Re: [DotGNU]csunit Error on mingw

[James Michael DuPont]

--- Rhys Weatherley <address@hidden> wrote:
> James Michael DuPont wrote:
> 
> > > But I don't like the idea of forking pnet into a Mingw32 branch
> into
> > > a
> > > seperate CVS ... That is a potential synchronization nightmare ,
> if I
> > >
> > > ever saw one ... (and mdupont, it would be good to remember that
> it
> > > took me
> > > 5 full months between my first major patch and getting commit
> access
> > > ...).
> > 
> > What ever your decision is dude, you have to live with it.
> > If you force me to do too much work, then I will just put it into
> my
> > cvs. Really I dont want to this.
> 
> I had a look at the patch a couple of days ago, and it mostly
> seemed ok.  Any patch that makes the code more portable is good.
> 
> Some of it was to aclocal.m4 and such, which isn't relevant any more
> since I removed that from CVS.  

That is good, I we will need to review that again. The issue with
config guess is one, and the issue with the subpackages being
configured int configure.in is another. We need to be able to pass
extra port specific commands to the individual configure commands. I
will review the newest cvs soon.

>And the lstat/stat thing could be
> handled a little better with a configure.in check and "HAVE_LSTAT".
> "#ifdef __MINGW32__" should be avoided if possible.
> 
> But I didn't have a beef with the changes so far.  They seemed
> relatively small and innocuous.
> 
> > Just think, if you scare me off, you loose a developer. You might
> not
> > like my style, my ideas might scare you. But I am a body, and I
> have a
> > good brain. I am prepared to help.
> 
> We just need to take this in small steps.  Big mega-patches always
> scare me because it is difficult to understand the implications.
> But small one-liners to fix lstat portability or whatnot are very
> likely to be applied.

That is fine, and you have gotten what I have so far.

I mailed the patch to the list, which you have, and updated the sources
on the debian package repository.

As soon as the guy doing the mingw native posts his changes, I will
test and integrate them. But without a place to put them, and if he is
not setup with cvs, we need to have a place to share.

> 
> Send me the latest that you have, diff'ed against the live CVS.
> And I'll probably apply them now.
OK, I have not made any more changes since the original posting. Did
you get those? yes, I see you have. 

Please take what is good and integrate that, we need to get the changes
from the mingw port, then I will retest.

>  The command-line expand thing
> is likely to be the only thing that I might find "scary".  Surely
> there is some easy way of doing globbing under mingw32?

Yes, I have read about that, win32 has the standard globbing, it is in
a n standard windows object file, but I did not persue that, because I
dont want to link in an object file without sources.

let me look into this.

and rhys, I am fine if you will take the time to review and intergrate
the changes. What do you think about making a debian-mingw32 directory,
and adding in my new debian stuff there? 

As long as you are in the middle, we dont need direct cvs access right
now, in fact, you are playing the role in that case of a hyper
intelligent cvs server, that understand the code as well.

the only thing we need is a place to upload the files to for
sharing...(untill we all have cvs)

I just dont want to be stuck on the outside in the rain, trying to help
but locked out. That I why I am asking for rights/help. Help is better
than rights, so lets try that out for a while.

best regards,
mike

=====
James Michael DuPont
http://introspector.sourceforge.net/

__________________________________________________
Do you Yahoo!?
Yahoo! Mail Plus - Powerful. Affordable. Sign up now.
http://mailplus.yahoo.com