wine-devel

wine-devel@winehq.org

2 participants
84469 discussions

dlls/msvcrt/file.c broken on FreeBSD 4.5
by Gerald Pfeifer 27 Apr '02

27 Apr '02

The following patch to dlls/msvcrt/file.c revision 1.23 date: 2002/04/26 19:05:16; author: julliard; state: Exp; lines: +4 -0 Better support for configure detection of missing types, added check for ssize_t. Removed a couple of no longer used portability functions. Various portability fixes in port.h. breaks FreeBSD 4.5: /usr/bin/gcc -c -I. -I. -I../../include -I../../include -g -O2 -Wall -mpreferred-stack-boundary=2 -fPIC -D__WINE__ -DUSE_MSVCRT_PREFIX -D_MT -D_REENTRANT -I/usr/X11R6/include -o file.o file.c In file included from file.c:43: ../../include/msvcrt/sys/stat.h:34: syntax error before `.' ../../include/msvcrt/sys/stat.h:35: syntax error before `.' ../../include/msvcrt/sys/stat.h:36: syntax error before `.' ../../include/msvcrt/sys/stat.h:48: syntax error before `.' ../../include/msvcrt/sys/stat.h:49: syntax error before `.' ../../include/msvcrt/sys/stat.h:50: syntax error before `.' file.c: In function `_fstat': file.c:693: structure has no member named `st_atimespec' file.c:695: structure has no member named `st_mtimespec' file.c:695: structure has no member named `st_ctimespec' file.c: In function `_stat': file.c:1181: structure has no member named `st_atimespec' file.c:1183: structure has no member named `st_mtimespec' file.c:1183: structure has no member named `st_ctimespec' file.c:1185: structure has no member named `st_atimespec' file.c:1185: structure has no member named `st_mtimespec' file.c:1185: structure has no member named `st_ctimespec' file.c: In function `_wstat': file.c:1242: structure has no member named `st_atimespec' file.c:1244: structure has no member named `st_mtimespec' file.c:1244: structure has no member named `st_ctimespec' file.c:1246: structure has no member named `st_atimespec' file.c:1246: structure has no member named `st_mtimespec' file.c:1246: structure has no member named `st_ctimespec' gmake[2]: *** [file.o] Error 1 gmake[2]: Leaving directory `/.amd_mnt/vexpert/files8/test/wine/dlls/msvcrt' gmake[1]: *** [msvcrt/msvcrt.dll.so] Error 2 Gerald

1 0

./tests/sock.c: No such file or directory
by Marcelo Welter 27 Apr '02

27 Apr '02

helo, ./configure make depend ./tests/sock.c: No such file or directory and now...? Marcelo

1 0

Correction for WWN 121
by Speeddymon 26 Apr '02

26 Apr '02

Hello, just had an update for WWN 121... Near the bottom it says that Alexandre rejected a patch for the ldconfig problem (the one i submitted) that puts $libdir/wine into ld.so.conf because users need to set $WINEDLLPATH... the corrections are this: users need to set $LD_LIBRARY PATH and The latest wineinstall patch I submitted removes the $libdir/wine addition and instead always runs ldconfig (just in case of modules being renamed) if anyone wants to append that to 121 or add it into the 122 __________________________________________________ Do You Yahoo!? Yahoo! Games - play chess, backgammon, pool and more http://games.yahoo.com/

2 1

Re: Correction for WWN 121
by vinn 26 Apr '02

26 Apr '02

> Hello, just had an update for WWN 121... Near the > bottom it says that Alexandre rejected a patch for the Dustin - I'll make the change and get it into 121. Sorry about that, I summarized that thread about halfway into it. -brian.

1 0

Name change
by Dustin Navea 26 Apr '02

26 Apr '02

Hello, just noticed that my name in WWN121 was posted as Speeddymon, I forgot that I had set it to that so many many years ago... This is my real name and this is what I will be posting under from now on... :) __________________________________________________ Do You Yahoo!? Yahoo! Games - play chess, backgammon, pool and more http://games.yahoo.com/

1 0

Re: Patch acceptance notification?
by Francois Gouget 26 Apr '02

26 Apr '02

On Fri, 26 Apr 2002, Speeddymon wrote: > If we submit a patch, and it is accepted (and > commited) it would be nice to know if any of my > patches will be in the next snapshot... If it is committed to CVS, it will be in the next Wine snapshot. This is automatic and systematic. The only thing that could 'prevent it' is if a huge bug was found between the commit and the snapshot release and the patch ended up being reverted, but that is pretty rare and cannot be forseen. -- Francois Gouget fgouget(a)free.fr http://fgouget.free.fr/ May your Tongue stick to the Roof of your Mouth with the Force of a Thousand Caramels.

