Parse C# unformated strings in source code then convert to string format

Question

There thousands of calls to log4net in the application which were done with the style of string concatenation instead of using string format like "{0} is greater then {1}".

So we wrote a program that will parse all the .cs files using Regex to find the log4net logging statements. It extracts the code between the parentheses () and then calls a method to reformat it and return the results. Then it rewrites the source code file.

This question relates to the method that reformats the log statement.

It receives an argument of string and returns string.

Here's a sample of the log statements:

"Column " + column + " Seq Cnt: " + sequentialCount + ": Seq Avg: " + (sequentialTotal / sequentialCount) 
"Opening file for writing copy protection failed. Retrying.", ex
"for assert " + value.ToString("X")
"ExpirationTime()"
"count = " + count + ", round << " + count + " = " + (round << count)
"Total Diff Bytes = " + (7*count)
(series.Count - i - 1) + " " + series.Time[i] + " O,H,L,C:" + series.Open[i].ToDouble() + "," +
                            series.High[i].ToDouble() + "," + 
                            series.Low[i].ToDouble() + "," + 
                            series.Close[i].ToDouble()
"Recovered orders from snapshot: \n" + OrderStore.OrdersToString()

Essentially, it seems that the plan should be to use Regex.Replace() with a MatchEvaluator.

What is the correct Regex expression for the Regex.Replace?

These seem to be the requirements:

Essentially find each interruption in the string "\s*+\s*(.*)\s+" (over simplified).
Replace each match with a token of the form {0} in the string and then put the specified value as an argument to the method.
Debug methods of the form log.Debug("message",ex) which have a reference to an exception must be identified and skipped.

Of course, the code will switch the calls to DebugFormat() InfoFormat() and so on.

The problem with the above Regex is that it matches this as the first match:

" + column + " Seq Cnt: " + sequentialCount + "

instead of:

" + column + "

I can't simply use ([^"]) instead of (.) or ([^+]*) since some of the values have additional plus or use quotes as arguments to methods.

So it needs some way to say match all characters except if the match the pattern \s+\s" which means plus sign followed by a quote separated by optional white space.

Just as a quick aside, if you have Resharper, you can just do a global code cleanup for this...it would be easier than a custom solution — Justin Pihony
– Justin Pihony, Commented Jun 15, 2012 at 14:57
I have resharper and searched on how to do this kind of batch cleanup. However, this doesn't work in resharper even on a single line because the method log.Debug() must be switched to log.DebugFormat(). Please help if you have any specific instructions on how to do this in Resharper. I'll give you the credit for the answer. — Wayne
– Wayne, Commented Jun 15, 2012 at 15:35
Couldn't you just do a global find/replace on log.Debug( to log.DebugFormat(? — Justin Pihony
– Justin Pihony, Commented Jun 15, 2012 at 15:45
I have Resharper 5.1 and it does do what you say if I click on it one line at a time. But I can't find it in the Code Cleanup section can you assist to find it? Or is that in a later version? — Wayne
– Wayne, Commented Jun 15, 2012 at 18:01

Qtax · Accepted Answer · 2012-06-15 14:58:53Z

1

You can make the quantifier lazy. For example:

"\s*\+\s*(\S.*?)\s*\+\s*"

answered Jun 15, 2012 at 14:58

Qtax

34k9 gold badges92 silver badges127 bronze badges

Sign up to request clarification or add additional context in comments.

Collectives™ on Stack Overflow

Parse C# unformated strings in source code then convert to string format

1 Answer 1

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

Comments

Your Answer

Sign up or log in

Post as a guest

Related