According to Ahmon Dancy: > It's definitely not a security patch. That was a miscommunication. > > >> I'd be glad to revise this *if* someone can confirm that EAGAIN is > >> portable and/or in some spec like POSIX. > >> > > The best I can do is confirm that EAGAIN exists on the following > operating systems: > > aix 4.2 > tru64 5.0a (where it is an alias for EWOULDBLOCK) > freebsd 4.0 (where EWOULDBLOCK is an alias for it) > hpux 10.20 > irix 6.5 (where EWOULDBLOCK is an alias for it)
I can add... NeXTSTEP 2.1 (old Mach/BSD based system) QNX 4.23 (POSIX-compliant RTOS - EWOULDBLOCK is an alias for it) SunOS 5.8 (where EWOULDBLOCK is an alias for it) Linux 2.0.36 (where EWOULDBLOCK is an alias for it) - also 2.2.14 & 2.4.9 I can't confirm it, but I vaguely recall seeing EAGAIN on even older SysV-based Concurrent RTU systems (though a hardcopy manual for this system suggests that EAGAIN is just for when fork() can't get a process table entry, while EWOULDBLOCK is for non blocking I/O, without access to the system anymore I can't see if one is an alias for another). I think it's been around a while. Indeed, if it was originally tied to fork(), it's probably been around since the 1970s. > You could also do: > > #ifdef EAGAIN > <my patch> > #endif > > or perhaps conditionalize on linux. Well, it's obviously not just on Linux systems, so I don't like the idea of limiting it to that. I was going to suggest the #ifdef EAGAIN conditional as well, which should remove any concerns about portability. -- Gilles R. Detillieux E-mail: <[EMAIL PROTECTED]> Spinal Cord Research Centre WWW: http://www.scrc.umanitoba.ca/~grdetil Dept. Physiology, U. of Manitoba Phone: (204)789-3766 Winnipeg, MB R3E 3J7 (Canada) Fax: (204)789-3930 _______________________________________________ htdig-general mailing list <[EMAIL PROTECTED]> To unsubscribe, send a message to <[EMAIL PROTECTED]> with a subject of unsubscribe FAQ: http://htdig.sourceforge.net/FAQ.html
