SBCL versus C++

Posted on 2015-08-06 by jcs

Drmeister has an interesting mini-benchmark that compares Common Lisp to C++ and Python. The real purpose of the post is to compare the performance of CCLASP Common Lisp to C++. CCLASP is a new Common Lisp system that drmeister is developing to tie together C++ and Common Lisp. The good news for him is that CCLASP has improved from being hundreds of times slower than C++ to just 4 times slower. It will be interesting to see where this ends up when the CCLASP compiler matures a bit more.

What I found most interesting about the post, though, was that SBCL was also tested in the benchmark and it outperformed C++. It’s a small test—calculating the 78th Fibonacci number 10,000,000 times—so it is by no means comprehensive. Still, it shows that the ridiculous FUD you still hear being dispensed about Lisp being slow is not true (and hasn’t been for years). See the post for the details, including the compile options that were used.

If you’re interested in the CLASP project, drmeister has a nice explanation for why he started it and what he hopes to accomplish. Judging from these intermediate results, he’s well on his way to a first rate Common Lisp implementation.

Posted in General | Tagged , | Leave a comment

Org Mode 8.3!

Posted on 2015-08-05 by jcs

As of my writing this, it hasn’t hit Elpa yet.

Posted in General | Tagged , | Leave a comment

Google and the Smoking Gun

Posted on 2015-08-05 by jcs

Regular readers know that I’ve been critical of Google in the past. But their business model doesn’t mean that some state attorney general with an agenda gets to use the power of government to do the bidding of the MPAA and movie studios. Nonetheless, it appears that that’s exactly what’s happening.

Knowledgeable observers have long assumed that Mississippi Attorney General Jim Hood has been carrying water for the MPAA and studios by pursuing fraudulent legal action against Google. Indeed, a federal judge has already ruled that Hood acted in bad faith to deny Google its constitutional rights. Google responded by subpoenaing documents from the MPAA and studios. Despite frantic resistance from the studios, Google has found a smoking gun in the form of an email trail between Hood’s staff and the MPAA.

The email, which is marked confidential, details plans to enlist various media outlets—including the Today Show and the Wall Street Journal—to place pieces critical of Google. If that fails, the memo says that Hood can issue a CID (basically a subpoena) written by the MPAA.

One hopes that when Google’s finished they’ll own a large portion of the MPAA’s bank account and that Hood will be removed from office. That’s probably too much to hope for but the collusion between Hollywood vested interests and an Attorney General to deprive a company of their constitutional rights should not be allowed to stand.

Added before publication: In the meantime, Hood refuses to take the hint.

Posted in General | Tagged | Leave a comment

Turning Off Annoying Magit Prompts

Posted on 2015-08-04 by jcs

As most of you know, I’m a big fan of Magit and have been using it for some time. One thing I didn’t like was that it asked for verification before staging all changed files (【S】 in magit-status) but being Emacs that was easily fixed. With the new Magit, that fix no longer works. We were warned that some of the variable names were changed so I thought, “No problem, I’ll just grep for the prompt string and see what the new magic incantation is.” Sadly, nothing was found. That’s because of the way prompts work in Magit 2.x; you have to add the prompt’s identifier to the magit-no-confirm list. For the stage all prompt, that identifier is stage-all-changes.1

I also noticed that Magit now wants me to confirm where I want to push my commits. There’s a long explanation of the reasons for this change in the doc string of the magit-push-always-verify variable. The TL;DR is that it’s an attempt to make everyone happy.

Here’s the code that configures Magit and eliminates both prompts: 

1: (use-package magit
2:   :ensure t
3:   :bind ("C-x g" . magit-status)
4:   :config
5:   (add-to-list 'magit-no-confirm 'stage-all-changes)
6:   (setq magit-push-always-verify nil)
7:   (setq magit-last-seen-setup-instructions "2.1.0"))

The important part for eliminating the Stage all changes? prompt is on line 5. You have to execute that after Magit is loaded. To get rid of the push verify prompt, use line 6.

Footnotes:

1

The prompt is derived from the identifier by removing the dashes, capitalizing the first character, and appending a ‘?’.

Posted in General | Tagged , | Leave a comment

A Diff on Quitting Emacs

Posted on 2015-08-03 by jcs

I used to know this and probably even blogged on it before but I’d forgotten about it and maybe you have too so it’s worthwhile repeating.

Posted in General | Tagged | Leave a comment

SBCL 1.2.14 Released

Posted on 2015-08-03 by jcs

The latest version of Steel Bank Common Lisp is available at the usual place. You can check the NEWS page for the changes in this iteration.

