Life’s but a walking shadow…

The <oXygen/> documentation has an example of setting up an “External Tool” to run Ant. The example is simple enough to illustrate its point, but there’s more that can be done, especially if you write the Ant build file knowing that it will be run from <oXygen/>.

This example is from the <oXygen/> “project” that I used for organising the exercises for my “Testing XSLT” tutorial at XTech 2008. Read the rest of this entry »

Running some JUnit tests with Ant 1.6.5 gave this error when the <junitreport> task ran:

build.xml:160: The following error occurred while executing this line:
build.xml:367: Could not find a valid processor version implementation
from net.sf.saxon.TransformerFactoryImpl

I know of two possible solutions:

• Upgrade to using Ant 1.7.0, or
• Set ANT_OPTS thus:

  declare -x \
  ANT_OPTS=-Djavax.xml.transform.TransformerFactory=com.sun.org.apache.xalan.internal.xsltc.trax.TransformerFactoryImpl

Concepts, Techniques, and Models of Computer Programming (ISBN 0-262-22069-5) is a big book at 900+ pages, and it covers a lot of ground. I expect it will take about two years to get through it, depending on how many of its exercises I do and how many other books I read at the same time.

It is natural to compare this book to Structure and Interpretation of Computer Programs (SICP) (ISBN 0-262-51087-1). That is the book that I still wish I’d first read in 1981 rather than in 2001. This book is not giving the same aha! moments (maybe just because I have read SICP). This book may in the end be of more practical use than the mind-expansion induced by SICP, if only because this book covers constraint programming, which I will find useful for xmlroff.

Now, the programming concepts book that I really want would be the successor to Lisp in Small Pieces (ISBN 0-521-56247-3), but AFAICT, it hasn’t been finished.

I have been selected to present my “Testing XSLT” tutorial at XTech 2008 in Dublin in May.

Since I do a lot of referring to Trac ticket numbers in email messages, most visibly on the xmlroff-list mailing list, I wrote an Emacs function that finds the Trac ticket references in the current buffer and inserts a sorted list of the ticket references and their URLs at the end of the buffer.

The code (if WordPress hasn’t done too much damage to it) is:

(defvar trac-base "http://xmlroff.org"
  "Base to use when inserting links to trac tickets in email.")

(defun trac-base-xmlroff ()
  (interactive)
  (setq trac-base "http://xmlroff.org")
  (message trac-base))

(defun insert-trac-links ()
  "Insert links for Trac links."
  (interactive)
  (let ((ticket-alist)
         (changeset-alist '()))
    (goto-char (point-min))
    (while
        (re-search-forward
         "#([0-9]+)" nil t)
      (setq ticket-alist
              (add-to-list 'ticket-alist
 		 (string-to-number
 		  (buffer-substring-no-properties
 		   (match-beginning 1)
 		   (match-end 1))))))
    (goto-char (point-max))
    (insert "\n")
    (dolist (ticket (sort ticket-alist '<))
      (insert (format
        "#%s %s/ticket/%s\n"
        ticket trac-base ticket)))))

I have a separate `trac-base-xxx' function for setting `trac-base' for each Trac that I regularly use, and I use GNUS customizations to call the appropriate `trac-base-xxx' function when I enter specific email folders.

xmlroff 0.5.4 is at http://xmlroff.org/download/xmlroff-0.5.4.tar.gz.

This release fixes some table bugs and adds linefeed-treatment and white-space-collapse properties (actually added in 0.5.3, but that release was only announced on the xmlroff-list).

I received a politely worded letter from Eircom telling me how happy they are to spam me with “special offers, price reductions and new products and services”. In fact, they’re so happy to do it that they’re going to do it even if I stop using Eircom.

Since I want to contribute to Eircom’s bottom line on my terms, not their’s, I’m sending back the opt-out form to the curiously named “Customer Suppressions Department”.

What, then, do they call the people who do the spamming? The “Customer Oppressions Department”?

Building xmlroff on Ubuntu 7.10 is straightforward once you install some build tools and the required ‘-dev’ packages.

Starting with a clean installed system, install the following packages (and their dependencies):

• libtool
• autoconf
• automake1.9
• libglib2.0-dev
• libxslt1-dev
• libcairo2-dev and/or libgnomeprint2.2-dev
• libpango1.0-dev
• libgtk2.0-dev (not libgdk-pixbuf-dev)

This is the second time that I’ve picked up a Tom Keneally book at Sydney Airport for the long flight out of Australia. The other was The Great Shame, and the great shame there is that I didn’t get around to reading it until several years later. By that time I was living in Ireland, so at least I could then better understand the Irish aspects of that account.

The Commonwealth of Thieves covers the period just before to just after the arrival of the First Fleet in Botany Bay. I enjoyed it, and I look forward to the likely future editions covering the next stages in Australia’s history.

Another computer, another instance of the Emacs 22 snapshot heading towards 100% CPU usage when Semantic is active.

The solution I tried before makes the Synaptic package manager complain about the Semantic version whenever I install any package. This time, I followed advice from the Emacs wiki and applied a patch to semantic-idle.el in place, byte-compiled the file, and moved the new .elc file over the existing .elc file (in a different directory).

It’s working fine, and Synaptic isn’t complaining.