From: Havard Eidnes <he@netbsd.org>
Subject: CVS commit: pkgsrc/net/p5-RPC-XML
To: pkgsrc-changes@NetBSD.org
Reply-To: he@netbsd.org
Message-Id: <20081118153753.03007175D0@cvs.netbsd.org>
Date: Tue, 18 Nov 2008 15:37:52 +0000 (UTC)
Sender: pkgsrc-changes-owner@NetBSD.org
Precedence: list


Module Name:	pkgsrc
Committed By:	he
Date:		Tue Nov 18 15:37:52 UTC 2008

Modified Files:
	pkgsrc/net/p5-RPC-XML: Makefile distinfo

Log Message:
Updated from version 0.60nb1 to 0.64.

Pkgsrc changes:
 o Add commented-out HOMEPAGE using search.cpan.org.

Upstream changes:

0.64	Mon Sep 29 04:20:02 PDT 2008

	t/40_server.t:
	Further flexibility on the test of $srv->url(). This should (finally)
	catch any variation of 127.* address and/or host/domain naming that
	aliases to "localhost".

	lib/RPC/XML/Parser.pm:
	t/20_parser.t:
	Lingering problem from RT ticket #30354, RPC::XML::Parser did not get
	the change committed that included parser-level support for <i8>. Also
	added tests for parsing RPC::XML::request objects and all the data-type
	objects to the parser test suite (had they been there before, I would
	have caught this myself).

0.63	Fri Sep 19 02:22:26 PDT 2008

	lib/RPC/XML.pm:
	Forgot to increment $VERSION before the previous commit. Won't show up
	in CPAN as an available update otherwise.

0.62	Fri Sep 19 02:12:02 PDT 2008

	t/40_server.t:
	Extended the die message when $srv fails to allocate. Some testers are
	getting a failure here but I have no idea how or why.

	lib/RPC/XML.pm:
	Testers-service reports showed that the new i8 type was not auto-boxing
	correctly on true 64-bit machines. Turns out I had bogus values for
	the maximums and minimums, not just for 8-byte ints but also for plain
	4-byte ones as well.

0.61	Mon Sep 15 01:38:44 PDT 2008

	t/00_load.t:
	Converted to Test::More due to some cpan-testers reports that showed
	test 2 failing. Hoping this might provide better feedback if it
	continues to fail.

	t/40_server.t:
	Changed a test for "localhost:$port" in the server URL method to
	optionally allow localhost.localdomain, for those systems whose
	/etc/hosts cause the former to convert to the latter. Was causing
	false failures.

	t/40_server.t:
	t/50_client.t:
	t/60_net_server.t:
	Test suites armored against server processes that die as a result of
	croak() or other calls within 3rd-party modules. The client and
	net_server suites were also converted to Test::More at this time.

	t/40_server.t:
	Fix for RT ticket #36078, fix the URL testing on the server object to
	be less strict. Instead of only allowing "localhost", now also allows
	"localhost.localdomain" and does a look-up of "localhost" for the 127.*
	IP address and permits that as well. In essence, this extends and
	supercedes the second change listed above after more invariant cases
	were discovered.

	lib/RPC/XML.pm:
	t/10_data.t:
	Per RT ticket #30354 and specific request from a large agency using
	this package, implemented an "i8" type as an extension to the official
	XML-RPC specification. Documentation and tests included. THIS SHOULD BE
	CONSIDERED EXPERIMENTAL.

	t/40_server.t:
	t/60_net_server.t:
	Some of the work in armoring the test suites against server-death
	changed the counts on tests and on call-counts in the system.status
	method. Some explicit re-starts didn't help.


To generate a diff of this commit:
cvs rdiff -r1.3 -r1.4 pkgsrc/net/p5-RPC-XML/Makefile
cvs rdiff -r1.2 -r1.3 pkgsrc/net/p5-RPC-XML/distinfo

Please note that diffs are not public domain; they are subject to the
copyright notices on the relevant files.