foo --bar

becomes “foo —bar”, which can look very similar to foo --bar if your contact is using a terminal that supports variable-width glyphs for exotic character points.

It’s a quick fix: stop double-hyphens being transformed to a long dash by right clicking the Adium input field, and unticking Substitutions > Smart Dashes from the menu.

Getting nice prompts is a common need and there is plenty of information out there on setting your PS1, but for various reasons I’ve had problems finding a ‘nice’ setup that’s just worked exactly the way I’ve wanted it to with titles. Here are the pertinent parts of my working configuration:

• I get the working directory and current command in the title
• I get the title in the screen windowlist if running in screen
• I get the same title in the WM_NAME of an rxvt or xterm, or the titlebar of an iTerm2, whether or not screen is involved

Here’s how I set my PS1 in .zshrc:

case "$TERM" in
  # [skipping some esoteric terminal emulators...]

  screen|screen.rxvt)
     # Set a coloured prompt
     PS1=$'%{\e[00;32m%}%*%{\e[00;34m%}%2~ %# %{\e[00m%}'
     ;;
  rxvt|rxvt-unicode|xterm|xterm-color)
     # Set the title, and a coloured prompt containing some useful info
     PS1=$'%{\e]0;%-3~\a\e\[00;32m%}%*%{\e[00;30m%}!%! %{\e[00;34m%}%2~ %# %{\e[0m%}'
     ;;
esac

I want an abbreviated form of the full command line so that it will fit in awesome's tasklist. I set the titles before a command is run in zsh's preexec hook. The screen window title and the xterm title are set with two different escapes.

function preexec() {
  local a=${${1## *}[(w)1]}  # get the command
  local b=${a##*\/}   # get the command basename
  a="${b}${1#$a}"     # add back the parameters
  a=${a//\%/\%\%}     # escape print specials
  a=$(print -Pn "$a" | tr -d "\t\n\v\f\r")  # remove fancy whitespace
  a=${(V)a//\%/\%\%}  # escape non-visibles and print specials

  case "$TERM" in
    screen|screen.*)
      # See screen(1) "TITLES (naming windows)".
      # "\ek" and "\e\" are the delimiters for screen(1) window titles
      print -Pn "\ek%-3~ $a\e\\" # set screen title.  Fix vim: ".
      print -Pn "\e]2;%-3~ $a\a" # set xterm title, via screen "Operating System Command"
      ;;
    rxvt|rxvt-unicode|xterm|xterm-color|xterm-256color)
      print -Pn "\e]2;%m:%-3~ $a\a"
      ;;
  esac
}

This requires the title to be reset for screen (since it’s not reset by PS1), using zsh’s precmd hook:

function precmd() {
  case "$TERM" in
    screen|screen.rxvt)
      print -Pn "\ek%-3~\e\\" # set screen title
      print -Pn "\e]2;%-3~\a" # must (re)set xterm title
      ;;
  esac
}

The beauty of this method (unlike methods that bind screen’s status line escapes to xterm Operating System Command escapes) is that no termcapinfo hacks in .screenrc are necessary.

Update: changed preexec to use zsh subscripting rather than splitting by IFS. Fixes duplication of 0-argument commands, e.g. “top top”.

I went through the HTML suggestions from the google webmaster tools and Google’s own SEO Starter Guide. I’ve pushed the search engine optimized cgit to my seo branch on github. You can see it in action at my git repositories. I’m testing all of this using an Apache ScriptAlias directive, I’m hoping it will still work alright with whatever other URL-processing schemes cgit supports. A short summary of the new SEO features so far:

• Use HTML h1 and h2 heading tags instead of custom-styled divs
• Much better title tags; commits have the commit subject, and the repo name has been added in a lot of places to avoid duplicate titles
• The bread-crumb has been integrated into the heading
• A configurable option to set nofollow relationships on links to non-HEAD commits, to avoid duplicate content being indexed

Of course, you could take the popular option of just using github instead of self-hosting your own git web interfaces… but even they don’t do quite a good a job IMO, they use the SHA1 in the web page titles, eww!

In RFC 1123 it is stipulated that:

a DNS resolver or server that is sending a non-zone-transfer query MUST send a UDP query first.

However, if you are doing a DNS UPDATE you really want the reliability that TCP offers, even if you don’t expect truncation to be an issue. The update is sent to the relevant authority server, so the arguments about load on root servers in the RFC aren’t applicable.

I’ve made the UDP implementation retry by default, but I think if you need more than 2 retries, you should be considering using TCP with its (much more advanced) retransmission algorithms.

Peter also found a bug in glibc’s res_send (actually in their send_dg function) whereby the resolver interprets the lack of the DNS “recursion available” flag in the header as an error. However, that flag isn’t even meaningful for DNS UPDATE responses; according to RFC 2136, those bits:

Should be zero (0) in all requests and responses. A non-zero Z field should be ignored by implementations of this specification.

As a result, glibc was setting errno to ECONNREFUSED or ETIMEDOUT even when the update was successful. I’ve hacked dudders to double-check after res_send, but it’s making me question the wisdom of using res_send at all, given that I’m constantly working around it.

Update: submitted glibc bug report #11950

To get dudders-1.04 on OpenWRT, simply update the official package feed and select dudders from the Net > DNS > dudders menu in the buildroot config. For systems other than OpenWRT, you can grab the source from sourceforge, or even github.