Hi Thomas, You are right. I also think a redesign of the cpp-wrapper would be the way to go. I was curiouse about the get_string() function and their behaviour of closing the device after retreiving the device-strings. So far I guess your points are a good way.
I'll take a look at some points but I suspect I won't be able to implement any big stuff in the next 2 month (have a little crunch time at work). But I'll stay tuned on this. Greets pit ----- Original Nachricht ---- Von: Thomas Jarosch <thomas.jaro...@intra2net.com> An: libftdi@developer.intra2net.com Datum: 05.02.2011 18:34 Betreff: Re: Bug in libftdipp Context::open() > Hello pit, > > On 01/20/2011 01:27 PM, pitpom...@arcor.de wrote: > > I found a bug in the cpp-wrapper for the ftdi-lib that concerns nearly all > open() functions of the Context-Class. All open-functions except the one > with the usb_device parameter return with -1 (and error string "all fine") > even if opening the device should be posssible (the ftdi-c version can open > the same device). The reason for that behavior is, that they all use for the > first connect to the device the correct ftdi-c-function concerning to the > given parameters (open(vendor,pid) uses ftdi_usb_open(ftdi, vendor, > product), open(description) uses ftdi_usb_open_string(ftdi, > description.c_str()) and so on) but after the first connect they all use the > ftdi_usb_get_strings function that needs as one parameter the usb_device. > this has not been set before when using these special open-functions so it > returns with an error and the open-function failes. > > I have already implemented a fix in my sources by adding to further > functions in the ftdi-c-lib that returns by reference the usb_device ande > changed the concerning calls in the cpp-wrapper. That all works fine for me > now. > > But is a problem in the ftdi-cpp-wrapper worth a change in the > ftdi-c-lib? > > I will try to commit it if possible today, otherwise next week. > > Thanks for the report and your patch! > > Looking at the C++ wrapper three years later, > I don't like the fact that I keeps an own "libusb_device *dev" > pointer -and- a pointer to the ftdi_context. > > I think it can be improved without bloating the open() > function of the C API. Something like this: > > - Remove the libusb_device pointer in the C++ wrapper > > - Provide a function in the C API to get the strings without > closing the device. Thinking about it some more, the C API > could do this by default, "wasting" some bytes for an > already open device is a non-issue. > > - Kill the get_strings_and_reopen() function > > - Implement is_open() in the C API and kill the > "open" flag, too. > > What do you think? > > Cheers, > Thomas > > -- > libftdi - see http://www.intra2net.com/en/developer/libftdi for details. > To unsubscribe send a mail to libftdi+unsubscr...@developer.intra2net.com > > > -- libftdi - see http://www.intra2net.com/en/developer/libftdi for details. To unsubscribe send a mail to libftdi+unsubscr...@developer.intra2net.com
