regexp: invalid UTF-8

[Carlos Cobo]

Hi there gophers,

I'm trying to remove a certain character pattern from a string.

The actual pattern is "(?i)\xa7[0-9a-fk-or]" but it doesn't work at all. I always get: error parsing regexp: invalid UTF-8: `�[0-9a-fk-or]`

That "\xa7" thingy is a "§" character, called Section sign.

It seems regexp doesn't translate my character

I tried using:

• "(?i)\u00a7[0-9a-fk-or]", \u00a7 instead of \xa7
  
• "(?i)\uc2a7[0-9a-fk-or]", '§' == '\u00a7'
• `(?i)\\u00a7[0-9a-fk-or]`, using \u00a7 fails (invalid escape sequence: `\u`)
• And a bunch more hoping something magically will work but none of them did actually work.

Here's the code:

http://play.golang.org/p/prSh39bIng

Thanks,

Carlos

[Paul Hankin]

\x inserts bytes into your string, whereas you want to insert the UTF-8 encoding. http://golang.org/ref/spec#String_literals

Using \u00a7 instead of \xa7 works. See: http://play.golang.org/p/Sq_C6qAUyq

-- 

Paul

[Carlos Cobo]

Yeah I figured that out when I tried translating the Section sign to []byte and to []rune.

The solution you suggest doesn't remove the characters but at least doesn't complaing about invalid UTF-8.

Copypasted from your code snippet, last 2 lines:

Before: "\xa7e--------- \xa7fHelp:"...

After: "\xa7e--------- \xa7fHelp:"...

It should be:
Before: "\xa7e--------- \xa7fHelp:"...
After: "--------- Help:"...

[Matt Harden]

Your test string is also invalid UTF-8. You should use \u00a7 instead of \xa7 in all strings.

[Carlos Cobo]

Second time I repeat.

I tried with both "\xa7" and "\u00a7". None of them work.

It seems the service producing this messages doesn't give a **** about UTF-8 so I'll have to do 2 passes. First to correct bytes, then to remove them.

[Peter]

Read up on string literals: http://golang.org/ref/spec#String_literals

There's some subtlety involved, but once you understand it you'll see it's quite consistent.

Have a look at http://play.golang.org/p/_rrMfmDKZh to make sure you can tell what's going on.

Hope this helps.