184523 - support Solaris Packaging in NSS 3.3.3

Status: RESOLVED FIXED

(NSS :: Build, enhancement, P1)

Description

[Kirk Erickson]

We need support for Solaris Packaging in NSS 3.3.3.
The differences between NSS_3_3_2_RTM and
NSS_3_3_2_SUN_PKG_BRANCH
NSS_3_3_2_RTM need to checked into
NSS_3_3_BRANCH.

Related bug: http://bugzilla.mozilla.org/show_bug.cgi?id=158683

Comment 1

[Wan-Teh Chang]

This is fine.  We should review the changes one last
time before landing them on the NSS_3_3_BRANCH.

Kirk, could you generate a patch?  Make sure you use
the -uN option for cvs diff.  This generates a unified
diff with the new files included.

Comment 2

[Kirk Erickson]

[e-mail] (Ty Lee) is generating the patch.
Sonja or I will attach it for review.  I attempted to
add Ty to the cc'list, but apparently he has no account.

Comment 3

[Kirk Erickson]

Attachment: proposed patch for NSS_3_3_BRANCH

Before apply, create new subdirs:
mkdir -p mozilla/security/nss/pkg/solaris/common_files/
mkdir -p mozilla/security/nss/pkg/solaris/SUNWtlsx
mkdir -p mozilla/security/nss/pkg/solaris/SUNWtls
patch <XXX.patch

I tested applying the patch on a different machine and
successfully built with all combinations of:

    BUILD_OPT=1
    USE_64=1

Ready for review by Wan-Teh.

Comment 4

[Sonja Mirtitsch]

We also need to generate a prototype file with the solarispackage, this has been
missed in the original implementation of the package for NSS 3.3.2 the prototype
is not crucial to the package itself, but it seems to be absolutely necessary to
generate a patch to the package. The combination of prototype_com and
prototype_sparc (or prototype_com and prototype_i386) is not accepted by buildpatch.
sorry about noticing this so late.

Comment 5

[Wan-Teh Chang]

Comment on attachment 108990 [details] [diff] [review]
proposed patch for NSS_3_3_BRANCH

Kirk,

You can go ahead and check this patch in on the
NSS_3_3_BRANCH.  I am qualified to review the
new files in mozilla/security/nss/pkg.	After you
check in this patch I can verify that the new
files are the same as the versions on the
NSS_3_3_2_SUN_PKG_BRANCH with the exception of the
version number.

Comment 6

[Wan-Teh Chang]

Sorry, that should read "I am NOT qualified to
review the new files in mozilla/security/nss/pkg".

Comment 7

[Sonja Mirtitsch]

--- pasting in Ty's email below. I'll check it in with the suggestions that Ty
has made. This will be a partial fix for this bug, because Ty is still working
on getting the prototype into the format that solaris buildpatch requires.
----
I have reviewed the NSS patch. It looked fine to me except the following
modifications must be made:
 

   1. In mozilla/security/nss/pkg/solaris/Makefile.com

              PRODUCT_VERSION = 3.3.2
              PRODUCT_NAME = NSS_3_3_2_RTM
              --- changed to --->
              PRODUCT_VERSION = 3.3.3
              PRODUCT_NAME = NSS_3_3_3

   2. Files under PKG has the CVS $Id expanded with information in
NSS_3_3_2_SUN_PKG_BRANCH. Please have them removed.
            Here is an example:               diff -b -r
mozilla/security/nss/pkg/solaris/SUNWtlsx/prototype_sparc
                   
/data/nss33Branch/mozilla/security/nss/pkg/solaris/SUNWtlsx/prototype_sparc
                    5c5
                    #ident        "$Id: prototype_sparc,v 1.1.2.1 2002/09/06
00:45:29 kirk.erickson%sun.com Exp $"
                    --- changed to --->
                    #ident        "$Id: $"

Thanks,

Ty

Comment 8

[Kirk Erickson]

Sonja - can we mark this fixed?

Comment 9

[Sonja Mirtitsch]

no, don't close it yet, there is another mistake in the packages, in that we do
not depend on specific versions. 

Comment 10

[Kirk Erickson]

Attachment: Earlier mail from Danek regarding Sun's dependency policy

Michael raised the question of introducing package dependencies
back in October.  I've attached Danek's response to the question
for reference.	Basically, the Sun convention is to bank on
linker smarts:

"The way it was explained to me was that this kind of dependency information
has, empirically here at sun, been shown to be better implemented as a
run-time dependency -- i.e., through proper use of linker hints and
dynamic, programmatic checks of the required interfaces (and stepping down
functionality as necessary).  We've found that this is far more robust
against all the weird things that customers end up doing to your packages."

Comment 11

[Kirk Erickson]

Not sure if this is done or not.
Sonja wanted to deal with package dependencies.
Reassigned to her.

Comment 12

[Sonja Mirtitsch]

Reassigning to Michael for evaluation.

Comment 13

[Wan-Teh Chang]

Marked the bug fixed because the work to support Solaris packaging
is done.

Please open a new RFE if we still need to add package dependencies
(see comment 8 and later).