Trial and error gets me to these steps that seem to work.

1. download py
2. unpack it to directory py147 (whatever you want to name it)
3. download pytest
4.  unpack it to directory pytest223 (again, whatever you want to name it)
5. install py (see below)
6. install pytest (see below)

In this example, I’ve got a symbolic link in my path called python27 that links to my python 2.7 executable.

> cd py147
> python27 setup.py install
> cd ../pytest223
> python27 setup.py install

# now test it by running one of the pytest included scripts
> cd testing
> python27 -m pytest test_python.py

Splendid.
Now getting on to the task of examining it to see if it’s functionality meets my needs.

Right now, I’m doing two things.
1. Writing my own framework.
2. Investigating pytest.

Hopefully I will have some results from both experiences to share.

However, what if you want to zero pad the number.

Well, that’s where iomanip comes in handy.

#include <iomanip>  // for setw, setfill
#include <iostream> // for cout
#include <sstream>  // for ostringstream
#include <string>   // for string

using namespace std;

int main ()
{
  int i = 1;

  ostringstream ss;
  ss << setw( 3 ) << setfill( '0' ) << i;
  string s = ss.str();

  cout << "i: " << i << endl;
  cout << "s: " << s << endl;
  return 0;
}

output:

i: 1
s: 001

A confusion arises sometimes, especially with those new to C++ rules, as to the variables at the top of a for loop.

In a for loop, such as for (int i = 0; i < 2; i++) { ... }, does the int i belong to the outer scope or the scope within the for loop.

The answer is that it belongs to the inner loop, and hides any outer variables, if they exist.
Of course, this is best shown with an example.

Three variables: i, j, k
Variable i is defined both in the outer scope and as a loop index variable defined in the top of the for statement.
Variable j is defined both in the outer scope and within the curly brackets of the for loop.
Variable k is only defined in the outer scope.

The treatment of i is the same as j. It’s as if all of the statements at the top of the for loop are inside the curly brackets of the for loop.

Sample Code:

#include
using namespace std;

void report(string label, int i, int j, int k)
{
  cout << label << " i:" << i << " j: " << j << " k: " << k << endl;
}

int main ()
{
  int i = 123;
  int j = 456;
  int k = 789;

  report("before loop", i, j, k);
  cout << endl;

  // inner scope i hides outer scope i
  for (int i = 0; i < 2; i++)
  {
    // inner scope j hides outer scope j
    // outer scope k not hidden
    int j = i;
    report("in for loop", i, j, k);
  }
  cout << endl;
  report("after loop ", i, j, k);
  return 0;
}

output:

 > g++ -Wall -o scope for_loop_scope.cpp
> ./scope
before loop i:123 j: 456 k: 789

in for loop i:0 j: 0 k: 789
in for loop i:1 j: 1 k: 789

after loop  i:123 j: 456 k: 789

Please leave a comment if you feel I’ve gotten something wrong here.
If your compiler does something different, I want to hear about it.

However, I occasionally have an idea that I’d like to test out by itself, isolated from the rest of the code.
My usual solution is to write a small test program and compile it on the command line, debugging with writes to the output stream.

As an example, I recently wanted to refresh my memory about how string::find works.
I saw an example program at cplusplus.com

So to start, I just grabbed the example and saved it to string_find.cpp.

Now to compile it.
Hmmm.
First attempt:

gcc -o string_find string_find.cpp
... massive list of errors omitted ...

Well, that wasn’t it.

Look it up somewhere, can’t remember where, probably on stackoverflow.

Here’s what I want:

g++ -Wall -o string_find string_find.cpp
> ./string_find.exe
first 'needle' found at: 14
second 'needle' found at: 44
'haystack' also found at: 30
Period found at: 51
There are two prepositions in this haystack with needles.

Awesome.
Now, post this to the blog so I don’t have to look it up again.
Whew!

Turns out that the gcc command will try to compile c++ code, but doesn’t link the standard library.
However, g++ does include the standard library, so stuff like are available.

Is happiness the secret of success? – CNN.com

Some people think if you are happy, you are blind to reality. But when we research it, happiness actually raises every single business and educational outcome for the brain. How did we miss this?

Quick read, and very worth it.

I’ve been trying it out for a while. Actually, I’m doing exactly what you shouldn’t do. I’m running it as my main theme with no child theme. Right now I’m just collecting notes on things I need to hack to get it to work.

I plan on creating a child theme at some point and moving my hacks/mods into the child theme. But for now, this works.

What I want to talk about now is how to add support for the subscribe to comments plugin.

By default, startbox does not support the plugin. However, this is really just a matter of styling.

Startbox has a line of css that hides the ‘Notify me of new posts by email’ text in the comment form. No worries, this can easily be overridden.

Add the following to your css, and all will work spledidly.

#commentform label.subscribe-label {display: inline;}

Label stuff with a new label.

What stuff?

Everything from my current directory down using my current config spec.

Here’s the code

cleartool mklbtype -nc 'my_new_label'
cleartool mklabel -recurse 'my_new_label' .

mklbtype : create a new label

mklabel: apply label

These are development branches that I don’t need anymore, and would rather not see in the default version tree.

My solution below works because I name all of my development/work package branches with a particular prefix.

Let’s say my prefix is dev_userName.

So, here’s the command.

alias ct='cleartool'
for b in $(ct lstype -kind brtype -short | grep dev_usrName)
do
  ct lock -obsolete brtype:$b
done