2 1

winsock2 unit test update
by Martin Wilck 26 Apr '02

26 Apr '02

A patch for the winsock 2 unit test posted yesterday. It now has an overlapped client. If you look at the code, it does pretty nasty things to wine, and still works :-) BUGS: probably doesn't compile under Windows, I'm using some gcc extensions. You just get too used to them ... If somebody was able to port these to Windows and test the outcome I would be really happy. Martin diff -ruNX ignore CVS/wine/dlls/winsock/tests/sock.c TMP/wine/dlls/winsock/tests/sock.c --- CVS/wine/dlls/winsock/tests/sock.c Fri Apr 26 20:58:26 2002 +++ TMP/wine/dlls/winsock/tests/sock.c Fri Apr 26 20:56:53 2002 @@ -26,7 +26,6 @@ #include <winsock2.h> #define MAX_CLIENTS 4 /* Max number of clients */ -#define NUM_TESTS 2 /* Number of tests performed */ #define FIRST_CHAR 'A' /* First character in transferred pattern */ #define BIND_SLEEP 10 /* seconds to wait between attempts to bind() */ #define BIND_TRIES 6 /* Number of bind() attempts */ @@ -86,6 +85,7 @@ test_params *general; DWORD sock_flags; int buflen; + int timeout; } client_params; /* This type combines all information for setting up a test scenario */ @@ -104,6 +104,7 @@ SOCKET s; struct sockaddr_in addr; sock_info sock[MAX_CLIENTS]; + int id; } server_memory; /* Thread local storage for client */ @@ -113,6 +114,7 @@ struct sockaddr_in addr; char *send_buf; char *recv_buf; + int id; } client_memory; /**************** Static variables ***************/ @@ -237,6 +239,7 @@ server_memory *mem = (LPVOID) LocalAlloc ( LPTR, sizeof (server_memory)); TlsSetValue ( tls, mem ); + mem->id = GetCurrentThreadId (); mem->s = WSASocketA ( AF_INET, gen->sock_type, gen->sock_prot, NULL, 0, par->sock_flags ); ok ( mem->s != INVALID_SOCKET, "Server: WSASocket failed" ); @@ -267,7 +270,7 @@ if ( mem->sock[i].s != INVALID_SOCKET ) closesocket ( mem->sock[i].s ); } - ok ( closesocket ( mem->s ) == 0, "closesocket failed" ); + ok ( closesocket ( mem->s ) == 0, "server (%x): closesocket failed", mem->id ); LocalFree ( (HANDLE) mem ); ExitThread ( GetCurrentThreadId () ); } @@ -283,6 +286,7 @@ WaitForSingleObject ( server_ready, INFINITE ); + mem->id = GetCurrentThreadId (); mem->s = WSASocketA ( AF_INET, gen->sock_type, gen->sock_prot, NULL, 0, par->sock_flags ); @@ -310,6 +314,39 @@ ExitThread ( GetCurrentThreadId () ); } +static SOCKET do_async_connect ( client_memory *mem ) +{ + HANDLE event; + WSANETWORKEVENTS wsa_events; + int tmp, err; + + event = WSACreateEvent (); + + WSAEventSelect ( mem->s, event, FD_CONNECT ); + + tmp = connect ( mem->s, (struct sockaddr*) &mem->addr, sizeof ( mem->addr ) ); + if ( tmp != 0 && ( err = WSAGetLastError () ) != WSAEWOULDBLOCK ) + ok ( 0, "do_async_connect (%x): connect error: %d", mem->id, err ); + + tmp = WaitForSingleObject ( event, INFINITE ); + ok ( tmp == WAIT_OBJECT_0, + "do_async_connect (%x): wait for connect event failed: %d", mem->id, tmp ); + err = WSAEnumNetworkEvents ( mem->s, event, &wsa_events ); + wsa_ok ( err, 0 ==, "do_async_connect (%x): WSAEnumNetworkEvents error: %d\n" ); + + err = wsa_events.iErrorCode[ FD_CONNECT_BIT ]; + ok ( err == 0, "do_async_connect (%x): connect error: %d", mem->id, err ); + + WSAEventSelect ( mem->s, (HANDLE) 0, 0 ); + WSACloseEvent ( event ); + + if ( err ) + return INVALID_SOCKET; + + trace ( "do_async_connect (%x) connected\n", mem->id ); + return mem->s; +} + /**************** Servers ***************/ /* @@ -323,11 +360,10 @@ id = GetCurrentThreadId(); char *p; - trace ( "simple_server (%x) starting\n", id ); - set_so_opentype ( FALSE ); /* non-overlapped */ server_start ( par ); mem = TlsGetValue ( tls ); + id = mem->id; wsa_ok ( set_blocking ( mem->s, TRUE ), 0 ==, "simple_server (%x): failed to set blocking mode: %d"); wsa_ok ( listen ( mem->s, SOMAXCONN ), 0 ==, "simple_server (%x): listen failed: %d"); @@ -352,7 +388,8 @@ ok ( n_recvd == n_expected, "simple_server (%x): received less data then expected: %d of %d", id, n_recvd, n_expected ); p = test_buffer ( mem->sock[0].buf, gen->chunk_size, gen->n_chunks ); - ok ( p == NULL, "simple_server (%x): test pattern error: %d", p - mem->sock[0].buf); + ok ( p == NULL, "simple_server (%x): test pattern error at %d; %c", + id, p - mem->sock[0].buf, (p ? *p : '*')); /* Echo data back */ n_sent = do_synchronous_send ( mem->sock[0].s, mem->sock[0].buf, n_expected, par->buflen ); @@ -381,16 +418,15 @@ int n_sent, n_recvd, n_expected = gen->n_chunks * gen->chunk_size, id; char *p; - id = GetCurrentThreadId(); - trace ( "simple_client (%x): starting\n", id ); /* wait here because we want to call set_so_opentype before creating a socket */ WaitForSingleObject ( server_ready, INFINITE ); - trace ( "simple_client (%x): server ready\n", id ); check_so_opentype (); set_so_opentype ( FALSE ); /* non-overlapped */ client_start ( par ); mem = TlsGetValue ( tls ); + id = mem->id; + trace ( "simple_client (%x) started\n", id ); /* Connect */ wsa_ok ( connect ( mem->s, (struct sockaddr*) &mem->addr, sizeof ( mem->addr ) ), @@ -414,7 +450,8 @@ /* check data */ p = test_buffer ( mem->recv_buf, gen->chunk_size, gen->n_chunks ); - ok ( p == NULL, "simple_client (%x): test pattern error: %d", id, p - mem->recv_buf); + ok ( p == NULL, "simple_client (%x): test pattern error at %d: %c", + id, p - mem->recv_buf, (p ? *p : '*')); /* cleanup */ read_zero_bytes ( mem->s ); @@ -429,37 +466,21 @@ { test_params *gen = par->general; client_memory *mem; - int id = GetCurrentThreadId(), n_expected = gen->n_chunks * gen->chunk_size, - tmp, err, n; - HANDLE event; + int n_expected = gen->n_chunks * gen->chunk_size, id, err, n; + HANDLE event = 0; WSANETWORKEVENTS wsa_events; char *send_last, *recv_last, *send_p, *recv_p; long mask = FD_READ | FD_WRITE | FD_CLOSE; - trace ( "event_client (%x): starting\n", id ); client_start ( par ); - trace ( "event_client (%x): server ready\n", id ); - mem = TlsGetValue ( tls ); + id = mem->id; + trace ( "event_client (%x) started\n", id ); - /* Prepare event notification for connect, makes socket nonblocking */ - event = WSACreateEvent (); - WSAEventSelect ( mem->s, event, FD_CONNECT ); - tmp = connect ( mem->s, (struct sockaddr*) &mem->addr, sizeof ( mem->addr ) ); - if ( tmp != 0 && ( err = WSAGetLastError () ) != WSAEWOULDBLOCK ) - ok ( 0, "event_client (%x): connect error: %d", id, err ); - - tmp = WaitForSingleObject ( event, INFINITE ); - ok ( tmp == WAIT_OBJECT_0, "event_client (%x): wait for connect event failed: %d", id, tmp ); - err = WSAEnumNetworkEvents ( mem->s, event, &wsa_events ); - wsa_ok ( err, 0 ==, "event_client (%x): WSAEnumNetworkEvents error: %d\n" ); - - err = wsa_events.iErrorCode[ FD_CONNECT_BIT ]; - ok ( err == 0, "event_client (%x): connect error: %d", id, err ); - if ( err ) goto out; - - trace ( "event_client (%x) connected\n", id ); + if ( do_async_connect ( mem ) == INVALID_SOCKET ) + goto out; + event = WSACreateEvent (); WSAEventSelect ( mem->s, event, mask ); recv_p = mem->recv_buf; @@ -470,7 +491,7 @@ while ( TRUE ) { err = WaitForSingleObject ( event, INFINITE ); - ok ( err == WAIT_OBJECT_0, "event_client (%x): wait failed", id, tmp ); + ok ( err == WAIT_OBJECT_0, "event_client (%x): wait failed: %d", id, err ); err = WSAEnumNetworkEvents ( mem->s, event, &wsa_events ); wsa_ok ( err, 0 ==, "event_client (%x): WSAEnumNetworkEvents error: %d\n" ); @@ -535,20 +556,183 @@ } ok ( send_p == send_last, - "simple_client (%x): sent less data then expected: %d of %d", + "event_client (%x): sent less data then expected: %d of %d", id, send_p - mem->send_buf, n_expected ); ok ( recv_p == recv_last, - "simple_client (%x): received less data then expected: %d of %d", + "event_client (%x): received less data then expected: %d of %d", id, recv_p - mem->recv_buf, n_expected ); recv_p = test_buffer ( mem->recv_buf, gen->chunk_size, gen->n_chunks ); - ok ( recv_p == NULL, "event_client (%x): test pattern error: %d", id, recv_p - mem->recv_buf); + ok ( recv_p == NULL, "event_client (%x): test pattern error at %d: %c", + id, recv_p - mem->recv_buf, (recv_p ? *recv_p : '*')); out: - WSACloseEvent ( event ); + if ( event ) WSACloseEvent ( event ); trace ( "event_client (%x) exiting\n", id ); client_stop (); } +/* + * ReadFile_client: A client using overlapped IO with ReadFile() / WriteFile() + */ +static void WINAPI ReadFile_client ( client_params *par ) +{ + test_params *gen = par->general; + client_memory *mem; + int id, n_expected = gen->n_chunks * gen->chunk_size, n_requests; + int tmp, err, i, sent_total = 0, recvd_total = 0; + char *send_last, *recv_last, *send_p, *recv_p; + LPHANDLE event = NULL, send_event, recv_event; + LPOVERLAPPED send_over, recv_over; + LPDWORD n_sent, n_recvd; + + client_start ( par ); + mem = TlsGetValue ( tls ); + id = mem->id; + trace ( "ReadFile_client (%x): started\n", id ); + + if ( do_async_connect ( mem ) == INVALID_SOCKET ) + goto out; + + if ( n_expected % par->buflen == 0 ) + n_requests = n_expected / par->buflen; + else + n_requests = n_expected / par->buflen + 1; + + event = (LPVOID) LocalAlloc ( LPTR, ( 1 + 2 * n_requests ) * sizeof ( HANDLE ) ); + send_over = (LPVOID) LocalAlloc ( LPTR, 2 * n_requests * sizeof ( OVERLAPPED ) ); + n_sent = (LPVOID) LocalAlloc ( LPTR, 2 * n_requests * sizeof ( DWORD ) ); + + send_event = event + 1; + recv_event = send_event + n_requests; + recv_over = send_over + n_requests; + n_recvd = n_sent + n_requests; + + event[0] = WSACreateEvent (); + WSAEventSelect ( mem->s, event[0], FD_CLOSE ); + + for ( i = 0; i < 2 * n_requests; i++ ) + event[ 1 + i ] = CreateEventA ( NULL, TRUE, FALSE, NULL ); + + /* Brutal: queue all requests right away */ + send_last = mem->send_buf + n_expected; + recv_last = mem->recv_buf + n_expected; + for ( i = 0, send_p = mem->send_buf, recv_p = mem->recv_buf; + i < n_requests; + i++, send_p += par->buflen, recv_p += par->buflen ) + { + tmp = min ( par->buflen, send_last - send_p ); + send_over[i].hEvent = send_event[i]; + err = WriteFile ( (HANDLE) mem->s, send_p, tmp, &n_sent[i], &send_over[i] ); + if ( !err ) tmp = GetLastError (); + ok ( err == TRUE || tmp == ERROR_IO_PENDING, + "ReadFile_client (%x): WriteFile error: %d", id, tmp ); + + tmp = min ( par->buflen, recv_last - recv_p ); + recv_over[i].hEvent = recv_event[i]; + err = ReadFile ( (HANDLE) mem->s, recv_p, tmp, &n_recvd[i], &recv_over[i] ); + if ( !err ) tmp = GetLastError (); + ok ( err == TRUE || tmp == ERROR_IO_PENDING, + "ReadFile_client (%x): ReadFile error: %d", id, tmp ); + } + trace ( "ReadFile_client (%x): all requests queued\n", id ); + shutdown ( mem->s, SD_SEND ); + + /* Wait for the FD_CLOSE event */ + err = WaitForSingleObject ( event[0], par->timeout ); + + ok ( err == WAIT_OBJECT_0 || err == WAIT_TIMEOUT, + "ReadFile_client (%x): error waiting for FD_CLOSE: %d", id, err ); + + if ( err == WAIT_OBJECT_0 ) + trace ( "ReadFile_client (%x): close event\n", id ); + else if ( err == WAIT_TIMEOUT ) + trace ( "ReadFile_client (%x): timeout\n", id ); + + /* Wait for read results only if TIMEOUT received - indicates not all data could be read */ + for ( i = 0, recv_p = mem->recv_buf; i < n_requests; i++, recv_p += par->buflen ) + { + err = GetOverlappedResult ( (HANDLE) mem->s, &send_over[i], &n_sent[i], FALSE ); + ok ( err == TRUE, "ReadFile_client (%x): write request %d error %d", + id, i, GetLastError () ); + sent_total += n_sent[i]; + + err = GetOverlappedResult ( (HANDLE) mem->s, &recv_over[i], &n_recvd[i], err == WAIT_TIMEOUT ); + ok ( err == TRUE, "ReadFile_client (%x): read request %d error %d", + id, i, GetLastError () ); + + /* If a read did not get all requested bytes, rearrange data */ + if ( mem->recv_buf + recvd_total < recv_p ) + memmove ( mem->recv_buf + recvd_total, recv_p, n_recvd[i] ); + recvd_total += n_recvd[i]; + } + + /* Try to read remaining data, if any */ + while ( recvd_total < n_expected ) + { + trace ( "ReadFile_client (%x): %d bytes missing, retrying ...\n", id, n_expected - recvd_total ); + recv_p = mem->recv_buf + recvd_total; + ResetEvent ( recv_over[0].hEvent ); + err = ReadFile ( (HANDLE) mem->s, recv_p, recv_last - recv_p, &n_recvd[0], &recv_over[0] ); + if ( err == FALSE ) + { + err = GetLastError (); + ok ( err == ERROR_IO_PENDING, "ReadFile_client (%x): ReadFile(2) error: %d", id, err ); + if ( err != ERROR_IO_PENDING ) break; + err = GetOverlappedResult ( (HANDLE) mem->s, &recv_over[0], &n_recvd[0], TRUE ); + ok ( err == TRUE, "ReadFile_client (%x): read request error %d", + id, GetLastError () ); + if ( err == FALSE ) break; + if ( n_recvd[0] == 0 ) + { + trace ( "ReadFile_client (%x): empty receive (overlapped)\n", id ); + break; + } + recvd_total += n_recvd[0]; + } + else + { + if ( n_recvd[0] == 0 ) + { + trace ( "ReadFile_client (%x): empty receive (immediate)\n", id ); + break; + } + recvd_total += n_recvd[0]; + } + } + + /* Wait for the FD_CLOSE event again - it should be signalled now */ + err = WaitForSingleObject ( event[0], par->timeout ); + + ok ( err == WAIT_OBJECT_0, + "ReadFile_client (%x): error waiting for FD_CLOSE: %d", id, err ); + if ( err == WAIT_OBJECT_0 ) + trace ( "ReadFile_client (%x): close event (2)\n" ); + + /* Check IO results */ + ok ( sent_total == n_expected, + "ReadFile_client (%x): sent less data then expected: %d of %d", + id, sent_total, n_expected ); + + ok ( recvd_total == n_expected, + "ReadFile_client (%x): received less data then expected: %d of %d", + id, recvd_total, n_expected ); + + recv_p = test_buffer ( mem->recv_buf, gen->chunk_size, gen->n_chunks ); + ok ( recv_p == NULL, + "ReadFile_client (%x): test pattern error at %d: %c", + id, recv_p - mem->recv_buf, (recv_p ? *recv_p : '*')); + + for (i = 0; i < 1 + 2 * n_requests; i++) + WSACloseEvent ( event[i] ); + + LocalFree ( (HANDLE)event ); + LocalFree ( (HANDLE)send_over ); + LocalFree ( (HANDLE)n_sent ); +out: + trace ( "ReadFile_client (%x) exiting\n", id ); + client_stop (); +} + /**************** Main program utility functions ***************/ static void Init (void) @@ -627,7 +811,7 @@ inet_addr: "127.0.0.1", \ inet_port: 9374 -static test_setup tests [NUM_TESTS] = +static test_setup tests [] = { /* Test 0: synchronous client and server */ { @@ -662,6 +846,25 @@ clt_params: { buflen: 128 } + }, + /* Test 2: ReadFile() overlapped client, synchronous server */ + { + general: { + STD_STREAM_SOCKET, + chunk_size: 2084, + n_chunks: 16, + n_clients: 2 + }, + srv: simple_server, + srv_params: { + buflen: 64 + }, + clt: ReadFile_client, + clt_params: { + sock_flags: WSA_FLAG_OVERLAPPED, + buflen: 128, + timeout: 1000 + } } }; @@ -670,9 +873,10 @@ START_TEST( sock ) { int i; - Init(); + int n_tests = sizeof (tests) / sizeof (test_setup); - for (i = 0; i < NUM_TESTS; i++) + Init(); + for (i = 0; i < n_tests; i++) { trace ( " **** STARTING TEST %d **** \n", i ); do_test ( &tests[i] );

