore tags file
Signed-off-by: Yegor Yefremov <yegorslists@googlemail.com>
Fabrice Fontaine [Tue, 21 Jan 2020 22:05:27 +0000]
cmake: do not check for g++ when FTDIPP is disabled
Signed-off-by: Samuel Martin <s.martin49@gmail.com>
[Retrieved (and slightly updated) from:
https://git.buildroot.net/buildroot/tree/package/libftdi1/0003-cmake-do-not-check-for-g-when-FTDIPP-is-disabled.patch]
Signed-off-by: Fabrice Fontaine <fontaine.fabrice@gmail.com>
Yegor Yefremov [Thu, 16 Jan 2020 07:42:21 +0000]
Fix typos
Signed-off-by: Yegor Yefremov <yegorslists@googlemail.com>
Dan Dedrick [Tue, 7 Jan 2020 16:18:30 +0000]
libftdi: pass eeprom strings as const
These strings are not modified and it doesn't make sense that they ever
would be. Setting these as const ensures that they will not be modified
and allows const strings to be passed in without special manipulation.
Specifically a call like the following failed to compile:
ftdi_eeprom_set_strings(&ftdi, "foo", "bar", "abc123");
The error will look something like this:
error: ISO C++ forbids converting a string constant to ‘char*’
[-Werror=write-strings]
Yegor Yefremov [Mon, 13 Jan 2020 14:08:44 +0000]
CMake: reorder Python detection
PythonInterp should be detected before PythonLibs otherwise
CMake exits with the following message:
Could NOT find PythonLibs (missing: PYTHON_LIBRARIES PYTHON_INCLUDE_DIRS)
Signed-off-by: Yegor Yefremov <yegorslists@googlemail.com>
Thomas Jarosch [Tue, 1 Oct 2019 14:17:09 +0000]
Fix build dependency if swig wrapper and documentation is both switched on
Seems to affect new cmake / swig versions only.
Reported by Richard Shaw <hobbes1069@gmail.com> during Fedora Rawhide build:
Ok, so far looks like I was right, you have to set the dependency on
ftdi1_doc.i before calling swig_add_library or it doesn't get it right:
[ 11%] Swig compile ftdi1.i for python
/usr/bin/doxygen /builddir/build/BUILD/libftdi-f49ee0d/build-py2/Doxyfile
cd /builddir/build/BUILD/libftdi-f49ee0d/build-py2/python && /usr/bin/cmake
-E make_directory /builddir/build/BUILD/libftdi-f49ee0d/build-py2/python
/builddir/build/BUILD/libftdi-f49ee0d/build-py2/python/CMakeFiles/_ftdi1.dir
/usr/bin/doxygen
/builddir/build/BUILD/libftdi-f49ee0d/build-py2/Doxyfile.xml
cd /builddir/build/BUILD/libftdi-f49ee0d/build-py2/python && /usr/bin/cmake
-E touch
/builddir/build/BUILD/libftdi-f49ee0d/build-py2/python/CMakeFiles/_ftdi1.dir/ftdi1PYTHON.stamp
cd /builddir/build/BUILD/libftdi-f49ee0d/build-py2/python && /usr/bin/cmake
-E env SWIG_LIB=/usr/share/swig/3.0.12 /usr/bin/swig -python -DDOXYGEN=YES
-outdir /builddir/build/BUILD/libftdi-f49ee0d/build-py2/python
-I/builddir/build/BUILD/libftdi-f49ee0d/src -I/usr/include/libusb-1.0
-I/usr/include/python2.7 -o
/builddir/build/BUILD/libftdi-f49ee0d/build-py2/python/CMakeFiles/_ftdi1.dir/ftdi1PYTHON_wrap.c
/builddir/build/BUILD/libftdi-f49ee0d/python/ftdi1.i
/builddir/build/BUILD/libftdi-f49ee0d/python/ftdi1.i:7: Error: Unable to
find 'ftdi1_doc.i'
make[2]: *** [python/CMakeFiles/ftdi1_swig_compilation.dir/build.make:67:
python/CMakeFiles/_ftdi1.dir/ftdi1PYTHON.stamp] Error 1
Frank Dana [Thu, 26 Sep 2019 18:57:25 +0000]
Simplify CMake policy guard
Signed-off-by: FeRD (Frank Dana) <ferdnyc@gmail.com>
Richard Shaw [Mon, 18 Feb 2019 18:42:30 +0000]
CMake/Swig updates for python wrapper
libftdi was FTBFS[1] in Fedora Rawhide after a GCC 9.0.X mass rebuild but
the problem turned out to be behavior changes[2] in CMake with regards to
SWIG[3].
