Andrew Ross http://blogs.gentoo.org/aross en-US http://blogs.law.harvard.edu/tech/rss 60 FBI fights zombie hordes http://blogs.gentoo.org/aross/2007/06/15/fbi_fights_zombie_hordes Fri, 15 Jun 2007 07:34:00 +0000 Andrew Ross ! Gentoo 1343@http://blogs.gentoo.org/ <p>Although <a href="http://news.bbc.co.uk/2/hi/technology/6752853.stm">the article itself</a> is a little lightweight, the headline is to die for, especially when viewed in a title-only RSS feed (such as the "Latest Headlines" live bookmark shipped by default with Firefox):</p> <p><img src="http://dev.gentoo.org/~aross/blog-support-files/fbi-fights-zombie-hordes.png" alt="News Headline: FBI tries to fight zombie hordes" /></p><div class="item_footer"><p><small><a href="http://blogs.gentoo.org/aross/2007/06/15/fbi_fights_zombie_hordes">Original post</a> blogged on <a href="http://b2evolution.net/">b2evolution</a>.</small></p></div> Although the article itself is a little lightweight, the headline is to die for, especially when viewed in a title-only RSS feed (such as the "Latest Headlines" live bookmark shipped by default with Firefox):

News Headline: FBI tries to fight zombie hordes

Original post blogged on b2evolution.

]]> http://blogs.gentoo.org/aross/2007/06/15/fbi_fights_zombie_hordes#comments Amazing arch teams! http://blogs.gentoo.org/aross/2007/04/25/amazing-arch-teams Wed, 25 Apr 2007 02:29:22 +0000 Andrew Ross Gentoo 1307@http://blogs.gentoo.org/ <p><a href="https://bugs.gentoo.org/show_bug.cgi?id=175814">Bug #175814 (dev-util/mercurial-0.9.3-r1 stable request)</a> is my first stabilization bug and, although it's a simple one, I'm still blown away by the rapid response from not one but five archs! Many thanks to all involved, and may all our stabilizations go as smoothly <img src="http://blogs.gentoo.org/rsc/smilies/icon_smile.gif" alt="&#58;&#41;" class="middle" /> .</p><div class="item_footer"><p><small><a href="http://blogs.gentoo.org/aross/2007/04/25/amazing-arch-teams">Original post</a> blogged on <a href="http://b2evolution.net/">b2evolution</a>.</small></p></div> Bug #175814 (dev-util/mercurial-0.9.3-r1 stable request) is my first stabilization bug and, although it's a simple one, I'm still blown away by the rapid response from not one but five archs! Many thanks to all involved, and may all our stabilizations go as smoothly :) .

Original post blogged on b2evolution.

]]> http://blogs.gentoo.org/aross/2007/04/25/amazing-arch-teams#comments 0-day bump requests http://blogs.gentoo.org/aross/2006/10/23/0-day-bump-requests Mon, 23 Oct 2006 11:42:27 +0000 Andrew Ross Gentoo 1034@http://blogs.gentoo.org/ <p><a href="http://planet.gentoo.org/developers/deathwing00/2006/10/16/just_another_0_day_bump_request">Ioannis is right</a> about the futility of filing 0-day bump requests.</p> <p>Just a few hours after Xen 3.0.3 was <a href="http://lists.xensource.com/archives/html/xen-users/2006-10/msg00473.html">announced on the xen-devel mailing list</a>, a version bump request was filed as <a href="http://bugs.gentoo.org/show_bug.cgi?id=151764">bug #151764</a>.</p> <p>As maintainers of a package, we're aware of when new relases are made - we're probably even following the development list and working on ebuilds for the release candidates. If a working ebuild doesn't appear in the tree within 3 hours of the upstream release announcement, it doesn't mean we need to be prodded! </p> <p>Oh, and since Chris White <a href="http://planet.gentoo.org/developers/deathwing00/2006/10/16/just_another_0_day_bump_request#c18320">didn't like Ioannis' choice of kitten</a>, I've taken the liberty of finding some <a href="http://images.google.com/images?q=beagle+puppy&amp;hl=en&amp;lr=&amp;client=safari&amp;rls=en&amp;pwst=1&amp;sa=X&amp;oi=images&amp;ct=title">puppies</a> and <a href="http://www.cutelittlekittens.com/kitten/6418.html">cute</a> <a href="http://en.wikipedia.org/wiki/Image:Stray_kitten_Rambo002.jpg">little</a> <a href="http://www.ratemykitten.com/pictures/000000960/11927/">kittens</a> to be sacrificed in response to this latest 0-day version bump request :-)</p> <p><a href="http://www.technorati.com/">Technorati</a> tags: <a href="http://technorati.com/tag/gentoo" rel="tag">gentoo</a>, <a href="http://technorati.com/tag/cute" rel="tag">cute</a>, <a href="http://technorati.com/tag/kittens" rel="tag">kittens</a>, <a href="http://technorati.com/tag/puppies" rel="tag">puppies</a></p><div class="item_footer"><p><small><a href="http://blogs.gentoo.org/aross/2006/10/23/0-day-bump-requests">Original post</a> blogged on <a href="http://b2evolution.net/">b2evolution</a>.</small></p></div> Ioannis is right about the futility of filing 0-day bump requests.

