>
>>Ok, I try to say differently:
>>everyone can read files, but only authorized people can modify files.
>If I can read the file, I can certainly modify it.
Of course you can, but you can only modify your local copy of files. If you
cannot do a commit you cannot modify the repository and what counts is what
is in the repository.
I believe that what you really want to do is to be sure that developers use
the right environment: you're looking for a way that minimize human
mistakes. From a theoretical point of view this may be possible: programmer
"A" developes (modifies) a set of files and uses another set of
files(source,libraries,DLL,...) to build an image only, without having the
possibility to modify them. The same for programmer "B","C", and so on.
>From a pratical point of view this is impossible to obtain.
The right way is define a process that verifies all jobs done:
1)someone releases a software putting files in repository
2)someone tests it (usually getting directly the image)
3)someone verify that is possible build software with files in the
repository and that software so built is the same released at point 1)and is
the same tested at point 2).
You can optimize previous points, but you cannot avoid them.
______________________________________________________
Get Your Private, Free Email at http://www.hotmail.com