The UseSWIG module in cmake 3.13+ will now use standard target names and not
generate the SWIG_MODULE_<target>_REAL_NAME variable.
Replace the deprecated swig_add_module with swig_add_library.
[Refinements from Intra2net:
- Use swig_add_library for cmake 3.8+ only
- Add cmake version guard for the new CMP0078 policy
- Use "NOT CMAKE_VERSION VERSION_LESS" construct to
support cmake versions below 3.7.0]
[1] https://bugzilla.redhat.com/show_bug.cgi?id=1675275
[2] https://cmake.org/cmake/help/latest/policy/CMP0078.html?highlight=cmp0078
[3] https://cmake.org/cmake/help/latest/module/UseSWIG.html
Yegor Yefremov [Fri, 7 Jun 2019 13:43:55 +0000]
Add SPDX identifiers to the core library and ftdi_eeprom tool
Signed-off-by: Yegor Yefremov <yegorslists@googlemail.com>
Yegor Yefremov [Thu, 22 Aug 2019 13:47:06 +0000]
swig: improve indentation
Fixes -Wmisleading-indentation warnings.
Claudio Lanconelli [Tue, 1 Oct 2019 13:11:54 +0000]
ftdi_eeprom: Add support for group0 drive levels on x232H chips
Yegor Yefremov [Tue, 6 Aug 2019 11:42:06 +0000]
swig: python: fix unresolved symbol issue
GCC later than 5.x produce _fdti1.so file with an undefined
symbol str2charp_size due to C99 inline semantics change. So
remove this keyword.
Signed-off-by: Yegor Yefremov <yegorslists@googlemail.com>
Fabrice Fontaine [Sat, 19 Jan 2019 20:42:01 +0000]
python/CMakeLists.txt: fix build with cmake < 3.7
VERSION_GREATER_EQUAL has been added only in cmake 3.7:
https://cmake.org/cmake/help/v3.7/release/3.7.html
So replace this statement by NOT CMAKE_VERSION VERSION_LESS
Fixes:
- http://autobuild.buildroot.org/results/
1091872e2b77d789e361d1ddefd235c738933c55
Signed-off-by: Fabrice Fontaine <fontaine.fabrice@gmail.com>
Thomas Jarosch [Fri, 18 Jan 2019 14:51:35 +0000]
Remove more example programs during rpm build
Eric Schott [Fri, 4 Jan 2019 00:55:59 +0000]
Add program to test flush (purge) functionality.
Eric Schott [Fri, 4 Jan 2019 00:54:40 +0000]
Improve Doxygen output for #defines in header files.
Eric Schott [Fri, 4 Jan 2019 00:52:02 +0000]
Implement tc[io]flush methods & deprecate broken purge_buffers methods.
The definitions for the SIO_RESET_PURGE_RX or SIO_RESET_PURGE_TX values are
with respect to the FTDI chip, not the CPU. That is, when the FTDI chip
receives a USB control transfer request with the command SIO_RESET_PURGE_RX,
the FTDI chip empties the FIFO containing data received from the CPU awaiting
transfer out the serial port to the connected serial device (e.g., a modem).
Likewise, upon reception of the SIO_RESET_PURGE_TX command, the FTDI chip
empties the FIFO of data received from the attached serial device destined to
be transmitted to the CPU.
Unfortunately the coding of the previous releases of libfti assumed these
commands had the opposite effect. This resulted in the function
ftdi_usb_purge_tx_buffer clearing data received from the attached serial
device. Similarly, the function ftdi_usb_purge_rx_buffer cleared the FTDI
FIFO containing data to be transmitted to the attached serial device. More
seriously, this latter function clear the libftid's internal buffer of data
received from the serial device, destined to the application program.
