P4
A DESCRIPTION OF THE REQUEST :
The annotation processor at the moment provides a very limited way to interact with files, as encapsulted by the Filer interface:
http://java.sun.com/j2se/1.5.0/docs/guide/apt/mirror/
Unfortunately, this makes it difficult to maintain some files over incremental bulids.
JUSTIFICATION :
For example, I am using annotations to build a list of services that are placed in my jar's META_INF/services directory. I want a file in that directory to contain a list of every class that implements a particular annotaiton. Unfortunately, every time I recompile and only some source files with the annotaion have been modified, only those class files will be read and the list will be replaced with this truncated list. I can only be sure thta my list is complete if I clean my class files and build the entire project.
It would be helpful for me to be able to read this text file before overwriting it so I can see what entries already exist in it. Then I can simply copy across the old entries and append the new ones, instead of loosing all the old information in the file.
The annotation processor at the moment provides a very limited way to interact with files, as encapsulted by the Filer interface:
http://java.sun.com/j2se/1.5.0/docs/guide/apt/mirror/
Unfortunately, this makes it difficult to maintain some files over incremental bulids.
JUSTIFICATION :
For example, I am using annotations to build a list of services that are placed in my jar's META_INF/services directory. I want a file in that directory to contain a list of every class that implements a particular annotaiton. Unfortunately, every time I recompile and only some source files with the annotaion have been modified, only those class files will be read and the list will be replaced with this truncated list. I can only be sure thta my list is complete if I clean my class files and build the entire project.
It would be helpful for me to be able to read this text file before overwriting it so I can see what entries already exist in it. Then I can simply copy across the old entries and append the new ones, instead of loosing all the old information in the file.