81204 - BSD/OS builds fail with --enable-crypto

Status: VERIFIED WORKSFORME

(SeaMonkey :: Build Config, defect)

Description

[lidl]

A while ago, I sent in patches to make the crypto stuff (NSS, et al)
compile under BSD/OS.

Last night, I attempted to turn on --enable-crypto, so I could finally
enjoy the wonders of browsing SSL enabled web sites.

Unfortunately, the build system for the main browser dosen't quite deal
with the build environment the same way that the NSS system does.
(I understand that this is being resolved.)

Anyway, the following attachement allows the system to build under
BSD/OS 4.2 with --enable-crypto turned on in .mozconfig file.

-Kurt

Comment 1

[lidl]

Attachment: patch for --enable-crypto building on BSD/OS 4.2

Comment 2

[cls]

The patch to configure should actually be against configure.in but I see what
you want.  Reassigning to javi since the security/ partition has been locked
down again.

Comment 3

[cls]

Index: configure.in
===================================================================
RCS file: /cvsroot/mozilla/configure.in,v
retrieving revision 1.858
diff -u -r1.858 configure.in
--- configure.in	2001/05/15 20:47:28	1.858
+++ configure.in	2001/05/17 00:22:44
@@ -449,20 +449,20 @@
 MOZ_COMPONENT_XPCOM_LIBS='-L$(DIST)/bin -lxpcom'
 
 if test -n "$CROSS_COMPILE"; then
-    OS_TARGET="${target_os}"
+    OS_TARGET=`echo $target_os | sed -e 's|/|_|g'`
     OS_ARCH=`echo $target_os | sed -e 's|/|_|g'`
     OS_RELEASE=
     OS_TEST="${target_cpu}"
     case "${target_os}" in
         linux*)       OS_ARCH=Linux ;;
         solaris*)     OS_ARCH=SunOS OS_RELEASE=5 ;;
-        openvms*)     OS_TARGET=`uname -s`
+        openvms*)     OS_TARGET=`uname -s | sed -e 's|/|_|g'`
                       OS_ARCH=`uname -s | sed -e 's|/|_|g'`
                       OS_RELEASE=`uname -r`
                       OS_TEST=`uname -Wh`;;
     esac
 else
-    OS_TARGET=`uname -s`
+    OS_TARGET=`uname -s | sed -e 's|/|_|g'`
     OS_ARCH=`uname -s | sed -e 's|/|_|g'`
     OS_RELEASE=`uname -r`
     OS_TEST=`uname -m`

Comment 4

[lidl]

Far be it for me to be a noodge -- but is this going to get reviewed
and checked in before the 0.9.1 freeze?

It would be really nice to have this work in the 0.9.1 release.

Comment 5

[Javier Delgadillo]

The changes to mozilla/security/manager/Makefile.in look fine to me.

Has everything else been checked in?  Or is it just the PSM 2 stuff that needs
to be checked in?

r=javi on the mozilla/security/manager/Makefile.in changes.

Comment 6

[lidl]

As far as I know, none of these changes have made it into the tree.
(Certainly my CVS pull from earlier today didn't have the fixes it in.)

I noticed that the top-level configure.in/configure changes were in
the tree briefly, but apparently got reverted back out.  Weird.

So, no, none of the changes are in the tree right now.  There are only
two files that need to be changed, and I guess now one of them has
been reviewed.  (Although cls doesn't need external review on build/config
issues, if I read those doc files correctly.)

Comment 7

[Javier Delgadillo]

cls:  If you give sr= here, I'll check this in when the tree opens.

Comment 8

[cls]

Won't these changes be superceded by the patch proposed by wtc in bug 81181 ? 
Anyway, sr=cls

Comment 9

[lidl]

Yes. If my reading of the proposed patches in bug 81181 are correct,
the suggested fixes for this bug are not going to be needed.
I wasn't aware of that bug's scope when I submitted this bug.

Comment 10

[lidl]

The checked in patches for bug 81181 solves the problem on BSD/OS
as well.  I'm going to close this bug, since I've verfied that
a build from this evening works for me.

Comment 11

[Jon Granrose]

verified.