Hello,
On 11/15/06, Wez Furlong <[EMAIL PROTECTED]> wrote:
We cannot change it in SAPI where we are not the .exe because we don't
own the process.
Messing with that flag affects the entire process and could break
things at a higher level.
I call this kind of thing "library abuse", where one pie
Hello,
On 11/15/06, Wez Furlong <[EMAIL PROTECTED]> wrote:
We cannot change it in SAPI where we are not the .exe because we don't
own the process.
Messing with that flag affects the entire process and could break
things at a higher level.
I call this kind of thing "library abuse", where one pie
We cannot change it in SAPI where we are not the .exe because we don't
own the process.
Messing with that flag affects the entire process and could break
things at a higher level.
I call this kind of thing "library abuse", where one piece of code
assumes that it is the only consumer of the librar
Hi Dmitry,
I think that changing global behavior (_fmode = _O_BINARY;) is not a good
idea.
May be this code should be removed from CGI and CLI too.
A big -1. It works and is known to work. It has been there since years.
May be we have a bug somewhere and _O_BINARY is not passed to open().
On 11/15/06, Pierre <[EMAIL PROTECTED]> wrote:
_fmode is global per process. We should set the mode of stdin/out:
_setmode( _fileno( stdin ), _O_BINARY );
If we need it and if it is already opened when we init the SAPI, but
someone has to verify this point.
I just checked it, we do set the m