Just a few hours after Xen 3.0.3 was announced on the xen-devel mailing list, a version bump request was filed as bug #151764.

As maintainers of a package, we're aware of when new relases are made - we're probably even following the development list and working on ebuilds for the release candidates. If a working ebuild doesn't appear in the tree within 3 hours of the upstream release announcement, it doesn't mean we need to be prodded!

Oh, and since Chris White didn't like Ioannis' choice of kitten, I've taken the liberty of finding some puppies and cute little kittens to be sacrificed in response to this latest 0-day version bump request :-)

Technorati tags: , , ,

Original post blogged on b2evolution.

]]> http://blogs.gentoo.org/aross/2006/10/23/0-day-bump-requests#comments Danger, Will Robinson! Hardened and xen-tools-3.0.2-r4 http://blogs.gentoo.org/aross/2006/10/14/hardened-and-xen-tools Sat, 14 Oct 2006 05:54:16 +0000 Andrew Ross Gentoo Xen 1005@http://blogs.gentoo.org/ <p>The motive for my <a href="http://planet.gentoo.org/developers/aross/2006/10/07/elog-no-excuses">previous post</a> is about to be revealed - I've just committed <a href="http://sources.gentoo.org/viewcvs.py/gentoo-x86/app-emulation/xen-tools/ChangeLog?rev=1.18&amp;view=markup">app-emulation/xen-tools-3.0.2-r4</a>, as <code>~x86</code>/<code>~amd64</code>, and hardened users should heed the ebuild's warning:</p> <blockquote> <p>xend may not work when python is built with stack smashing protection (ssp). If 'xm create' fails with '&lt;ProtocolError for /RPC2: -1 &gt;', see bug #141866</p> </blockquote> <p>While -r4 includes fixes for <a href="http://bugs.gentoo.org/buglist.cgi?query_format=advanced&amp;short_desc_type=allwordssubstr&amp;short_desc=&amp;long_desc_type=substring&amp;long_desc=&amp;bug_file_loc_type=allwordssubstr&amp;bug_file_loc=&amp;status_whiteboard_type=allwordssubstr&amp;status_whiteboard=&amp;keywords_type=allwords&amp;keywords=&amp;bug_status=NEW&amp;bug_status=ASSIGNED&amp;bug_status=REOPENED&amp;bug_status=RESOLVED&amp;emailtype1=exact&amp;email1=&amp;emailtype2=substring&amp;email2=&amp;bugidtype=include&amp;bug_id=137137%2C148628%2C149138%2C151014%2C144057%2C143999%2C148486%2C147876%2151014&amp;chfieldfrom=&amp;chfieldto=Now&amp;chfieldvalue=&amp;cmdtype=doit&amp;order=Reuse+same+sort+as+last+time&amp;field0-0-0=bug_id&amp;type0-0-0=anyexact&amp;value0-0-0=">quite a few bugs</a>, for those using a hardened profile the biggest change is a relaxation of the restriction I introduced in -r3 to combat <a href="http://bugs.gentoo.org/show_bug.cgi?id=141866">bug #141866 (RPC Protocol Error with xen-tools and hardened profile)</a>.</p> <p>In response to this bug I made -r3 die if python was built with <a href="http://www.research.ibm.com/trl/projects/security/ssp/">SSP</a>, since it appeared that everyone using xen with a hardened profile was having the same problem. However, after -r3 hit the tree <a href="http://bugs.gentoo.org/show_bug.cgi?id=141866#c5">some</a> <a href="http://bugs.gentoo.org/show_bug.cgi?id=141866#c8">users</a> <a href="http://bugs.gentoo.org/show_bug.cgi?id=141866#c9">reported</a> that they were unaffected by this problem, hence the relaxation in -r4.</p> <p>If you're using xen on a hardened system please test xen-tools-3.0.2-r4 and let me know (via the <a href="http://bugs.gentoo.org/show_bug.cgi?id=141866">above-mentioned bug</a>) if your 'xm create' fails with '&lt;ProtocolError for /RPC2: -1 &gt;' or not.</p> <p><a href="http://www.technorati.com/">Technorati</a> tags: <a href="http://technorati.com/tag/gentoo" rel="tag">gentoo</a>, <a href="http://technorati.com/tag/hardened" rel="tag">hardened</a>, <a href="http://technorati.com/tag/xen" rel="tag">xen</a> </p><div class="item_footer"><p><small><a href="http://blogs.gentoo.org/aross/2006/10/14/hardened-and-xen-tools">Original post</a> blogged on <a href="http://b2evolution.net/">b2evolution</a>.</small></p></div> The motive for my previous post is about to be revealed - I've just committed app-emulation/xen-tools-3.0.2-r4, as ~x86/~amd64, and hardened users should heed the ebuild's warning:

