anyone already tried that?
I'm getting some errors when compiling wine which I was able to fix, but World of Warcraft won't start... Will icc be supported at some time in the furure?
tom
Tomas Carnecky wrote:
anyone already tried that?
I'm getting some errors when compiling wine which I was able to fix, but World of Warcraft won't start... Will icc be supported at some time in the furure?
tom
last time I checked, icc didn't support stdcall calling convention, which is rather important for compiling wine (on x86) didn't check for its presence lately
A+
On 10/16/05, Tomas Carnecky [email protected] wrote:
I'm getting some errors when compiling wine which I was able to fix, but World of Warcraft won't start... Will icc be supported at some time in the furure?
Maybe, especially if people who want it supported pitch in by running the Wine regression tests and reporting any problems.
Dan Kegel wrote:
On 10/16/05, Tomas Carnecky [email protected] wrote:
I'm getting some errors when compiling wine which I was able to fix, but World of Warcraft won't start... Will icc be supported at some time in the furure?
Maybe, especially if people who want it supported pitch in by running the Wine regression tests and reporting any problems.
as icc has other cmd-line options, the compiler spits out lots of warnings, like this one:
icc -c -I. -I. -I../../include -I../../include -D__WINESRC__ -DNO_LIBWINE_PORT -DWINE_UNICODE_API="" -D_REENTRANT -fPIC -Wall -pipe -mpreferred-stack-boundary=2 -fno-strict-aliasing -gstabs+ -Wdeclaration-after-statement -Wpointer-arith -g -O2 -o sortkey.o sortkey.c iccbin: Command line warning: ignoring option '-mp'; no argument required iccbin: Command line warning: ignoring option '-g'; no argument required iccbin: Command line warning: ignoring option '-W'; no argument required
futhermore, lots of compiler warnings a la: ../../include/wine/unicode.h(106): remark #810: conversion from "int" to "WCHAR={unsigned short}" may lose significant bits return ch + wine_casemap_lower[wine_casemap_lower[ch >> 8] + (ch & 0xff)]; relay.c(178): remark #1684: conversion from pointer to same-sized integral type (potential portability problem) fprintf( outfile, "\t.byte 0x64\n\tmovw %%sp, (%d)\n", STACKOFFSET );
and much more...
and finally fails on file dlls/kernel/thread.c with: /tmp/iccbinN9FwHFas_: Assembler messages: /tmp/iccbinN9FwHFas_:2812: Error: symbol `SetLastError' is already defined /tmp/iccbinN9FwHFas_:2840: Error: symbol `GetCurrentThreadId' is already defined
tom
That's expected. To support icc, one would have to adjust the commandline options a bit depending on the compiler being used. In particular, one would want to suppress many compiler warnings like the ones you mentioned. icc provides a nice way of suppressing any desired set of warnings.
Dan Kegel wrote:
That's expected. To support icc, one would have to adjust the commandline options a bit depending on the compiler being used. In particular, one would want to suppress many compiler warnings like the ones you mentioned. icc provides a nice way of suppressing any desired set of warnings.
I used something like CFLAGS="-g -Wbrief -wd108,144,167,181,188,193,279,556,810,869,981,1287,1418,1419,1572" last time I tried A+
Eric Pouech wrote:
Dan Kegel wrote:
That's expected. To support icc, one would have to adjust the commandline options a bit depending on the compiler being used. In particular, one would want to suppress many compiler warnings like the ones you mentioned. icc provides a nice way of suppressing any desired set of warnings.
I used something like CFLAGS="-g -Wbrief -wd108,144,167,181,188,193,279,556,810,869,981,1287,1418,1419,1572"
problem is, even if I specify CFLAGS on the command-line, wine adds its own flags. So the actual compile command looks as follows:
icc -c -I. -I. -I../../include -I../../include -D__WINESRC__ -DNO_LIBWINE_PORT -DWINE_UNICODE_API="" -D_REENTRANT -fPIC -Wall -pipe -mpreferred-stack-boundary=2 -fno-strict-aliasing -gstabs+ -Wdeclaration-after-statement -Wpointer-arith -g -Wbrief -wd108,144,167,181,188,193,279,556,810,869,981,1287,1418,1419,1572 -o c_28604.o c_28604.c
I'd like to get rid of "-Wall -pipe -mpreferred-stack-boundary=2 -fno-strict-aliasing -gstabs+ -Wdeclaration-after-statement -Wpointer-arith"... how can I do that?
thanks tom
Eric Pouech wrote:
Dan Kegel wrote:
That's expected. To support icc, one would have to adjust the commandline options a bit depending on the compiler being used. In particular, one would want to suppress many compiler warnings like the ones you mentioned. icc provides a nice way of suppressing any desired set of warnings.
I used something like CFLAGS="-g -Wbrief -wd108,144,167,181,188,193,279,556,810,869,981,1287,1418,1419,1572"
ok.. I've found what I have to do.. override EXTRACFLAGS...
now I can compile the wine without warnings.. but here comes the first compiler error:
tom@eiger ~/cvs/wine/libs/wine $ icc -c -I. -I. -I../../include -I../../include -D__WINESRC__ -DBINDIR=""/usr/bin"" -DDLLDIR=""/usr/lib/wine"" -D_REENTRANT -fPIC -g -Wbrief -wd108,144,167,181,188,193,279,556,810,869,981,1287,1418,1419,1572 -o ldt.o ldt.c /tmp/iccbinFR1dsuas_: Assembler messages: /tmp/iccbinFR1dsuas_:2339: Error: symbol `wine_get_fs' is already defined /tmp/iccbinFR1dsuas_:2363: Error: symbol `wine_set_fs' is already defined tom@eiger ~/cvs/wine/libs/wine $
tom
On Sat, 2005-10-29 at 14:58 +0200, Tomas Carnecky wrote:
I used something like CFLAGS="-g -Wbrief -wd108,144,167,181,188,193,279,556,810,869,981,1287,1418,1419,1572"
I suppose these disables some warnings. Wouldn't they be useful for Wine if fixed?
Dimi Paun wrote:
On Sat, 2005-10-29 at 14:58 +0200, Tomas Carnecky wrote:
I used something like CFLAGS="-g -Wbrief -wd108,144,167,181,188,193,279,556,810,869,981,1287,1418,1419,1572"
I suppose these disables some warnings. Wouldn't they be useful for Wine if fixed?
well, not all of them as you can see #108 bit field on signed type #111 statement is unreachable #144 assign to function pointer a value of type void* #167 cast from pointer to pointer, but of different type #181 incompatible pointer type for %p (char*) #188 enumerated type used as integer #193 0 used in some pre-processor when ID is not defined #279 controlling expression is constant (all are debug macros fire this) #310 old-C style parameter list #556 cast from function pointer to void* #593 variable set but never used #810 conversion to integral type of smaller size #869 parameter to func not referenced in func #981 arguments (to func) are evaluated in an unspecified order #1287 unknown __attribute__ to function definition #1418 global var/func without prior definition #1419 extern declaration in .c file #1572 comparison of floating point values
-
Dan Kegel -
Dimi Paun -
Eric Pouech -
Tomas Carnecky