> Good afternoon,
>
> On Saturday, January 9, 1999, Jason Salem wrote:
> >I've been working with the server for a bit, getting my "sea-legs", and
> >deciding how to do some of the things we plan on implementing. At first,
> >I'd run on linux, but lately I've been wanting to work with the windows
> >port. Since the current version has problems, I'd been playing
> around with
> >compiling under cygwin -- but the gdbm port behaves strangely
> under there,
> >requiring odd mounting behaviors that I haven't spent enough
> time to ferret
> >out completely. I also haven't gotten past the gdbm port part.
>
>
> I've heard about this error before but don't believe that someone has
> resolved it yet. It'd be interesting to see what would happen if the
> BerkeleyDB from SleepyCat were used in place of gdbm.
>
I've looked at the Sleepycat site, and their product looks good, but I
haven't been able to download it yet. I'm working with someone there to be
able to take a look at it. It would be nice to be able to use a DLL.
> >At this point, I'd like to ask if cygwin is something that you
> worked with
> >at any time to try porting, or if you had only worked with borland's
> >compiler on windows. If you'd tried it and discarded it, I don't want to
> >merely walk down that same path. If you haven't, then I may continue
> >experimenting with it to see if I can streamline it. Aside from the odd
> >mounting, it looks like the environment might have some merit.. running
> >configure and make with all the native-appearing unix tools (cp,
> rm, etc.)
> >and yet still having the ability to add windows-specific code (like the
> task
> >tray icon I've been told would be useful, or the ability to run as a
> service
> >under NT).
>
>
> Currently, the Win32 port is entirely done by a couple of other
> people, but
> coordinated by me, and lives in the CVS tree with the unix code.
> The reason
> for the use of the Borland compiler was only that was what people had on
> hand that were willing to do the work. VisualC++ doesn't implement
> opendir()/readdir()/closedir(), but they can be easily implemented given a
> short bit of time. I've been busy enough with work and with the unix
> versions of Genesis that I've been leaving the Win32 port in the hands of
> those interested in working on it.
>
> However, I'd like to see some of what you discuss, namely the
> ability to run
> as a service under NT. Also, allowing the source to be compiled under
> cygwin would be cool since that'd allow a larger number of people to work
> with the source. What is the task tray icon?
>
I don't have a problem with the Borland compiler, per-se, though I don't
have a copy of it (not since Turbo C for Windows). I have visual c++ 5, and
I've noticed that porting unix software to it is not something I enjoy.
Perhaps one portion of hell will consist of a pile of unix source code and
the MS compiler.
I decided to try cygwin because of the environment -- having bash etc. makes
it easier to take the code from a source snapshot and put out a port
quickly. Part of me balks at the thought of two sets of people porting to
the same platform with different compilers. I suppose the end test is which
works better and brings ports over more quickly.
And a task tray icon is probably not the "official" term.. but it's one of
the programs that when it runs generates an icon in the tray of the task bar
in the win95/NT4 interface. It can provide a simple way of seeing if a
service is running, and an ability to interface with that service easily.
> Feel free to cc: genesis@cold.org on your response. I didn't want to send
> your post there without checking with you first.
>
> - Bruce
>
>
|