Commit message (Collapse) | Author | Age | Files | Lines | ||
---|---|---|---|---|---|---|
... | ||||||
* | set DLL API call declarations to static | Pete Batard | 2010-10-26 | 1 | -1/+2 | |
| | | | | | * prevents redefinition warnings when an app is statically linking against SetupAPI.lib in MSVC for instance. | |||||
* | switched MSVC project files from MBCS to Unicode | Pete Batard | 2010-10-26 | 3 | -13/+13 | |
| | | | | | | * MBCS (which is different from UTF-8) only makes sense if supporting Windows 95/98, which we don't * also switched all string functions to use char* | |||||
* | merged latest from official | Pete Batard | 2010-10-18 | 1 | -53/+59 | |
| | | | | 678c242705b29f9cb6cc421b5625a6485ae58047 & 9cd9059bf9baac69a9d6c909f4c1e000592fa260 | |||||
* | fixed potentially non unique return value for htab_hashpbr314 | Pete Batard | 2010-10-14 | 3 | -7/+8 | |
| | | | | * also increased htab size to 1021, from a .h macro | |||||
* | fixed trac #68 (incorrect buffer size for HID reports)pbr313 | Pete Batard | 2010-10-13 | 2 | -12/+7 | |
| | | | | | * this regression was introduced in pbr301 * reported by gorlik | |||||
* | fixed missing malloc check in htab_hash()pbr312 | Pete Batard | 2010-10-12 | 2 | -6/+6 | |
| | | | | | * spotted by Michael Plante * also some cleanup | |||||
* | bumped internal versionpbr311 | Pete Batard | 2010-10-12 | 1 | -1/+1 | |
| | ||||||
* | hash function for session id now uses a proper hash table | Pete Batard | 2010-10-12 | 1 | -9/+169 | |
| | | | | * hash table functions modified from glibc (LGPL) | |||||
* | addressed TODOs from new_enum | Pete Batard | 2010-10-11 | 1 | -87/+40 | |
| | | | | | | * use realloc for unref * pick VID/PID from PCI HCD * cleanup | |||||
* | Windows enumeration overhaul | Pete Batard | 2010-10-11 | 2 | -855/+673 | |
| | | | | | | | * changed device enumeration to rely on session ids (hashed device ID) * all enumeration passes now occur in a single call * also removed the separate HCD list (allows PCI hotplug of HCDs) * additional cleanup | |||||
* | bumped internal versionpbr310 | Pete Batard | 2010-10-06 | 1 | -1/+1 | |
| | ||||||
* | fixed missing $M_LDFLAGS from libusb/Makefile.am | Pete Batard | 2010-10-06 | 1 | -1/+1 | |
| | | | | * problem was introduced by recent -version-info patch from official | |||||
* | bumped internal versionpbr309 | Pete Batard | 2010-10-05 | 1 | -1/+1 | |
| | ||||||
* | fixed missed index -> _index in poll_windows.c debug messages | Pete Batard | 2010-10-05 | 1 | -2/+2 | |
| | | | | * follow up from cygwin shadow warning fix | |||||
* | bumped internal versionpbr308 | Pete Batard | 2010-10-05 | 1 | -1/+1 | |
| | ||||||
* | fixed index shadow warnings from cygwin (from official) | Pete Batard | 2010-10-05 | 2 | -86/+86 | |
| | ||||||
* | small fixup to be in check with proposed official | Pete Batard | 2010-10-05 | 2 | -2/+2 | |
| | ||||||
* | Linux: fix build failure when CLOCK_MONOTONIC does not exist | Mike Frysinger | 2010-10-05 | 1 | -5/+6 | |
| | | | | This is not available on all platforms. | |||||
* | Update libtool version info | Daniel Drake | 2010-10-05 | 1 | -0/+1 | |
| | | | | | | | With input from various people on the mailing list, update the libtool versioning info and start to update this on every release. The next libusb release will not need a change here. All following ones will. | |||||
* | bumped internal versionpbr307 | Pete Batard | 2010-10-04 | 1 | -1/+1 | |
| | ||||||
* | fixed erroneous checks on SetupDi calls | Pete Batard | 2010-10-04 | 1 | -2/+2 | |
| | ||||||
* | bumped internal versionpbr306 | Pete Batard | 2010-09-21 | 1 | -1/+1 | |
| | ||||||
* | force the handling of version stringification in the .rc | Pete Batard | 2010-09-21 | 3 | -2/+4 | |
| | | | | * this removes the version string from libusb_version.h | |||||
* | fixed missing def entries for function calls with no args | Pete Batard | 2010-09-21 | 1 | -0/+57 | |
| | | | | * this is needed for libusb_getversion() in MinGW32 | |||||
* | bumped internal versionpbr305 | Pete Batard | 2010-09-21 | 1 | -2/+2 | |
| | ||||||
* | minor .rc improvements | Pete Batard | 2010-09-21 | 1 | -10/+4 | |
| | | | | | * remove no longer needed debug filename * add NUL terminators & official URL | |||||
* | handle versioning through libusb_version.h | Pete Batard | 2010-09-21 | 6 | -937/+66 | |