1 0

Improved FD_CLOSE scheme for sockets
by Martin Wilck 26 Apr '02

26 Apr '02

Just a small patch to the server sock.c - it tries to implement exactly the strategies that I described in my previous FD_CLOSE posting. Seems to work pretty well with my test cases. It makes a difference to my previous implementation only in rare cases, but I am sure it will be better at preventing server busy-loops in these cases. Martin diff -ruNX ignore CVS/wine/server/sock.c TMP/wine/server/sock.c --- CVS/wine/server/sock.c Fri Apr 26 20:58:26 2002 +++ TMP/wine/server/sock.c Fri Apr 26 20:36:44 2002 @@ -188,8 +188,6 @@ /* We need this to delay FD_CLOSE events until all pending overlapped requests are processed */ if ( !events || async_active ) return; - if (events & FD_CLOSE) sock->hmask |= FD_CLOSE; - if (sock->event) { if (debug_level) fprintf(stderr, "signalling events %x ptr %p\n", events, sock->event ); @@ -284,7 +282,7 @@ { /* incoming data */ sock->pmask |= FD_READ; - sock->hmask |= FD_READ; + sock->hmask |= (FD_READ|FD_CLOSE); sock->errors[FD_READ_BIT] = 0; if (debug_level) fprintf(stderr, "socket %d is readable\n", sock->obj.fd ); @@ -330,6 +328,7 @@ if ( event & ( POLLERR|POLLHUP ) ) sock->state &= ~(FD_WINE_CONNECTED|FD_WRITE); sock->pmask |= FD_CLOSE; + sock->hmask |= FD_CLOSE; if (debug_level) fprintf(stderr, "socket %d aborted by error %d, event: %x - removing from select loop\n", sock->obj.fd, sock->errors[FD_CLOSE_BIT], event); @@ -390,7 +389,7 @@ if (mask & FD_WRITE || (sock->flags & WSA_FLAG_OVERLAPPED && IS_READY (sock->write_q))) ev |= POLLOUT; /* We use POLLIN with 0 bytes recv() as FD_CLOSE indication. */ - if (sock->mask & ~sock->hmask & FD_CLOSE && !(sock->hmask & FD_READ) ) + if (sock->mask & ~sock->hmask & FD_CLOSE) ev |= POLLIN; return ev; @@ -447,6 +446,7 @@ { case ASYNC_TYPE_READ: q = &sock->read_q; + sock->hmask &= ~FD_CLOSE; break; case ASYNC_TYPE_WRITE: q = &sock->write_q; @@ -478,7 +478,7 @@ async_insert ( q, async ); } } - else if ( async ) destroy_async ( async ); + else if ( async ) destroy_async ( async ); else set_error ( STATUS_INVALID_PARAMETER ); pollev = sock_reselect ( sock ); @@ -788,6 +788,8 @@ sock->pmask &= ~req->mask; /* is this safe? */ sock->hmask &= ~req->mask; + if ( req->mask & FD_READ ) + sock->hmask &= ~FD_CLOSE; sock->state |= req->sstate; sock->state &= ~req->cstate;

1 0

Re: wine/ dlls/winsock/socket.c dlls/winsock/ws2_3 ...
by Martin Wilck 26 Apr '02

26 Apr '02

Alexandre, -static int FILE_GetUnixHandleType( HANDLE handle, DWORD access, enum fd_type *type, int *flags ) +static int FILE_GetUnixHandleType( HANDLE handle, DWORD access, enum fd_type *type, int *flags_ptr ) { - int ret, fd = -1; + int ret, flags, fd = -1; - ret = wine_server_handle_to_fd( handle, access, &fd, type, flags ); + ret = wine_server_handle_to_fd( handle, access, &fd, type, &flags ); + if (flags_ptr) *flags_ptr = flags; This is of course the most robust solution - but because FILE_GetUnixHandleType is internal, you might as well require all functions calling it to pass a valid flags pointer, and let FILE_GetUnixHandle call wine_server_handle_to_fd directly, at a minimum performance gain. Martin -- Martin Wilck Phone: +49 5251 8 15113 Fujitsu Siemens Computers Fax: +49 5251 8 20409 Heinz-Nixdorf-Ring 1 mailto:Martin.Wilck@Fujitsu-Siemens.com D-33106 Paderborn http://www.fujitsu-siemens.com/primergy

2 2

bug in wine install
by Yannick Fillatrau 26 Apr '02

26 Apr '02

I just have downloaded the source of wine v. 20020411, I launched ./tools/wineinstall as a normal user and I was told to enter the root password, the install crashed. Here is a sample of stdout: Wine build complete. --- Performing 'make install' as root to install binaries, enter root password Mot de passe: /bin/bash: - : unrecognized option Either you entered an incorrect password or we failed to run 'make install' correctly. If you didn't enter an incorrect password then please report this error to wine-devel(a)winehq.com. Installation failed, aborting. --- I'm sure the password was correct...

2 1

FD_CLOSE, again
by Martin Wilck 26 Apr '02

26 Apr '02

I still have problems with FD_CLOSE signalling. AFAICS FD_CLOSE must be signalled when either of the three following conditions hold: 1 Network error (Linux signals POLLERR) 2 Graceful close by the peer (Linux signals POLLHUP) 3 shutdown (SD_SEND) by the peer (Linux signals POLLIN with 0 bytes to read). ** Condition 3 will only be noticed if we tell poll() we're interested ** ** in POLLIN. ** In that case, if 1 or 2 occurs, POLLIN will also be signalled, whether or not there are still data to be read. However, we cannot simply activate POLLIN always, because if there _is_ data to be read on a normal connection and the app does not read it immediately, the server would busy-loop. We can take advantage of the fact that after every recv(), winsock will notify the server and try to re-enable FD_READ. (This does not hold for ReadFile() operations, as I pointed out in my recent "RFC..." posting). Thus what we could do is enable polling for POLLIN whenever FD_READ is re-enabled, even if then app has not registered for FD_READ notification, and disable it when POLLIN is received. Furthermore, finishing an asynchronous read request would also reenable POLLIN, whether or not there are more requests in read queue. Only an app that uses synchronous ReadFile() calls would never see FD_CLOSE. Does that seem correct, or am I again overlooking something?? It is a completely different story _at what time_ FD_CLOSE should be signalled when one ove the above 3 conditions is actually encountered. The Winsock2 spec says different things: In docs of WSAEventSelect: "FD_CLOSE should only be posted after all data is read from a socket, but an application should check for remaining data upon receipt of FD_CLOSE to avoid any possibility of losing data". This suggests that we could signal FD_CLOSE if POLLHUP is received, even if there is still data to be read, although we "should" not. In docs of WSAAsyncSelect: "[FD_CLOSE is posted] after remote system initiated graceful close, when no data currently available to receive (note: if data has been received and is waiting to be read when the remote system initiates a graceful close, the FD_CLOSE is not delivered until all pending data has been read)" This suggests the opposite - signal FD_CLOSE only after the app has read all data, even if this is far later than the actual POLLHUP. Actually this suggests that FD_CLOSE is _equivalent_ with condition 3 above, regardless if 1 or 2 are in effect. This is the behaviour I have tried to implement in my latest patches. Again, am I overlooking something? Martin -- Martin Wilck Phone: +49 5251 8 15113 Fujitsu Siemens Computers Fax: +49 5251 8 20409 Heinz-Nixdorf-Ring 1 mailto:Martin.Wilck@Fujitsu-Siemens.com D-33106 Paderborn http://www.fujitsu-siemens.com/primergy

