Searching For String Literals

Question

In the quest for localization I need to find all the string literals littered amongst our source code. I was looking for a way to script this into a post-modification source repository check. (I.E. after some one checks something in have a box setup to check this stat) I'll probably use NAnt and CruiseControl or something to handle the management of the CVS (Well StarTeam in my case :( ) But do you know of any scriptable (or command line) utility to accurately cycle through source code looking for string literals? I realize I could do simple string look up based on regular expressions but want a little more bang for my buck. (Maybe analyze the string or put it into categories) Because a lot of times the string may not necessarily require translation. Any ideas?

Answer 1

Visual Studio 2010 and earlier:

1. Find In Files (CTRL+SHIFT+F)
2. Use: Regular Expressions
3. Find: :q (quoted string)
4. Find All

Find Results window will now contain a report of all files, with line numbers and the line itself with the quoted string.

For Visual Studio 2012 and later search for ((\".+?\")|('.+?')) (reference, hat-tip to @CincauHangus)

Answer 2

It uses the compiled binary instead of source, but Sysinternals' Strings app might be useful.

Answer 3

To find all Text="textonly" instances use the following Regular Expression when searching:

(Text=)(")([a-z])

This is help for finding Text="*" but excluding text that's already been converted to use resource files:

Text="<%$ Resources:LocalizedText, KeyNameFromResourceFile%>"

Also (>)([a-z]) can be used to find literals between tags like so:

<h1>HeaderText</h1>

Answer 4

There's a C# parser on CodePlex that you can probably use.