Hi.

I need construction advice on the following (for java and windows filesystem):

#1: Convert full java class name to filename:
  Example:  com.example.ClassName      -> com/example/ClassName.java
         :  com.example.ClassName.java -> com/example/ClassName.java


#2: Convert image in classpath to filename (file extension may vary):
  Example:  com.example.images.Image.png -> com/example/images/Image.png

#3: Add path and wildcard to a filepath (if it does not start with "XXXX." where XXXX can be any characters upper and lower case) :
  Example: p_databasePackage.pkb       -> *.p_databasePackage.pkb
         : CHDEV.p_databasePackage.pkb -> CHDEV.p_databasePackage.pkb

The original values are passed as a list of filetype + filename from a web app, and I need to validate their existance in a SCM system. The filetypes and how to resolve their path is stored in a database. I would really like to include a regexp per filetype on further resolving the filename, since users do not enter filenames as such.

I would really appreciate all and any help given on this.

I would approach these as follows:

For the case where the ".java" is missing, use the normal string matching functions to see if the suffix is ".java" - if not then append the required string. While it may be possible to come up with a regex, it will be very complex as detecting the absence of a prefix is messy (to say the least). The best you can do is to use a regex to find the ".java" suffix and then use your programming language to append what you want if the match fails - but you can do all this with simple string operations far more easily.

To change all but the last "." to "/", try the pattern:

\.(?=.*?\.)

and the replacement string

/

This looks for a literal period that is followed by another one and replaces it with a "/". The last period will not be followed by another one and so will not be replaced. This covers the 2nd case of example #1 as well as example #2.

For example #3 (I assume that the generic match is to present "*." if there is only a single period in the string) use the pattern:

^(\w+\.\w+)$

and replacement string of

*.$1

(You may need to use '\1' or something else to refer to the captured text form match group #1 in the replacement string). If the composition of your filename induces characters that are not in the '\w' class (such as "$" and the like) then you will need to alter these parts.

Each of these regex replacements will need a separate pass as you cannot tell during the replacement part which replacement string to use.

Susan