On Sun, Apr 22, 2012 at 12:07 AM, Tormod Volden wrote: > On Sat, Apr 21, 2012 at 11:01 PM, Stefan Schmidt wrote: >> >>> dfu-suffix: Check for availability of truncate() function >> >> You happen to know which platforms are missing this? > > No, but I would guess everything non-posix. I believe, unless you are > using cygwin, the posix layer on Windows is not so complete. I know > that when compiling with MinGW, truncate() is not available. I just > found this http://stackoverflow.com/questions/584639/truncate-file so > it looks like we can use SetEndOfFile() on Windows instead. I will > look into that another day.
I found ftruncate() on MinGW, so I changed to use that (with help of fileno() to avoid reopening the file this got much cleaner too). For non-MinGW Windows, it seems like _chsize_s is an equivalent (forget about SetEndOfFile() which acts on Windows file handles). Since I can not test that variant, and MinGW works fine, I am not including it. > >> >>> Check for availability of getpagesize() >>> Replace usleep() and sleep() by milli_sleep() macro >> >> Is HAVE_WINDOWS_H really the preferred way to check for windows? Satz >> used ENV_WINDOWS (which also looked odd to me). Could be that it >> really depends on what you are using to create the executable. MinGW >> vs. Windows compiler. Anyway, as you have tested this under Windows >> and I also have no problems with it under Linux thats fine. > > http://msdn.microsoft.com/en-us/library/windows/desktop/ms686298%28v=vs.85%29.aspx > says that you should include windows.h to use the Sleep() function. So > maybe not the preferred way to check for Windows but seems a good > guess for checking for Sleep(). Since we are using HAVE_WINDOWS_H it implies auto-tools is being used, so it could seem like checking for Sleep() would be the proper way. However that does not work when cross-compiling on MinGW. Similar to the getpagesize() issue this might be a bug in auto-tools or I don't know how to use it properly. >> The squashing of the bug fix in the previous change is the only thing >> that would block me from pulling this in. If you could fix this up I >> would happily pull in an updated pull request. >> > > Done! Done again, rebased with the ftruncate fix. Before repulling, do a fetch so you can git diff the small difference. > >> From my side we would be ready for a 0.6 release afterwards. :) > > We could maybe wait to see if I can get suffix truncation work on > Windows easily, that would save some documentation and make dfu-util > fully consistent across architectures. Should be fine now. Tormod _______________________________________________ devel mailing list devel@lists.openmoko.org https://lists.openmoko.org/mailman/listinfo/devel