| | | | | | | * libusb_version.h.in is now used to generate libusb_version.h * removes the need for libusb-1.0.rc.in and libusbi.h.in * new LIBUSB_PACKAGE_VERSION added as .rc can't stringify macros | |||||
* | fixed previous patch and updated .def for new callspbr304 | Pete Batard | 2010-09-21 | 2 | -4/+22 | |
| | ||||||
* | Fix a race condition | Graeme Gill | 2010-09-21 | 3 | -18/+42 | |
| | | | | See http://sourceforge.net/mailarchive/message.php?msg_name=4BD920C5.3090909%40argyllcms.com | |||||
* | Add a missing argument to libusb_wait_for_event | Ludovic Rousseau | 2010-09-21 | 1 | -1/+1 | |
| | | | | | | | In the libusb_wait_for_event() sample code for section "Letting other threads do the work for you" libusb_wait_for_event() was used without the struct timeval *tv parameter. The sample code was invalid and could not be compiled. | |||||
* | Linux: Fix endian handling for usbfs/sysfs | Hector Martin | 2010-09-21 | 1 | -2/+2 | |
| | | | | | | | | usbfs endian-swaps, sysfs doesn't, not the other way around. This fixes configurator descriptors in big-endian systems when using sysfs. usbfs is untested. Signed-off-by: Hector Martin <hector@marcansoft.com> | |||||
* | fixed .def generation broken by previous commitpbr303 | Pete Batard | 2010-09-20 | 1 | -63/+63 | |
| | ||||||
* | merged latest official changes | Pete Batard | 2010-09-20 | 7 | -92/+180 | |
| | | | | * up to e65f69a7ef3de357c867ddaac9598efe407078b6 | |||||
* | replaced LIBUSB_API with LIBUSB_CALL | Pete Batard | 2010-09-20 | 5 | -118/+118 | |
| | ||||||
* | updated libusb-1.0.def for libusb_getversion() | Pete Batard | 2010-09-20 | 1 | -0/+9 | |
| | ||||||
* | added libusb_getversion() | Pete Batard | 2010-09-20 | 4 | -0/+895 | |
| | | | | | * new struct libusb_version has also been defined * libusbi.h.in is now being used to fill in the version from configure.ac | |||||
* | added libusb version nano | Pete Batard | 2010-09-20 | 1 | -2/+2 | |
| | ||||||
* | fixed missing winioctl.h header for cygwinpbr301 | Pete Batard | 2010-09-06 | 1 | -0/+1 | |
| | ||||||
* | use IOCTLs for HID reports (input, output, feature) | Pete Batard | 2010-09-06 | 2 | -153/+59 | |
| | | | | | | * fixes feature reports not providing actual read size (reported by Axel Rohde) * removes the USE_HIDD_FOR_REPORTS macro * IOCTL usage inspired from HIDAPI by Alan Ott | |||||
* | workaround for unavailable _strdup on MinGW32pbr300 | Pete Batard | 2010-08-21 | 1 | -1/+6 | |
| | ||||||
* | fixed possible buffer overflow in windows_error_str()pbr299 | Pete Batard | 2010-08-19 | 1 | -4/+9 | |
| | | | | * also added removal of CR/LF | |||||
* | added full support for filter drivers | Pete Batard | 2010-08-19 | 2 | -35/+66 | |
| | | | | | * precedence is driver > upper filter > lower filter and defined in struct driver_lookup lookup from set_device_paths() | |||||
* | added limited support for filter driverspbr298 | Pete Batard | 2010-08-18 | 1 | -45/+61 | |
| | | | | | * first lower filter driver only * driver currently has precedence over filter driver | |||||
* | removed root from list of DDK include pathspbr297 | Pete Batard | 2010-08-17 | 1 | -1/+1 | |
| | | | | * fixes an issue where DDK would use the config.h from MinGW/cygwin in root instead of using msvc's | |||||
* | fixed errcode shadow warning in MinGW-w64 (windows backend)pbr296 | Pete Batard | 2010-08-17 | 1 | -8/+8 | |
| | ||||||
* | fixed errcode shadow warning in MinGW-w64 (core) | Pete Batard | 2010-08-17 | 1 | -2/+2 | |
| | ||||||
* | restrict scope of previous workaround and return error when possible | Pete Batard | 2010-08-17 | 3 | -9/+18 | |
| | ||||||
* | only apply previous workaround for 32 bit buildspbr295 | Pete Batard | 2010-08-16 | 3 | -0/+17 | |
| | ||||||
* | added Workaround for MinGW-w64 multilib bug | Pete Batard | 2010-08-15 | 3 | -5/+38 | |
| | | | | | | * current MinGW-w64 32 bit headers (winbase.h) are missing the WINAPI qualifier on Interlocked### calls * this results in missing decorations on symbols and failed linking as a result * this workaround hooks into kernel32.dll for these function calls to alleviate the issue | |||||
* | remove unneeded sanity check in cache_config_descriptorspbr294 | Pete Batard | 2010-08-10 | 2 | -8/+3 | |
| | | | | | | | * issue reported by Benjamin Dobell (with additional input from Tim Roberts) * unlike interface numbers, bConfigurationValue are not required to be in order * also produces actual bConfigurationValue in debug output * small whitespace fix in windows_usb.h |