在 2011年6月15日 下午7:56，Daniel Stenberg <dan...@haxx.se>写道：
> On Wed, 15 Jun 2011, zl liu wrote: > > yes,1.2.8. the test is running on windows 7 . >> > > Ok. We have received bug reports previously on Windows that we haven't been > able to repeat elsewhere and afaik nobody has fixed. > > > after i modify the misc.c::445 like : fprintf(stderr, "%s, %d\n", buffer, >> GetLastError()); when -7, the system errno is 10035. >> > > Can you see if you can figure out how that can happen? > -7 haven't return to the caller. just the internal errno. > > 10035 is WSAEWOULDBLOCK from what I understand, and > src/misc.c:_libssh2_send() is written so that it converts such an error to > return -EAGAIN. The code in src/transport.c that use _libssh2_send() then > checks the return code for -EAGAIN and should not return anoter error than > LIBSSH2_ERROR_EAGAIN (-37) for that case. > > I think we should focus on the first error that happens as I suspect the > ones following may be results of the first problem. after debug the libssh2,i get the following information: the function call stack: > libssh2.dll!send_existing(_LIBSSH2_SESSION * session=0x017d0068, const unsigned char * data=0x014f1e00, unsigned int data_len=9, long * ret=0x00035410) 行605 C libssh2.dll!_libssh2_transport_send(_LIBSSH2_SESSION * session=0x017d0068, const unsigned char * data=0x014f1e00, unsigned int data_len=9, const unsigned char * data2=0x00000000, unsigned int data2_len=0) 行696 + 0x15 字节 C libssh2.dll!_libssh2_channel_receive_window_adjust(_LIBSSH2_CHANNEL * channel=0x014f1c10, unsigned int adjustment=3932160, unsigned char force=0, unsigned int * store=0x00000000) 行1581 + 0x1b 字节 C libssh2.dll!_libssh2_channel_read(_LIBSSH2_CHANNEL * channel=0x014f1c10, int stream_id=0, char * buf=0x017dda08, unsigned int buflen=4) 行1882 + 0x12 字节 C libssh2.dll!sftp_packet_read(_LIBSSH2_SFTP * sftp=0x017dd9e8) 行193 + 0x25 字节 C libssh2.dll!sftp_packet_requirev(_LIBSSH2_SFTP * sftp=0x017dd9e8, int num_valid_responses=2, const unsigned char * valid_responses=0x1002c03c, unsigned int request_id=1, unsigned char * * data=0x000359ec, unsigned int * data_len=0x000359e0) 行404 + 0x9 字节 C libssh2.dll!sftp_read(_LIBSSH2_SFTP_HANDLE * handle=0x017dd3c0, char * buffer=0x00035e68, unsigned int buffer_size=1024000) 行1204 + 0x25 字节 C libssh2.dll!libssh2_sftp_read(_LIBSSH2_SFTP_HANDLE * hnd=0x017dd3c0, char * buffer=0x00035e68, unsigned int buffer_maxlen=1024000) 行1311 + 0x11 字节 C the value in this point: data: 0x01391E00 5d 00 00 00 00 00 3c 00 ].....<. 0x01391E08 00 00 00 00 0b 00 00 00 ........ p->odata: 0x01391E18 5e 00 00 00 00 00 00 00 ^....... 0x01391E20 1d 00 00 00 00 00 00 00 ........ data_len = p->olen = 9 i haven't send anything, it is a internal mistake? > another question , why not define the MACRO :LIBSSH2DEBUG in the debug >>>> mode. when i want to enable trace，i must define the macro LIBSSH2DEBUG >>>> manually and compile it again. >>>> >>>> That's what we have "./configure --enable-debug" for! >>> >>> i compile on win32 .it can not execute ./configure. >> > > Ok. But how is defining LIBSSH2DEBUG somewhere not roughly the equivalent? > What can we do to make it even simpler? > > -- > thanks! > > / daniel.haxx.se > _______________________________________________ > libssh2-devel http://cool.haxx.se/cgi-bin/mailman/listinfo/libssh2-devel > >
_______________________________________________ libssh2-devel http://cool.haxx.se/cgi-bin/mailman/listinfo/libssh2-devel
