Blog posts tagged "perl"

Sharing from within Google Reader

January 4th, 2008

Collapsing the GTalk buddy list, and Reader sharing list was a serious blunder, and one that could use a bit more ink spilled about it. But one click sharing is one of my favorite Reader features.

GData Won’t Save You

Except there is a bit of a problem. I don’t really want to share with other Google Reader users, I’m not even sure I’m destined to be a long time Reader user. I want to share links the way I’m already doing it, through

No problem, Reader has an Atom feed of shared items. A really good feed, with the source info maintained, well formed, nicely done. Simplest thing in the world to parse the feed, and write the entries back to And I can tag any post in Reader, which is perfect, easy Ajaxy sharing into with a few minutes work.

Except for reasons I can’t fathom Reader isn’t including my tags in the Shared Items feed. Which all of a sudden makes my data feel a bit more locked up and trapped then I’d really like.

For Our Sins

Casting around a bit for a solution, I noticed the “Email” button, which allows me to send a link via email, along with a short note, and so “Email to” was born.

Super quick and dirty Perl script that:

  1. Parse the Google Reader HTML email for the relevant URL (no semantic markup, alas)
  2. Pull the link description from the subject
  3. Look for a line beginning “tags: ” followed by a space separated list of tags.
  4. Look for a line beginning “note: ” for the extended description.

Add the following rule to /etc/aliases file, and away you go.

to_del: | /home/you/

Takes 10-15 seconds vs 1 second to share, but much more flexible.

And Perl is still unbeatable when it comes to these kind of scripts.

Status Update: XML::RSS, Magpie, cvs2rss, wp-agg, etc

May 28th, 2005

Dan from GeekUprising is now the official maintainer of XML::RSS. I imagine his first major change will be to incorporate the patch he sent me last Fall splitting the RSS creation logic from the RSS parsing logic (while still being transparent to people who don’t want that kind of thing). Congratulations, and welcome.

cvs2rss has been missing since the last server move, but well placed nudge from Phil got me to dig up the archives today, so voila, as they.

My aggregator for WordPress, affectionately known as wp-agg is officially superseded by FeedWordPress. Charles has done a great job growing the idea to something useful to someone other then me, and polished it off by actively maintaining it. Nice work. I’ll be switching the Magpie blog as soon as I get a chance.

Sourceforge stats after a 6 month hiatus (or was it 18 months? now I’m forgetting), are finally back, now with largely useless graphs. I had a hunch that usage had continued to grow and it has, with 5,000 page views a day, and 6,000 downloads a month. Wow. Really need to revive the website to handle the enlarged community. Need a good domain name first.

You know, now that I’m a contractor again, if someone wanted to hire me for 20-25 hours of work I’ve got a list of features that I’d love to have time to roll out as MagpieRSS version 0.8.

Tagged: Uncategorized , , , , , ,

XML::RSS Free to Good Home

April 5th, 2005

I’ve been trying to give away XML::RSS for a long time now. Private emails, comments, etc have gone nowhere. The drive towards a 2.0 version based on libxml2/DOM, and a rich object model stalled. (though I’ve got a nascent version I use privately for scripts like my rss2delicious, and the RNC protest infoline).

XML::RSS has a number of fatal flaws that make it hard to grow, and hard to support. Its handling of character encoding will always be fatally flawed (its a hard problem), and its parser model has been evolving over many many years.

It is also the oldest RSS parser around, hugely popular, widely used, and with an amazing number of really excellent patches sitting in the queue. Work has been done to fix the encoding issues, re-factor out the RSS writing, and generally spiff it up. And I’m the road block to that all getting rolled out. I’ve got my solution, and I’m not having fun hacking on XML::RSS, time to pass it on.

Want it?

Tagged: Uncategorized ,

Craig on ChangeThis

February 8th, 2005

My buddies at ChangeThis got Slashdotted last Friday, and apparently held up well under the load. (damn, I wish we had actually finished the re-write of we started that code base for) And today they have, “Why Craigslist Works”, by Craig. I haven’t quite gotten over the conceptual leap of having to dowload a PDF, but I’m noting this for when life is a little less crazy.

Tagged: Uncategorized , ,

Ruby and Perl

December 2nd, 2004

Spent a little more time looking at Ruby tonight, and at least looking at the basics of the language it is very striking how Perl like it is, so much so that at times the changes in syntax feel arbitrary. (is %w{} really nicer then qw{}?)

Looking at it (but not really doing much coding) I’m struck that Ruby can be even more punctuation heavy then Perl, which really contributes to the “it looks like line noise” critique.

But my favorite quote from Prog Ruby so far is:

In general, there’s a move away from some of the Perlisms in the Ruby community. If you run your programs with the -w flag to enable warnings (you do run with warnings enabled, don’t you?), you’ll find the Ruby interpreter catches most of them. (ed. emphasis mine)

-w certainly looks like a Perlism to me. :)

Tagged: Uncategorized , , ,

Migrating Kwiki metabase/rcs

October 23rd, 2004

A part of the’s recent move to a new server, we’ve moved towards less custom compiled packages, and a more standard Debian install. One of those changes was that apache no longer runs as the user ‘apache’, but rather as ‘www-data’. Well and fine, except Wifimug is using rcs to provide versioning, and every file is strictly locked by the user apache, who no longer exists. What to do?

First mass unlock all the files with:

sudo rcs -u *,v

Then mass re-lock with:

sudo -u www-data rcs -l *,v

Hat tip to Nat for pointing out the problem.

Tagged: Uncategorized , ,