1 0

RFC: How should ReadFile() on sockets behave ??
by Martin Wilck 26 Apr '02

26 Apr '02

If there is a socket for which event notification has been requested by the application with WSAEventSelect(), the behaviour is as follows: - When data arrives, FD_READ is signalled _ONCE_. - FD_READ is only delivered again if the app explicitly reenables it with WSAEventSelect/WSAAsyncSelect or if the application issues a recv() call. - The latter is accomplished in wine by calling _enable_event ( FD_READ, ...) after every recv() operation in the winsock code. - If an app uses ReadFile() rather than a winsock call to retrieve data, the FD_READ event is of course not reenabled, because the current implementation in file.c has no concept of socket events. Question: should ReadFile() behave like recv(), i.e. reenable FD_READ, or not? If yes, we would have a certain dependency of the ReadFile() code on winsock, because FD_READ etc are winsock-specific macros. If no, an app that relies on ReadFile() reenabling FD_READ will obviously fail. The best thing would probably be to do this in the server, but we currently can't, because the server does not notice when a read operation is finished. Opinions? Martin -- Martin Wilck Phone: +49 5251 8 15113 Fujitsu Siemens Computers Fax: +49 5251 8 20409 Heinz-Nixdorf-Ring 1 mailto:Martin.Wilck@Fujitsu-Siemens.com D-33106 Paderborn http://www.fujitsu-siemens.com/primergy

2 2

file.c:FILE_ReadFileEx - fd is not overlapped
by Marcelo Welter 26 Apr '02

26 Apr '02

Helle friends Still fighting with my serial printers.. :) I upgrade my wine sources today with CVS and more advantages are in this new code (i'm very happy) but with one printer i receive the messages: trace:file:fd is not overlapped i comment the lines //if ( ! (flags & FD_FLAG_OVERLAPPED) ) { // WARN ( "fd is not overlapped\n" ); // SetLastError ( ERROR_INVALID_PARAMETER ); // goto error; //} in files/file.c and my printer communicate good. What i can change in my application to work with serial in this case See in +comm,+file trace:comm:ClearCommError handle 48 cbInQue = 2 cbOutQue = 0 trace:file:ReadFile 48 0x10025308 2 0x40566828 (nil) trace:file:FILE_TimeoutRead 48 0x10025308 2 0x40566828 trace:file:FILE_ReadFileEx file 48 to buf 0x10025308 num 2 0x40566794 func (nil) warn:file:FILE_ReadFileEx fd is not overlapped trace:comm:ClearCommError handle 48 cbInQue = 2 cbOutQue = 0 trace:file:ReadFile 48 0x10025308 2 0x40566828 (nil) trace:file:FILE_TimeoutRead 48 0x10025308 2 0x40566828 trace:file:FILE_ReadFileEx file 48 to buf 0x10025308 num 2 0x40566794 func (nil) warn:file:FILE_ReadFileEx fd is not overlapped Marcelo

2 1

Byte Swapping in storage32.c for solaris
by Warren_Baird＠cimmetry.com 26 Apr '02

26 Apr '02

I've been running into some problems with byte swapping in ole related code - it looks like the relevant stuff is in storage32.c - I've modified StorageUtl_{Read,Write}Word and DWord to byteswap on bigendian platforms, and that seems to help, but I'm also running into problems with the wchar property names in StorageImpl_{Read,Write}Property - the individual characters are showing up in the wrong order, so "Data" is showing up as 0x44006100 0x74006100 instead of 0x00440061 0x00740061 as I would expect... Is it safe to use the same approach here? should ReadProperty byteswap each wchar and then WriteProperty byteswap it again? There seem to be a lot of places things like this happen: StorageImpl_ReadProperty(This, propertyIndex, &chainProperty); chainProperty.startingBlock = bbHeadOfChain; StorageImpl_WriteProperty(This, propertyIndex, &chainProperty); and I'm not sure how else to handle these cases... Also - at the moment the only byteswapping macros actually appear in tools/wrc/wrctypes.h - would it make sense to move them to one of the top-level include directories, and use them in both storage32.c and in wrc? If so, which include file should I drop them in? Thanks! Warren

1 0

Bug 572
by Andriy Palamarchuk 26 Apr '02

26 Apr '02

Alexandre, could you look at bug 572: http://bugs.codeweavers.com/show_bug.cgi?id=572 It is possible that this bug is caused by your patch. Thanks, Andriy Palamarchuk __________________________________________________ Do You Yahoo!? Yahoo! Games - play chess, backgammon, pool and more http://games.yahoo.com/

2 2

existing cooperation with mono?
by Freddy BL 26 Apr '02

26 Apr '02

Hi Wine-Team! Existing cooperations between the Wine-project and the Mono-project (http://www.go-mono.org)? At the moment it seams, that in the future there exists under MS-Windows three types of programs: programs which are in native-code (that can be run on Wine and can be portet to Linux by re-compiling with Wine), programs which are in MS-IL-code (that can in the future be run on Mono I hope) and mixed programs (IL-Code-programs, which integrate Native-Code DLLs and Native-Code-programs, which integrate IL-Code DLLs). And I think, that for the last type of programs a cooperation between the Wine-Devels and the Mono-Devels is very advantageous. So, existing this cooperations? (One of the results of an coopration could be an installer like that by CodeWeavers, which can be install both (Wine and Mono) for example) Greatings Freddy BL _________________________________________________________________ Senden und empfangen Sie Hotmail über Ihr Mobilgerät: http://mobile.msn.com

3 2

[Germany] Linuxtag
by Uwe Bonnes 26 Apr '02

26 Apr '02

Hallo, is there anybody beside me willing to care for a booth at Linuxtag June 6-9 in Karlsruhe? Please give feedback so I can apply for a booth. Beside our normal program winex and crossover might be good objects to demonstrate. Bye -- Uwe Bonnes bon(a)elektron.ikp.physik.tu-darmstadt.de Institut fuer Kernphysik Schlossgartenstrasse 9 64289 Darmstadt --------- Tel. 06151 162516 -------- Fax. 06151 164321 ----------

