From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of Bill Middleton
Sent: mardi 10 mai 2005 20:19
To: J閞鬽e Laban; mono-devel-list@lists.ximian.com
Subject: Re: [Mono-devel-list] Mono-1.1.7 - NetBSD 2.0.x Support

This looks promising.  Nice job.

A couple of suggestions:

1. Consider submitting your patch against the current SVN code.
 I will. I'm talking to Miguel about that.

2. Don't remove the existing *-*-*netbsd* case from configure.in.  Instead, change it to *-*-*netbsd*1* , and even better, see if you can get things working with 1.6. and update the existing entry. :)
I can leave the default netbsd. However, I am sure mono will never run properly on NetBSD 1.6 because of the lack of native threads support. I'll make the changes anyway.
 
3. Don't use __NetBSD__ in libgc (dyn_load.c)  libgc sets up definitions for NETBSD in gcconfig.h.  If there is a precompiler definition exclusive to NetBSD2 (I hope) then make a new definition, and || where appropriate, unless you can test against 1.6 simultaneously. (although I  know __thread wont work on 1.6).  It also looks like GC_NETBSD_THREADS are also specific to 2.x - perhaps GC_NETBSD2_THREADS?

I'm trying to leave the NetBSD 1.6 folks some room to work, is all. 
I'll try and make the change too but here too, I'm quite sure NetBSD 1.6 won't be supported.
 
4. The changes in mono/mini that you've made should be redone against SVN.  Most of them were taken care of with the FreeBSD changes, I hope.   Otoh, I'm pretty clueless about the registers on NetBSD - but are you sure you want to redefine:

+# define SC_EAX sc_eip
+# define SC_EBX sc_edx
+# define SC_ECX sc_eax
+# define SC_EDX sc_ecx

?  Those seem odd.
Indeed, these redefinitions are odd, but the register indices are not in the same orders under NetBSD, hence the shuffle. I'll check if this is still necessary, since it is a port from an 1.1.4 patch. About the FreeBSD changes, I'll check against the SVN although the 1.1.7 is not really that old.
 
Thanks Bill.
 
-- Jerome

Attachment: smime.p7s
Description: S/MIME cryptographic signature

Reply via email to