RE: [Evolution] evolution folder layout (Off-topic)

On Mon, 10 Apr 2000, Paul Bleisch wrote:

Am I the only one that is confused as to why the naming convention of
folders *has* to be tied to the naming conventions for the underlying
filesystem?  It seems that as long as this is the case, Evolution will
require some amount of code devoted to verifying that some folder name
is "legal" and fixing it up if it is not.

  I believe that the folder naming convention should follow the
filesystem's conventions because it reduces the number of conventions the
user has to memorize and creates a 1-to-1 mapping between folder names in
the application and on the disk.  
  Also, my earlier statement requires some clarification.  When I said
that the folder names should have the same convention as any other file
would, I didn't intend for anyone to write code in Evolution that performs
the same checks as the OS.  On the contrary, I would expect Evolution to
have *no* restrictions on filenames.  That is, a user should type in a
file or folder name, Evolution should submit it to the OS via a system
call, and the system call should blow up if the name is illegal.  Thus,
Evolution will conform to the filesystem's conventions regardless of what
the filesystem or OS happens to be.

Someone suggested that there be a map in evolution that maps
"Evolution folder names" to "filesystem directory (folder) names". I
can't recall why this was shot down (and I couldn't find any comment
on it in the archives).

  How is this better or more simple than following the well-established
conventions currently in practice?  And what sort of characters are you
looking for that you *need* this map?  My philosophy here is to not add
more code than is needed, especially since you have perfectly good system
calls that will handle filename restrictions for you.


[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]