1 0

More BUG fixes for winsock2 patch series
by Martin Wilck 26 Apr '02

26 Apr '02

Patch: BUGS in my recently submitted patch series This applies over my complete patch series - I will send clean patches tomorrow. The fixes in server/file.c and server/serial.c relate to Rein Klazes' problem. Rein: please tell me if they fix it!! Patch http://www.winehq.com/hypermail/wine-patches/2002/04/0163.html is also needed. Modfied files: dlls/winsock: socket.c (allocation error) server: sock.c (remove socket from select loop for blocking sockets, too) file.c (try to serve async requests immediately) serial.c (dito) diff -ruNX ignore CVS/wine/dlls/winsock/socket.c TMP/wine/dlls/winsock/socket.c --- CVS/wine/dlls/winsock/socket.c Thu Apr 25 20:20:37 2002 +++ TMP/wine/dlls/winsock/socket.c Thu Apr 25 19:46:05 2002 @@ -2479,7 +2479,7 @@ goto error; } - iovec = WS_ALLOC ( dwBufferCount * sizeof (struct iovec) ); + iovec = HeapAlloc (GetProcessHeap(), 0, dwBufferCount * sizeof (struct iovec) ); if ( !iovec ) { diff -ruNX ignore CVS/wine/server/file.c TMP/wine/server/file.c --- CVS/wine/server/file.c Thu Apr 25 20:20:36 2002 +++ TMP/wine/server/file.c Thu Apr 25 20:18:57 2002 @@ -389,6 +389,8 @@ if ( status == STATUS_PENDING ) { + struct pollfd pfd; + if ( !async ) async = create_async ( obj, current, ptr ); if ( !async ) @@ -397,6 +399,15 @@ async->status = STATUS_PENDING; if ( !async->q ) async_insert( q, async ); + + /* Check if the new pending request can be served immediately */ + pfd.fd = obj->fd; + pfd.events = file_get_poll_events ( obj ); + pfd.revents = 0; + poll ( &pfd, 1, 0 ); + + if ( pfd.revents ) + file_poll_event ( obj, pfd.revents ); } else if ( async ) destroy_async ( async ); else set_error ( STATUS_INVALID_PARAMETER ); diff -ruNX ignore CVS/wine/server/serial.c TMP/wine/server/serial.c --- CVS/wine/server/serial.c Thu Apr 25 20:20:36 2002 +++ TMP/wine/server/serial.c Thu Apr 25 20:27:11 2002 @@ -282,6 +282,8 @@ if ( status == STATUS_PENDING ) { + struct pollfd pfd; + if ( !async ) async = create_async ( obj, current, ptr ); if ( !async ) @@ -293,6 +295,19 @@ async_add_timeout(async,timeout); async_insert(q, async); } + + /* Check if the new pending request can be served immediately */ + pfd.fd = obj->fd; + pfd.events = serial_get_poll_events ( obj ); + pfd.revents = 0; + poll ( &pfd, 1, 0 ); + + if ( pfd.revents ) + /* serial_poll_event() calls set_select_events() */ + serial_poll_event ( obj, pfd.revents ); + else + set_select_events ( obj, pfd.events ); + return; } else if ( async ) destroy_async ( async ); else set_error ( STATUS_INVALID_PARAMETER ); diff -ruNX ignore CVS/wine/server/sock.c TMP/wine/server/sock.c --- CVS/wine/server/sock.c Thu Apr 25 20:20:36 2002 +++ TMP/wine/server/sock.c Thu Apr 25 15:10:54 2002 @@ -250,7 +250,6 @@ sock->errors[FD_CONNECT_BIT] = sock_error( sock->obj.fd ); if (debug_level) fprintf(stderr, "socket %d connection failure\n", sock->obj.fd); - set_select_events( &sock->obj, -1 ); } } else if (sock->state & FD_WINE_LISTENING) @@ -269,7 +268,6 @@ sock->pmask |= FD_ACCEPT; sock->errors[FD_ACCEPT_BIT] = sock_error( sock->obj.fd ); sock->hmask |= FD_ACCEPT; - set_select_events( &sock->obj, -1 ); } } else { @@ -335,11 +333,17 @@ if (debug_level) fprintf(stderr, "socket %d aborted by error %d, event: %x - removing from select loop\n", sock->obj.fd, sock->errors[FD_CLOSE_BIT], event); - set_select_events( &sock->obj, -1 ); } } - sock_reselect( sock ); + if ( sock->pmask & FD_CLOSE || event & (POLLERR|POLLHUP) ) + { + if ( debug_level ) + fprintf ( stderr, "removing socket %d from select loop\n", sock->obj.fd ); + set_select_events( &sock->obj, -1 ); + } + else + sock_reselect( sock ); /* wake up anyone waiting for whatever just happened */ if ( sock->pmask & sock->mask || sock->flags & FD_FLAG_OVERLAPPED ) sock_wake_up( sock, event );

3 2

Last header patch for Mingw32 I think (wingdi.h)
by Steven Edwards 26 Apr '02

26 Apr '02

cvs diff -u wingdi.h Index: wingdi.h =================================================================== RCS file: /home/wine/wine/include/wingdi.h,v retrieving revision 1.69 diff -u -r1.69 wingdi.h --- wingdi.h 3 Apr 2002 22:08:27 -0000 1.69 +++ wingdi.h 25 Apr 2002 16:48:19 -0000 @@ -24,6 +24,14 @@ extern "C" { #endif +#ifdef __MINGW__ +#undef _MAX_PATH +#undef _MAX_DIR +#undef _MAX_EXT +#undef _MAX_DRIVE +#include "stdlib.h" +#endif + typedef struct _ABCFLOAT { FLOAT abcfA; FLOAT abcfB; "Every revolution was once a thought in one man's mind" - Ralph Waldo Emerson

