Usage of typename

What is wrong with the following code?

The issue is very simple but hard to notice. If you try to compile this, you will get the following errors:

    main.cpp:24:5: error: need 'typename' before 'OuterStruct<T2>::InnerStruct' because 'OuterStruct<T2>' is a dependent scope
    OuterStruct<T2>::InnerStruct mUsingInner;
    main.cpp: In function 'int main(int, char**)':
    main.cpp:34:13: error: 'struct InnerStruct_Wrapper<int>' has no member named 'mUsingInner'
    wrapper.mUsingInner = innerStrct;

The Issue

At least it straight away tells you something is wrong with InnerStruct_Wrapper. Here’s what is happening:

  • The compiler does not know that mUsingInner in that line is actually a variable of type “OuterStruct::InnerStruct”

  • InnerStruct will only be known later when it is being instantiated

  • Compiler cannot figure out what InnerStruct means here. It could be a type or a member variable

The Fix

The only way here is to help the compiler by telling it beforehand that is going to be a class or struct type. The way to tell the compiler that something like this is supposed to be a type name is to throw in the keyword “typename”. The “typename” keyword has to be used everytime you want to tell the compiler that it should expect a Type in its place.

One place where this is used quite a lot is when templates have to implement iterators.

Full Code

The full sample code, along with the fix is provided on Stacked-Crooked.

Setting up your VIM... Your style

Writing code using VIM can be a bit overwhelming but it helps creating a few shortcuts there to make your life easier. Here I will discuss a few.

  • First of all, it is always a good idea to set up your tabs and spaces
  • Set up shortcuts for:
    • Toggling of Auto-Indent while pasting source text; very useful in code (F2 in my provided script)
    • Toggling of Auto-Scroll of Split windows (F3 in my provided script)
    • Macro for Folding functions within the source code (@q in my provided script)
    • Smart case search - based on whether the input is in caps or not
    • Toggling of Mouse Usage in VIM window - to allow quick selection of text (F12 in my provided script)

Please see the provided script below to have a look.

Getting a group of lines from a file

I’ve had this need quite a few times to pull out a section of logs that would begin with a particular line and end with another. grep is not exactly useful there because it only prints out sections based on line counters (using -A/B/C) lines based on a single search pattern.

I came up with a script that can run at almost same speeds using grep/cat/awk. awk is used to decide whether the end of the section has been reached or not. Some features of the script are:

  • Since awk script only toggles one variable it works seamlessly without delaying the actual work

  • Works on gzip files as well

  • If you do not want to depend on grep or are unsure how many lines may be between begin and end keyword then replace gunzip with zcat and grep with cat.

  • Usage: group filename printBeginKeyword printUntilKeyword NumOfLines

Let’s have a look at the script….

32 bit vs 64 bit

When to use 64-bit?

If your application needs more than 2-4gb of data

  • If your application intensively uses 64-bit arithmetic
  • * 32-bit x86 compiled applications are restricted to x86 instruction set, even when run on on a 64-bit machine
  • x64 supports 16 registers compared to just 8 in x86. If your code is computation intensive this may help a great deal

When to use 32-bit?

After reading so many great things about 32 bit, why would anybody still want to code against 32-bit? Simple reason - code compiled in 32-bit takes lesser memory. For example, a pointer in 64-bit machine is 8 bytes compared to 4 bytes with 32-bit. Now imagine how many pointers do you have going around in your code?

When to not use 64-bit?

Sometimes if your code is running on a machine that does not have too much of space then it is a never a great idea to use 64-bit compilation because the same task can probably be done much better using 32-bit version. L2 and L3 caches will be utilized a lot more in 64-bit code (and a lot earlier as well) compared to 32-bit; on such systems.


Always benchmark your code and ensure that the application actually needs more than 2-4Gb of space.

Tokenize a String using C++

Here’s a short snippet to split a string into multiple tokens; into a vector. As you will see that, if you run the code, boost version performs better because you can choose a number of delimiters to split your string instead of the vanilla version using the normal C++ code. Of course, you may also write your own code to do something like this but I was looking to do some short snippets.

I have posted my code on Stacked-Crooked which you can view along with the output as well. It shows C++ doesn’t perform so well.

Heap Sort vs Merge Sort vs Insertion Sort vs Radix Sort vs Counting Sort vs Quick Sort

I had written about sorting algorithms (Tag: Sorting) with details about what to look out for along with their code snippets but I wanted a do a quick comparison of all the algos together to see how do they perform when the same set of input is provided to them. Hence I started working on a simple implementation for each one of them. I have now put together all of them in a single project on GitHub. I ensured that they all have the same set of procedures during their run. Some of the items I wanted to ensure was:

  • Same input array

  • Same number of iterations. Each iteration having the same input

  • Each algo being timed the exact same way as another

Some of the algorithms being tested were:

How did we ensure Equality?

  • Created a simple base class for all algorithms: AlgoStopwatch

    • Responsible for benchmarking everything

    • Provide a function called doSort() that would allow derived classes to implement their algorithm

    • Ensures that every algorithm has a name and description - to help us distinguish

  • Another class to help manage the testing of all the algorithms: AlgoDemo

    • All instances are created here for the algorithms

    • The input array is provided by this class to all algorithms


As usual the code for the project is available here:

It can be run using Visual Studio without any changes.


Looking at the numbers below, it may be hard to compare the actual values. Hence I decided to normalize them by calculating how much time will be required to sort 100 numbers using the same rate as the actual numbers. They are provided for all algorithms on the right-most column.

