>> But, to your main concern, I did absolutely nothing that wasn't
>> surrounded by #ifdefs.
> :-) I'm also concerned that things are properly #ifdef HAVE_FEATURE and not > #ifdef THIS_PARTICULAR_SYSTEM
I have indeed. And I'll configure/compile it under Cygwin here, Solaris/gcc at work, and Digital UNIX on my ISP before I turn it back in. I gave up early (perhaps too early) on Cygwin/mingw32 (b20 with Mumit Khan's patches) -- and I believe that tree has now unforked, so they're the same codebase again. I may give that a try. I've seen some tests in configure files for cygwin, but -- perhaps because of the former fork -- they haven't worked too well, though this has shown up mostly on C++ which (thank God) we haven't got.
Unfortunately, that means somebody has got to write the code in configure that tests for cygwin, mingw32, getpwuid(), getopt(), etc.
I have this sinking feeling that I know who that somebody is. ;-)
Which is why I asked earlier about m4. I don't think we want to change configure, but rather what goes in one end of autoconf so that configure comes out the other end. Or I assume that, since my ignorance about autoconf, etc. is only exceeded by my ignorance of how my wife thinks. Hopefully, curing the former won't take 33 years, which is how long I've worked on the latter problem without finding a solution. ;-)
Btw, on another topic, I discovered that TheBat!, which is otherwise a very admirable email client for Windows, puts numbers by its Re:'s (e.g., the subject line above).
Sigh. Looks like something else we have to cope with. I'll take it on, since I'm the one who's made the archives look crummy. Basically it boils down to s/Re(.*):/Re:/ followed by s/Re: Re:/Re:/ iterated. That may cure a couple of other email client peculiarities as well (e.g., "Re: RE:", which I haven't seen in the wild, but suspect might happen when somebody with a proper email client responds to a reply from an Exchange user).
Btw, I may not have access to MSVC++ for a while at work, so I'm definitely looking for a volunteer who's got that to give it a try when I get the Windows build for hypermail ready for tentative release.
Incidentally, I think we should definitely consider supplying the Win32 binaries for hypermail, but I do *not* want to see us get into the situation I've seen in other open source projects where the Win32 version is a separate branch (and therefore requires separate hand maintenance) from the Unix version. That is, I think I should keep going along the path we've discussed in this thread, but just plan to release a binary version for folks who, due to their unfortunate choice of OS, aren't able to cope with distributions in source form.
I'll volunteer to do the Win32 builds for the forseeable future, since LCCWin32 produces smaller binaries than MSVC++.
-- Rev. Bob "Bob" Crispen crispen at hiwaay dot net 666k - Retirement plan of the BeastReceived on Wed 03 Jan 2001 01:48:23 AM GMT
This archive was generated by hypermail 2.3.0 : Sat 13 Mar 2010 03:46:12 AM GMT GMT