Hi!
18-Июн-2006 06:43 [EMAIL PROTECTED] (Blair Campbell) wrote to
[EMAIL PROTECTED]:
BC> +++ lfnfuncs.c 18 Jun 2006 06:43:52 -0000 1.2
BC> +#define _PUSH_DS __emit__(0x1E)
BC> +#define _PUSH_ES __emit__(0x06)
BC> +#define _POP_DS __emit__(0x1F)
BC> +#define _POP_ES __emit__(0x07)
BC> char * getshortfilename( const char *longfilename )
BC> + _PUSH_DS;
BC> + _PUSH_ES;
BC> _DS = FP_SEG( longfilename );
BC> _SI = FP_OFF( longfilename );
BC> _ES = FP_SEG( shortfilename );
Note: saving ES _isn't_ required. Only DS should be preserved.
BC> + return( ( _CFLAG || _AX == 0x7100 ) ?
BC> + strcpy( shortfilename, longfilename ) : shortfilename );
May you explain, why you copy longfilname to shortfilename? Especially,
there is latent bug, because shortfilename is local buffer, which will be
overwritten by subsequent calls to getshortfilename(). May be, this is your
way to eliminate error message (about returning const string as non-const)?
But why you not add const modifier for return type?
BC> +static void __creat_or_truncate( const char * filename, int mode )
BC> + if( _CFLAG || ( handle = _AX ) == 0x7100 )
BC> + handle = creatnew( filename, mode );
BC> + _close( handle );
Same question: why you try to mix assign with comparision? This is one
line shorter in source, but harder to read (and understand) and less
code-efficient, than:
if (_CFLAG || _AX == 0x7100)
_AX = creatnew (filename, mode);
_close (_AX);
BC> -FILE * lfnfopen( const char *filename, const char *mode )
BC> +FILE * lfnfopen( const char *filename, char *mode )
Why?
BC> + if( strchr( mode, 'a' ) != NULL )
BC> + __creat_or_truncate( filename, 0 );
BC> + else if( strchr( mode, 'w' ) != NULL )
BC> + __creat_or_truncate( filename, 0 );
if (strchr( mode, 'a' ) ||
strchr( mode, 'w' ))
__creat_or_truncate (filename, 0);
BC> int lfnopen( const char *filename, int access, ... )
BC> + if( access & O_CREAT ) {
BC> + unsigned int mode = 0;
BC> + mode = va_arg( vargs, unsigned );
BC> + access |= O_CREAT;
BC> + __creat_or_truncate( filename, mode );
BC> + }
First: `access' is already contains O_CREAT bit set, so |= above is
redundant. Or there is bug and this bit should be cleared?
Second, optimization all of this block:
if (access & O_CREAT)
__creat_or_truncate (filename, va_arg (vargs, unsigned));
Using Tomcat but need to do more? Need to support web services, security?
Get stuff done quickly with pre-integrated technology to make your job easier
Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronimo
http://sel.as-us.falkag.net/sel?cmd=lnk&kid=120709&bid=263057&dat=121642
_______________________________________________
Freedos-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/freedos-devel
