Definite's Extractor

My findings on Life, Linux, Open Source, and so on.

Monthly Archives: February 2010

Open source scrum tools

Last week I went to a scrum training course, which is very insightful and interesting.
The course also suggested some scrum utilities, but none of them are open source.

After the course, I spent some time on finding a open source scrum software, preferably free. My criteria are:

  • Open source and free.
  • Able to be hosted on Linux
  • Provide web interface

I’ve tested express, it has very beautiful and professional interface. Unfortunately, I cannot make it work beyond the registration screen, as it keeps telling me registration mail is failed to send.

Caius suggests agilo, a plug works upon trac. But given I never like the interface of trac, so I won’t bother to install it.

Then I find icescrum2. At the current stage, well, it is not for the end users yet. Still need quite a good deal of work to polish ,such as:

  • No intuitive way to activate and start a sprint.
  • Expected home directory is in /usr/share/tomcat/.icescrum
  • Expected location of in the [webapps]/[icescrum2]/,
    yet the actual location of that file is in [webapps]/[icescrum2]/META-INF/icescrum2/config/

It also forces user to start on sprint 1, which is inconvenient for scrum project which just migrate to icescrum2.

I can post detail of installation instruction for F-12 if anyone interested.
Maybe even make rpm for it.

RHEL5 repo build on Fedora 12

I created a RHEL-5 repo on my Fedora-12 box.
But my RHEL-5 box cannot load the repo due to “Error performing checksum”.
After some Googling, the common solution is to use either sha or md5 as checksum.

createrepo -s sha
createrepo -s md5

However, that solution is not complete, especially if you have tested-run yum with the ill-checksumed repo.

Solution to that is simple, just run yum clean all to clean the cache.

[Input Style] What’s Over-the-spot, on-the-spot … etc?

Last Friday I saw an ibus issue about input style support in ibus-anthy. The maintainer, Fujiwara insist that editing in the candidate window is not "over-the-spot".

Ok, time to do literature review:
According to sun and Mozilla, preedit area is INSERTED into the inputing spot in on-the-spot, the text after the input spot WILL be pushed to the right when preedit area expend; while preedit area is PUT OVER the inputing spot in over-the-spot mode, the text after the input spot WILL NOT be pushed to the right.

Reference from IBM tells different story. Over-the-spot, as the page states, is the mode that candidate window closely followed the input spot, but the preedit string is formed in candidate window.

Java also has its own definition. Below-the-spot is the term for IBM’s over-the-spot.


After the intensive web search and discussion, we conclude that we should use some thing like “Embedded preedit in client application” to avoid confusion.

rpmlint tip

This tip is very simple, I should have RTFM instead of Googling.

If you don’t know the exact meaning of an error or warning message,
either use -i for displaying the detail description of error/warning message, or use -I for checking individual error message.