2 3

Re: Send the first two mingw32 patches to /dev/null and use these instead
by Alexandre Julliard 26 Apr '02

26 Apr '02

"Steven Edwards" <Steven_Ed4153(a)yahoo.com> writes: > Index: windef.h > =================================================================== > RCS file: /home/wine/wine/include/windef.h,v > retrieving revision 1.62 > diff -u -r1.62 windef.h > --- windef.h 10 Mar 2002 00:02:34 -0000 1.62 > +++ windef.h 25 Apr 2002 02:09:36 -0000 > @@ -157,8 +157,9 @@ > #define _MAX_EXT _MAX_FNAME > #define _MAX_PATH 260 > #endif > +#if ! defined (__MINGW__) > #define MAX_PATH _MAX_PATH > - > +#endif This seems to be a mingw32 bug, it is not supposed to define MAX_PATH, only _MAX_PATH. At least that's how it's done under Windows. -- Alexandre Julliard julliard(a)winehq.com

2 1

First attempt at unit testing. Victim of choice: _hread()
by Jakob Eriksson 25 Apr '02

25 Apr '02

Thanks to Vincent Béron and Francois Gouget I've managed to do my own little unit test. I post it here hoping for comments of any kind: possible improvements, additions, advice etc. (I guess I could add some more limit-tests, trying to read many small chunks in a row see that I still end up in the right place in the file. Things like that.) -------- #include "winbase.h" #include "winerror.h" #include "wine/test.h" static void test__hread( void ) { LPCSTR filename = "testfile.xxx"; LPCSTR sillytext = "en larvig liten text dx \033 gx hej 84 hej 4484 ! \001\033 bla bla..."; HFILE filehandle; char buffer[10000]; long bytes_read; UINT bytes_wanted; UINT i; filehandle = _lcreat( filename, 0 ); ok( HFILE_ERROR != filehandle, "_lcreat complains." ); ok( HFILE_ERROR != _hwrite( filehandle, sillytext, strlen( sillytext ) ), "_hwrite complains." ); ok( HFILE_ERROR != _lclose(filehandle), "_lclose complains." ); filehandle = _lopen( filename, OF_READ ); bytes_read = _hread( filehandle, buffer, 2 * strlen( sillytext )); ok( strlen( sillytext ) == bytes_read, "file has wrong size."); for (bytes_wanted = 0; bytes_wanted < strlen( sillytext ); bytes_wanted++) { ok( 0 == _llseek( filehandle, 0, FILE_BEGIN ), "_llseek complains." ); ok( _hread( filehandle, buffer, bytes_wanted ) == bytes_wanted, "erratic _hread return value." ); for (i = 0; i < bytes_wanted; i++) { ok( buffer[i] == sillytext[i], "that's not what's written." ); } } ok( HFILE_ERROR != _lclose( filehandle ), "_lclose complains." ); ok( DeleteFileA( filename ) != 0, "DeleteFile complains." ); } START_TEST(file) { test__hread( ); } -- regards, Jakob Eriksson The wages of sin is debugging. -- Ron Jeffries

1 0

Re: wine/ dlls/winsock/socket.c dlls/winsock/ws2_3 ...
by Michael Cardenas 25 Apr '02

25 Apr '02

do these patches include his fixes from today? On Thursday 25 April 2002 05:34 pm, you wrote: > ChangeSet ID: 1019770464826040353932516 > CVSROOT: /opt/cvs-commit > Module name: wine > Changes by: julliard(a)wine.codeweavers.com 02/04/25 16:34:24 > > Modified files: > dlls/winsock : socket.c ws2_32.spec > files : file.c > include : winbase.h > > Log message: > Martin Wilck <Martin.Wilck(a)fujitsu-siemens.com> > Implemented overlapped WSARecvFrom(), WSASendTo(), and > WSAGetOverlappedResult(). > Fixed shutdown() to work correctly for overlapped sockets (do not > actually do a shutdown() system call as long as there are pending > overlapped requests). > Also added correct error reporting when trying to read or write on a > shut down connection, also for ReadFile()/WriteFile() calls. > > Patch: http://cvs.winehq.com/patch.py?id=1019770464826040353932516 > > Revision Changes Path > 1.92 +679 -122 wine/dlls/winsock/socket.c > 1.20 +1 -1 wine/dlls/winsock/ws2_32.spec > 1.143 +22 -5 wine/files/file.c > 1.142 +1 -1 wine/include/winbase.h -- michael cardenas lead windows compatibility engineer lindows.com "Be the change you want to see in the world" -Mahatma Gandhi

2 1

cprof Profiler
by John Levon 25 Apr '02

25 Apr '02

Hi, saw your question on the weekly news. I noticed nobody had pointed out that what was cprof is now available under the eazel-tools module of the Gnome CVS repository (I believe). It also has some changes from the old cprof version like shared library support regards john -- "I continue to be amazed at what Andrei can make templates do. Some of it still makes my head hurt." - Herb Sutter

1 0

Re: [PATCH UPDATE] tools/wineshelllink *METHOD 4*
by Paul Millar 25 Apr '02

25 Apr '02

On Thu, 25 Apr 2002, Speeddymon wrote: > + if [ -d "/usr/lib/menu" -a "$mode" = "menu" -a ! `whoami` = "root" ] > + then [...] > + elif [ `whoami` = "root" ] > + then > + mdk_entry >> "/usr/lib/menu/wine" > + fi Sorry, I haven't been following this but isn't the elif condition redundant? Also, AFAIK although its unusual, there's no reason an account name cannot have a space in it, so `whoami` should be "`whoami`" [spent too much time debugging strings in scripts ;^] ---- Paul Millar

4 3

Memori and Swap Full after some hours
by Marcelo Welter 25 Apr '02

25 Apr '02

I have a windows application run in a very old machine Pentium 133Mhz 32 Mb RAM (believes this still exist) The application running fine, but after some hours (3 or 4) the memory and swap are full and the machine run very slowly. I Close the wine and all back to normal I dont know what are happen with system, but is a very hard trouble to me. Anybody have any idea to apply in this case Thanks Marcelo Welter

4 6

Jump to page:

2025

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

2006

2005

2004

2003

2002

2001

wine-devel