When I first tried to download it from Sourceforge, it wasn’t available and Sourceforge offered me version 1.2.13 instead. This despite the fact that the link from sbcl.org was specifically for the new version. I tried later and everything was OK. I wouldn’t have thought much of it except for the recent contretemps involving Sourceforge.

As usual, it compiled without incident and this time when I ran the regression tests there were no failures so the OS X problem in the last couple of releases has been fixed.

I really love SBCL and if you are looking for a good Common Lisp development system, you should definitely check it out. I say that every month, of course, but it’s true nonetheless.

Posted in General | Tagged , | Leave a comment

What Security Experts Do Differently

Posted on 2015-08-02 by jcs

Over at the Google Security Blog, Lulia Lon, Rob Reeder, and Sunny Consolvo have an interesting post on some research they did on what security experts do differently from non-expert to stay safe on the Internet. The post is the TL;DR for the longer (20 pages) paper on the research.

The most important thing you can do, according to the experts, is to keep your system up to date by installing updates but non-experts don’t do that and sometimes even consider it unsafe. The other big item experts do is to use a password manager, also resisted by non-experts. The password manager insures—or, at least, can insure—that two other important criteria are met: the use of unique, strong passwords. That’s hard without a password manager but trivial with one. When I set a password for a new site I just let my password manager choose a long, random, unique password. I don’t even bother looking at it because there’s no chance I could remember it—that’s what the password manager is for.

The fifth thing that experts recommend is to use two-factor authentication. Sadly, most sites don’t offer that but it’s obviously a big win for security.

Although the paper is a bit long, it’s interesting and fills in many details that the short post doesn’t cover. It’s also instructive to see the things that non-experts believe are effective in keeping them safe.

Posted in General | Tagged | Leave a comment

Come Over to the Dark Side

Posted on 2015-08-01 by jcs

and achieve great power.

Posted in General | Tagged , | Leave a comment

The Washington Post Gets it Right

Posted on 2015-08-01 by jcs

Last Wednesday, Irreal excoriated the Washington Post for their uninformed—ignorant really—editorial supporting backdoors in crypto systems allowing the FBI and other law enforcement agencies to read encrypted communications when they feel it’s necessary. My post and the TechCrunch article it pointed to was harsh so I’m delighted to point to a WAPO Op Ed that gets it right.

An opinion piece entitled Why the fear over ubiquitous data encryption is overblown, makes the case that on balance ubiquitous encryption allowing communication unreadable by anyone other than the intended parties is a good thing, the fears of the law enforcement community notwithstanding. That’s the libertarian/crypto-geek position, of course, so you might think the authors are from the Pirate Party or some other (by U.S. standards) fringe group. The authors, however, have surprising backgrounds. Mike McConnell was the Director of National Intelligence and a former director of the NSA. Michael Chertoff is the former Secretary of Homeland Security. William Lynn was a Deputy Secretary of Defense. These are not the usual suspects. They’re respected members of the government elite who understand the issues and might be expected to support the backdoors. Yet they’re taking on the FBI and others who insist that the fate of the nation depends on the government’s ability to spy on us.

Read the editorial. It’s not so much that it says anything you haven’t heard before but that it’s said by knowledgeable members of the establishment. We can only hope that congress will listen.

Posted in General | Tagged , | Leave a comment

Drag and Drop Into an Org Buffer

Posted on 2015-07-31 by jcs

John Kitchin has a very nice post in which he presents at bit of Elisp that enables us to drag and drop images into an Org buffer. The code is short and easy to understand. I just cut it out of his post, pasted it into my init.el, and evaluated it. After that I could drag and drop just as in his video demonstration. The only problem I had was with the 【Ctrl】 drag and drop. When dragging from the Finder, 【Ctrl】 clicking on an item brings up a menu. What you have to do is drag it over to the Org buffer and then press control before you release the mouse button.

Kitchin’s code was developed on a Mac and according to the comments has some problems with Linux machines. If you experience any trouble, you might want to try abo-abo’s drag and drop solution. If you’re interested, check out his demonstration video.

I really liked that when he first demonstrated it but, as I recall, there was some indication that it wouldn’t work on a Mac so I never tried it. After reading Kitchin’s post, I downloaded and tried abo-abo’s org-download to see how it compared to Kitchin’s solution.

I had a couple of glitches but to be fair I didn’t do more than a cursory trial. I could have probably resolved my problems with a bit of code reading. On the other hand, abo-abo’s solution allows you to grab images from Web pages while Kitchin’s code works only with images in the file system. Between the two solutions, you should find one that works for you and meets your needs.

Posted in General | Tagged , | Leave a comment