PureBasic Forums - English

Firstly well done for adding Unix style Regular Expressions in PB, potentially very useful since Regular Expressions are v-e-r-y powerful.

So - anyways - here is my wishlist:

1) Could we perhaps shorten/truncate the commands since is a bit of a mouthful - perhaps to 2) MatchRegularExpression() currently only returns 0 or 1 - it would be more useful to return the position of the first matched string - eg: 3) Knowing the number of matched Regular Expressions would be really useful too:
To make best use of this powerful library we need new commands to:
4) Search & replace and return the position of matched Regular Expressoins - eg: 5) Transpose strings using Regular Expressions (like the Unix 'tr' command - eg:

Regex term is often used.

naw wrote:1) Could we perhaps shorten/truncate the commands since

Code: Select all

CreateRegularExpression() / MatchRegularExpression()

is a bit of a mouthful - perhaps to

Code: Select all

CreateRegExpr() / MatchRegExpr() 

For this point, a macro do this :

> Could we perhaps shorten/truncate the commands since
I use auto-completetition, so there's o problem with long commands. Remember that PureBasic must stay easy/readable also fpr beginners. You can also use macros, as Progi1984 said.

> MatchRegularExpression() currently only returns 0 or 1 - it would be more useful to return the position of the first matched string
I already asked this, I even asked Fred this via PM and he will NOT implement that, because of performance-reasons.
But I agree, it would be very very very good, if we had an additional command for that!

>Knowing the number of matched Regular Expressions would be really useful too
I also asked Fred this. His solution is to misuse ExtractRegularString(), as it returns the number of results.
@ Fred: I find this a bit too ressource-consuming as you need an additional array which costs memory and time.

> Search & replace and return the position of matched Regular Expressoins
Do I misunderstand you? There is already such a command: ReplaceRegularExpression(). The only thing that must be implemented is that ReplaceString$ can be a RegExp, too.

> the Unix 'tr' command
I only used 'tr' once, but IMHO you can do this already with ReplaceRegularExpression(). Doesn't 'tr' only handles single characters? If so, can't you use ReplaceRegularExpression(0, "Hello World", "[a-z]") or do I have a thinking-error?

// Edit:
5) Well this is a good example for ReplaceRegularExpressions() to support RegExp.

@ Fred:
I just got the following idea:
ExtractRegularExpression() redims thegiven Array, so it has to know the number of results, hasn't it? If so, there must be a kind of CountRegularExpressionMatches()...
Or does ExtractRegularExpressionError() redim the array each time an element is added?

@AND51 & progi1984,
re: xxxxxRegExpr() - I'm an ex Unix admin, so I just like my commands to be short and cryptic eg: sed, awk, grep, tr, bc

re: MatchRegularExpression() counts the number of matches - Sorry - I guess I should have (re)searched the forums first - but still the RE is incredibly powerful, but we need the tools to exploit it

re: ReplaceRegularExpression() - yes would be great to use RE for the ReplaceString - just like the tr command eg: Generally the RegExpr library implementation is a bit cumbersome (IMO).
I dont understand why CreateRegularExpression() & FreeRegularExpression() have been implemented - presumably there must be some advantage in pre-loading RE's into an array, but I cant imagine what or why?

Things would be much neater if we could replace:
with: