issue 24 is fixed by making sure the time fields use a static width

several noticable recent changes

changes changes changes

All test targets now run 'make all' before they prcoeed with the actual
testing so that all test files are build first properly. David Byron reported.

Gisle Vanem's djgpp/MS-DOS updates

fix to figure out the "real" windows path when built and run with mingw
Andr�s Garc�a helped out!

re-indented to use curl-standard source formatting

netinet/tcp.h may require netinet/in.h to be include before

use the correct struct

Tor fixed a left-over from the ip argument to setnodelay

Gisle Vanem:

A patch to bypass MS' sillyness with regard to IPv6 and getaddrinfo().

The CURLDEBUG part is to avoid redefinition warning caused by memdebug.h. If
ENABLE_IPV6 isn'

Gisle Vanem:

A patch to bypass MS' sillyness with regard to IPv6 and getaddrinfo().

The CURLDEBUG part is to avoid redefinition warning caused by memdebug.h. If
ENABLE_IPV6 isn't enabled, it doesn't matter since we never call
getaddrinfo(). Allthough we could to support weird protocols like SOCK_RDM
that Win-2K/XP has.

show more ...

some more password blurb

check for netinet/tcp.h precense before actually including it

removed the ip number from the notcpdelay function

get the version number from the new curlver.h header file

localtime and gmtime are not thread-safe on newer AIXes either so we force
a check for there *_r-versions too

force recent AIX versions to check for strerror_r

only output one line about the nodelay even if it fails

win32 doesn't need and even doesn't build if we extern declare sys_nerr

include the strerror.h file without curl_ prefix

strerror without prefix

cut off 'curl_' from the strerror file names

tcp-nodelay patch by Joe Halpin

so there are at least two different strerror_r() versions and our brand
new configure script detects them and now this code acts according to what
API that was detected

Tor Arntsen fixed how this is invoked