This patch adds the ftdi_tciflush, ftdi_tcoflush, and ftdi_tcioflush functions
which emulate the Linux serial port tcflush(3) function.
Jordan Rupprecht [Thu, 20 Sep 2018 18:05:56 +0000]
Use a separate Py_ssize_t var instead of casting an int pointer as a Py_ssize_t pointer.
Casting is fine if Py_ssize_t == int, but not when Py_ssize_t == long.
Signed-off-by: Jordan Rupprecht <rupprecht@google.com>
Eneas U de Queiroz [Tue, 31 Jul 2018 20:44:57 +0000]
CMake: use find_package (SWIG) for cmake >= 3.0.0
There's a workaround for a bug (fixed in cmake 3.0.0) that does not
work in CMake 3.12. Only use the workaround with cmake < 3.0.0.
Signed-off-by: Eneas U de Queiroz <cote2004-github@yahoo.com>
Uwe Bonnes [Fri, 1 Jun 2018 09:26:52 +0000]
Improve new async read/write example
Uwe Bonnes [Thu, 31 May 2018 13:22:05 +0000]
examples/async.c : Short example using async read/write
Thomas Jarosch [Wed, 28 Feb 2018 20:14:49 +0000]
Bring back ftdi_eeprom by default
Yegor Yefremov [Wed, 3 Jan 2018 12:46:23 +0000]
CMake: move options to a dedicated file
Also disable all options having extra dependencies.
If an option is selected makes its dependencies as REQUIRED.
Signed-off-by: Yegor Yefremov <yegorslists@googlemail.com>
Thomas Jarosch [Tue, 6 Feb 2018 20:18:17 +0000]
Add Robby McKilliam to AUTHORS
Robby McKilliam [Wed, 13 Dec 2017 22:54:35 +0000]
Add kernel driver auto attach/detach
Introduce new detach mode:
AUTO_DETACH_REATACH_SIO_MODULE
Tested on linux (Ubuntu 16.04 and Raspbian Stretch) with an FT230.
Thomas Jarosch [Fri, 27 Oct 2017 15:15:27 +0000]
Access ftdi->max_packet_size after checking ftdi context != NULL
Reported via email by Mr. Alex Tvema.
Thomas Jarosch [Mon, 25 Sep 2017 07:17:25 +0000]
Start ChangeLog for next version
Thomas Jarosch [Mon, 25 Sep 2017 07:16:13 +0000]
Add Pawel Jewstafjew to AUTHORS
Pawel Jewstafjew [Wed, 20 Sep 2017 18:38:54 +0000]
fix support for XON/XOFF flow control
Thomas Jarosch [Tue, 19 Sep 2017 14:19:37 +0000]
Fix ftdi_stream.c compilation on WIN32
Rolf Fiedler [Fri, 25 Aug 2017 09:33:35 +0000]
Fix small memleak when ftdi_init() fails
If libusb open fails (e.g. permission problem),
then ftdi_init() leaks the memory for the eeprom data.
Thomas Jarosch [Mon, 7 Aug 2017 14:52:39 +0000]
Increase version to 1.4 final
Matthias Janke [Fri, 21 Jul 2017 14:10:15 +0000]
follow-up on c++ open() fix. fixes logic in open() with long argument list.
Thomas Jarosch [Wed, 26 Jul 2017 15:17:09 +0000]
Sync README with ChangeLog
Thomas Jarosch [Wed, 26 Jul 2017 15:03:29 +0000]
Increase version to 1.4rc2
Thomas Jarosch [Fri, 21 Jul 2017 10:00:58 +0000]
C++ API: Increase .so library version
We have an ABI break in this release (get_strings_and_reopen() changes),
so programs using the C++ wrapper need to be recompiled.
Thomas Jarosch [Fri, 21 Jul 2017 09:58:31 +0000]
Add Eric Schott to AUTHORS, update ChangeLog
Eric Schott [Fri, 21 Jul 2017 09:56:07 +0000]
C++ API: Correct the purge Direction and ModemCtl enumerations definitions
Do not use value of zero for bitmasks. The new enumeration values
are defined so programs compiled with the old ftdi.hpp include
would work as previously (doing nothing when the enumeration
with the zero value was bit-or'ed into the argument).