[kaffe] kaffe is not compiling on cygwin

[Dalibor Topic]

Hi all,

Jim Pick wrote:
> On Fri, 5 Dec 2003 05:15:48 -0800 (PST)
> Michael Franz <mvfranz at yahoo.com> wrote:
> 
> 
>>--- Michael Nascimento Santos <mister__m at hotmail.com>
>>wrote:
>>
>>>I am trying to install kaffe 1.1.2 or kaffe 1.0.7 on
>>>cygwin, but I'm getting
>>>compiler errors with both versions. I'd like to know
>>>if these versions have
>>>been tested with cygwin and what I should do to make
>>>it work for me. I have
>>>googled through the archives but couldn't find an
>>>answer to my problem. Any
>>>help will be appreciated.
> 
> 
>>When was the last time cygwin was supported?  I was
>>looking at this the other day and thought it was funny
>>that I got this error.  Was support stopped or did the
>>build somehow change?
> 
> 
> Yep, Kaffe is broken on Cygwin.  
> 
>   http://www.kaffe.org/pipermail/kaffe/2003-November/044378.html
> 
> There has been work done on Cygwin support in the past, but I'm not sure
> what is a good version to work with.  I did get it to compile and run
> some stuff around the time of the 1.0.7 release, with the interpreter
> only, but it failed a whole lot of regression tests.   I'm sure it
> worked better at other times.
> 
> We sort of need somebody to "adopt" the port, and try to fix it. 
> Volunteers?
> 
> It's definitely something I want to see happen, I'll probably even do it
> myself someday, since I actually care about Windows somewhat (I
> occasionally play games on it, and my family uses it).  Actually, I'd
> like to see a mingw32 port working as well -- I did one back at
> Transvirtual with their in-house version of Kaffe.  Keep in mind, my
> timetable is pretty relaxed, so that may mean it'll be a year before I
> cycle around to it. :-)

I've got a patch that fixed the build for Cygwin *before* the change 
that broke all systems without signal context handling for good posted here:

http://www.mail-archive.com/kaffe@kaffe.org/msg02802.html

though I don't think I actually checked it in.

With respect to fixing the the build errors on platforms where we don;t 
have access to the exception/signalling information I proposed to get 
rid of the EXCEPTIONFRAME call here: 
http://www.kaffe.org/pipermail/kaffe/2003-August/043433.html

There wasn't much discussion, and I don't know if it would actually 
work. But it may be worth a try to try defining EXCEPTIONFRAME away as 
it is done in kaffe/kaffevm/systems/unix-jthreads/signal.c for the 
interpreter.

cheers,
dalibor topic