Daniel Stenberg [Wed, 19 Mar 2003 09:09:40 +0000 (09:09 +0000)]
Rename getdate.c to getdate.c.cvs, since the "normal" build procedure do
imply that yacc/bison exists and can generate this file. Those without one
of those tools can then checkout and rename the getdate.c.cvs file.
Daniel Stenberg [Sat, 15 Mar 2003 16:51:43 +0000 (16:51 +0000)]
Sort out the ENGINE problems people seem to be having. Now we put all ENGINE
related stuff within HAVE_OPENSSL_ENGINE_H and we don't make any private
typedef or similar if the header is missing...
Daniel Stenberg [Sat, 15 Mar 2003 16:43:58 +0000 (16:43 +0000)]
* use the pid returned back from test-servers and kill them before starting
them the first time
* verify that the server we start really comes up fine and works as
expected before continue
* count test cases where the server can't be run (for whatever reason)
* prefix lots of messages with RUN: to make it easier to realize which script
is saying what when running tests verbose
* remove the generic sleep(1) from each test, makes the suite fly! ;-)
I hope these changes will make the tests run somewhat more reliably on more
platforms.
Daniel Stenberg [Thu, 13 Mar 2003 23:02:33 +0000 (23:02 +0000)]
No longer halts operation if select or socket are missing, since in most
cases this is wrong... and if they're truly missing, we won't succeed to
link later on anyway.
Daniel Stenberg [Tue, 11 Mar 2003 18:55:34 +0000 (18:55 +0000)]
Richard Gorton improved the random_the_seed() function for systems where
we don't find/know of a good random source. This way, we get a better
randomness which in turn should make SSL connections more secure.
Daniel Stenberg [Mon, 10 Mar 2003 14:52:33 +0000 (14:52 +0000)]
AAAARG
libtool 1.4.3 is scary as hell and caused just about every build on all sorts
of platforms to stop working, thanks to the fact that it ruquires a SED
variables somehow set by the configure script. It works fine on my linux
running autoconf 2.57 and automake 1.7 but others seem not to do as fine.
Reverting back to the ltmain.sh we had previously, which I believe is 1.4.2
including handmade patches for FreeBSD.
ALERT ALERT ALERT before we try 1.4.3 or similar versions again, check the
${SED} stuff and similar carefully.
Daniel Stenberg [Mon, 10 Mar 2003 12:25:32 +0000 (12:25 +0000)]
Include sys/types.h as well. Ray DeGennaro reports successful compiling on
AIX when this fix is applied and I cannot see how this will break any
systems.
Daniel Stenberg [Mon, 3 Mar 2003 22:30:25 +0000 (22:30 +0000)]
Detect AIX 4.3 or later, and if found disable the check for the thread-safe
*_r() functions as they're not needed (and if fact mess things up for us).
Brought to our attention by the friendly Troels Walsted Hansen in bug report
#696217.
Daniel Stenberg [Mon, 3 Mar 2003 06:40:36 +0000 (06:40 +0000)]
Add (void) on our uses of the swrite() macro when we don't read the return
code as this makes compiler warnings. We *should* fix the code to deal with
the return codes instead...
Daniel Stenberg [Sun, 2 Mar 2003 17:43:42 +0000 (17:43 +0000)]
Init postdata properly before issuing a request, so that there isn't any
lingering POST-stuff that confuses GET requests. Juan F. Codagnone reported
this problem in bug report #653859.
Daniel Stenberg [Thu, 27 Feb 2003 12:50:54 +0000 (12:50 +0000)]
It appears that there are FTP-servers that return size 0 for files
when SIZE is used on the file while being in BINARY mode. To work
around that (stupid) behavior, we attempt to parse the RETR response
even if the SIZE returned size zero.
Debugging help from Salvatore Sorrentino on February 26, 2003.
Daniel Stenberg [Wed, 26 Feb 2003 12:42:25 +0000 (12:42 +0000)]
No longer loop to read multiple times before returning back from the transfer
function, as this could easily end up looping for a very long time (more or
less until the whole transfer was done) and no library-using app would want
that.
Daniel Stenberg [Mon, 24 Feb 2003 16:53:53 +0000 (16:53 +0000)]
Fixes to bring back the the "Expect: 100-continue" functionality. If the
header is used, we must wait for a 100-code (or timeout), before we send the
data. The timeout is merely 1000 ms at this point. We may have reason to set
a longer timeout in the future.