xend may not work when python is built with stack smashing protection (ssp). If 'xm create' fails with '<ProtocolError for /RPC2: -1 >', see bug #141866

While -r4 includes fixes for quite a few bugs, for those using a hardened profile the biggest change is a relaxation of the restriction I introduced in -r3 to combat bug #141866 (RPC Protocol Error with xen-tools and hardened profile).

In response to this bug I made -r3 die if python was built with SSP, since it appeared that everyone using xen with a hardened profile was having the same problem. However, after -r3 hit the tree some users reported that they were unaffected by this problem, hence the relaxation in -r4.

If you're using xen on a hardened system please test xen-tools-3.0.2-r4 and let me know (via the above-mentioned bug) if your 'xm create' fails with '<ProtocolError for /RPC2: -1 >' or not.

Technorati tags: , ,

Original post blogged on b2evolution.

]]> http://blogs.gentoo.org/aross/2006/10/14/hardened-and-xen-tools#comments elog - endless possibilities and no excuses. http://blogs.gentoo.org/aross/2006/10/07/elog-no-excuses Sat, 07 Oct 2006 02:33:26 +0000 Andrew Ross Gentoo 983@http://blogs.gentoo.org/ <p>I don't believe in holding the hands of users running <em>~arch</em> - even if they only have a few <em>~arch</em> ebuilds on an otherwise stable system. While <strong>ALL</strong> Gentoo users should be reading the messages delivered via <code>ewarn</code>, for <em>~arch</em> users this can be critical.</p> <p>With the introduction of the elog logging framework in portage-2.1.1, there's no excuse for missing these messages. <a href="http://gentoo-wiki.com/Make.conf.example">make.conf.example</a> shows how to automatically save ebuild messages to disk, or email them to an address of your choosing. <a href="http://bugs.gentoo.org/show_bug.cgi?id=136496">elogviewer</a> provides a <a href="http://maxima.wordpress.com/files/2006/06/elogviewer.png">nice GUI</a> for reviewing messages, while <a href="http://www.gentoo-portage.com/app-portage/portage-mod_jabber">portage-mod_jabber</a> allows you to receive them via XMPP (Jabber).</p> <p>Should you want to, there's no reason why you couldn't easily receive ebuild messages via SMS or pager - they could even be posted to a blog using the <a href="http://en.wikipedia.org/wiki/MetaWeblog">MetaWeblog API (MWA)</a>, or sent directly to your workstation screen using <a href="http://www.pykota.com/software/pykoticon">PyKotIcon</a> or <a href="http://remotegrowl.erlang.no/">Remote Growl</a>.</p> <div class="item_footer"><p><small><a href="http://blogs.gentoo.org/aross/2006/10/07/elog-no-excuses">Original post</a> blogged on <a href="http://b2evolution.net/">b2evolution</a>.</small></p></div> I don't believe in holding the hands of users running ~arch - even if they only have a few ~arch ebuilds on an otherwise stable system. While ALL Gentoo users should be reading the messages delivered via ewarn, for ~arch users this can be critical.

With the introduction of the elog logging framework in portage-2.1.1, there's no excuse for missing these messages. make.conf.example shows how to automatically save ebuild messages to disk, or email them to an address of your choosing. elogviewer provides a nice GUI for reviewing messages, while portage-mod_jabber allows you to receive them via XMPP (Jabber).

Should you want to, there's no reason why you couldn't easily receive ebuild messages via SMS or pager - they could even be posted to a blog using the MetaWeblog API (MWA), or sent directly to your workstation screen using PyKotIcon or Remote Growl.

Original post blogged on b2evolution.

]]> http://blogs.gentoo.org/aross/2006/10/07/elog-no-excuses#comments