Question

I’m trying to parse web data coming from a server, and I’m trying to find a more stl version of what I had.

My old code consisted of a for() loop and checked each character of the string against a set of escape characters and used a stringstream to collect the rest. As I’m sure you can imagine, this sort of loop leads to being a high point of failure when reading web data, as I need strict syntax checking.

I’m trying to instead start using the string::find and string::substr functions, but I’m unsure of the best implementation to do it with.

Basically, I want to read a string of data from a server, different data, separated by a comma. (i.e., first,lastname,[e-mail]) and separate it at the commas, but read the data in between.

Can anyone offer any advice?

Answer 1

I’m not sure what kind of data are you parsing, but it’s always a good idea to use a multi layer architecture. Each layer should implement an abstract function, and each layer should only do one job (like escaping characters).

The number of layers you use depends on the actual steps needed to decode the stream

for your problem I suggest the following layers:

1st: tokenize by ‘,’ and ‘\n’: convert in to some kind of vector of strings

2nd: resolve escapes: decode escape characers

you should use std::stringstream, and process the characters with a loop. unless your format is REALLY simple (like only a single separator character, without escapes), you can’t really use any standard function.