Merge Sort

# of Items in Array Time Taken Average for 100 numbers
Random 10 0.047464 0.47464
1K 5.41906 0.541906
1M 8444.11 0.844411
Sorted 10 0.027155 0.27155
1K 4.47016 0.447016
1M 6323.05 0.632305

### Merge Sort 2 **# of Items in Array** **Time Taken** **Average for 100 numbers**
**Random** 10 0.033668 0.33668
1K 3.89374 0.389374
1M 7076.04 0.707604
**Sorted** 10 0.019034 0.19034
1K 2.7833 0.27833
1M 4664.16 0.466416
### [Insertion Sort]( **# of Items in Array** **Time Taken** **Average for 100 numbers**
**Random** 10 0.006095 0.06095
1K 0.369859 0.0369859
1M 323.878 0.0323878
**Sorted** 10 0.005022 0.05022
1K 0.11696 0.011696
1M 122.427 0.0122427
### [Heap Sort 1]( **# of Items in Array** **Time Taken** **Average for 100 numbers**
**Random** 10 0.013543 0.13543
1K 3.22043 0.322043
1M 4770.14 0.477014
**Sorted** 10 0.007507 0.07507
1K 0.625425 0.0625425
1M 633.142 0.0633142
### [Heap Sort 2]( **# of Items in Array** **Time Taken** **Average for 100 numbers**
**Random** 10 0.019352 0.19352
1K 3.86284 0.386284
1M 8914.22 0.891422
**Sorted** 10 0.011289 0.11289
1K 3.49712 0.349712
1M 6661.45 0.666145
### [Heap Sort 3]( **# of Items in Arraytrong>** **Time Taken** **Average for 100 numbers**
**Random** 10 0.016266 0.16266
1K 0.968032 0.0968032
1M 900.004 0.0900004
**Sorted** 10 0.012845 0.12845
1K 3.08637 0.308637
1M 3839.23 0.383923
### [QuickSort]( **# of Items in Array** **Time Taken** **Average for 100 numbers**
**Random** 10 0.072972 0.72972
1K 2.74698 0.274698
1M 4640.77 0.464077
**Sorted** 10 0.042773 0.42773
1K 1.84335 0.184335
1M 2473.42 0.247342
### [Counting Sort]( **# of Items in Array** **Time Taken** **Average for 100 numbers**
**Random** 10 0.022982 0.22982
1K 1.21822 0.121822
1M 1823.85 0.182385
**Sorted** 10 0.026815 0.26815
1K 1.19146 0.119146
1M 1612.58 0.161258
### [Radix Sort]( **# of Items in Array** **Time Taken** **Average for 100 numbers**
**Random** 10 0.033351 0.33351
1K 3.22004 0.322004
1M 5650.9 0.56509
**Sorted** 10 0.020659 0.20659
1K 3.26273 0.326273
1M 5683.91 0.568391

Global Search in VIM

You must be knowing about regular VIM search


However, sometimes you do not want to replace something but just see all instances of the a word or phrase. In such cases, global search is really useful when on vim. The syntax is simpler than normal search-replace:


This will show all instances of the “SEARCH” term within the VIM window. There’s another version of the same command which is as follows:


This is similar to the previous command but the only difference is that this command contains “!” which signifies that the command will be executed on all lines NOT matching the pattern.

One example here would be to delete all lines in a file containing a particular word:


There could be lot more actions (regular vim stuff) that you could do like yanking or indenting.

Reading Excel in C#

Recently had to work on an Excel and update its content when some events were triggered on a Bloomberg terminal. I couldn’t find any code for reading the excel file using C# so thought of putting that up.

There were a few things to learn:

  • Excel support is not available by default so make sure to add Microsoft.Office.Interop.Excel as a reference

  • Sometime indexing to reference the Workbooks or Worksheets may not work so try out other values

    • I remember this was not a problem with Visual Studio 2010 but something must have happened in Visual Studio 2012
  • It is also possible to try Workbook names to try and retrieve them

Please follow the shortened code to read an Excel file after the break.

Ideas for Shared Memory in Linux

I have been doing extensive research today regarding shared memory in linux before I embark on a new project in which shaving time off process communication latency is extremely important. I found some very interesting links as well. Let’s break them down:

Topic Description Link
Theory/Explanation Some articles that explain how to get started open-std
Content in the first column Content in the second column lkd
IPC Mechanisms – some recommendations. Mainly talks about softwares like Boost, dBus dBus used to work using kernel levels but lately have come up with a patch that bypasses kernels Stackoverflow
What every programmer should know about memory LWN
Cache Friendly Code Stackoverflow
Helpful Articles Which linux IPC technique to use? Stackoverflow
Inter process communication – talks about shared memory ALP
Linux Poll Events on Shared Memory Stackoverflow
Using semaphores in Shared Memory Stackoverflow
Sharing semaphores between Shared Memory Stackoverflow
Sharing memory between processes Stackoverflow
Reserving memory at kernel boot up (DMA) Stackoverflow
Extra IPC Shared memory samples linuxgazette

Oracle SQL Queries

Recently I found myself in a situation where I had to view all the tables in the Oracle Server.

I did some digging and came up with these helpful queries.

Query Remarks
Select * from All_Tables View all the tables in the Oracle Server
Select * from all_Views View all the views in the Oracle Server
DESC table View all the columns of a given table. This function is used instead of the previous one mostly, but when this one doesn’t work make sure you try the other one - I bet it will work.

Here’s a short code for viewing all the columns of a given table.