URI regex ?

Hi all !

I have to rewrite URI in many files :

  * xsl
  * schemas
  * RNG/RNC grammars
  * schematrons
  * nvdl
  * DTD
  * etc...


I exactly know the protocol(s) of the URI I have to rewrite, so I can 
detect all URIs by a starting string
As I have to work on various file types, and as URI may be in known tags 
(xsl:import, sch:validate, ...) as in variables or anywhere else, I 
would like to process the files as text files.

I have 2 questions :

  * files may use various encoding. Is there a pre-defined encoding
    where URI are always correct ? Or should I check the file encoding
    before reading it ?
  * Is there a regex I can use to detect the end of the URI, knowing the
    start of the URI ? Reading RFC3987 is not so easy...


Thanks in advance,
Christophe