31 May 17 20:34, you wrote to me:

 ac>> when compiled using Cygwin.  To date there has been no attempt to
 ac>> port it back to OS/2 or DOS.  Contributions are welcome though!  The
 ac>> latest version is stable under Linux, FreeBSD, BeOS and Mac OS X.

 RW> Hmm, what was the OS/2 version compiled with?  Now that OS/2 Warp is
 RW> back on the scene with ArcaOS 5, it might be feasible to bring the OS/2
 RW> port back 

I neglected to reply to this months ago, so here's a more comprehensive answer.

I have no interest in running ArcaOS, but I do finally have a properly working
virtual machine running OS/2 Warp 4.52 under VirtualBox. For several years I've
 been unable to get Warp 4.52 to connect to any SMB shares, making it pretty
inconvenient to do any development on, but just last month I belatedly took
another look at it and finally fixed the several issues I was having.

The DOS & OS/2 versions of timEd 1.10 (released by the original author Gerard
van Essen in December 1995) were built with Watcom C++ 10.6:

http://bbs.ozzmosis.com/tim110.zip   (DOS 16-bit)
http://bbs.ozzmosis.com/tim110x.zip  (DOS 32-bit)
http://bbs.ozzmosis.com/tim110p.zip  (OS/2 32-bit)

These versions contain Y2K bugs.

It (and NetMgr, etc) was open sourced in November 2000 (released as
ART_SRC.ZIP), and that code contains about a year's worth of changes from the
original author since his 1995 release, including (unfinished, IIRC) code to
port it to Windows NT:

http://bbs.ozzmosis.com/art_src.zip

I don't recall if it actually builds, and at this point it still has Y2K bugs.

Some time between 2001-2003 it was ported to Linux by Oliver Grimm, with Y2K
bugs removed, but IIRC remained unusable due to some outstanding bugs.

In 2003 Gerrit Kuehn sent me Oliver Grimm's code. I ported it to FreeBSD, fixed
 any serious bugs I found, then backported it to Linux and later OS X.

The issue with the UNIX fork now is that much of the DOS, OS/2 & NT code was
removed altogether (by either Oliver or me) to make the UNIX port easier and
would need to be carefully merged back in. I suspect most of this involves
screen writing and directory handling, but there may be enough other
differences between the two code bases for this to be difficult.

The UNIX fork also has a dependency on XMSGAPI, which could be an issue, though
 XMSGAPI already builds in OS/2 & Windows, so maybe not.

Interestingly the November 2000 source code release also includes the complete
source to timEd 1.00 (but not 1.01), from 1994-04-08 (TIM100S.ARJ).

Both the DOS & OS/2 versions of timEd 1.00 were built with Borland C++, not
Watcom.

--- GoldED+/BSD 1.1.5-b20170303
 * Origin: Blizzard of Ozz, Melbourne, Victoria, Australia (3:633/267)