Question

I am using SVN to develop some software in a team of four people.

Whenever I update my working copy, it creates far more conflicts than I would expect to be created, a large number of these conflicts are like so:

If the initial was

 import Class1 

and I change it to

import Class1 import Class2 

while another team member changes the code to

import Class1 import Class3 

Subversion will not merge all three lines.

I suppose this is because subversion isn’t aware of the semantic content of the code, so couldn’t decide when it would/wouldn’t be appropriate to do this.

Another conflict I get is if I replace

processA(); processB(); 

with :

if(x){   processA();   processB(); } 

the whole thing is marked as a change, is there no way to get it to highlight the fact that only if(x) was added above/below. In cases like this i have to analyse the entire lines inside the new block, and worry about missing small changes in the lines inside.

The question is, is there a more effective way for my team to use SVN in these cases, a large portion of my time is spent merging.

Sorry for being so long winded in my expression, thanks for reading.

Answer 1

Since SVN doesn’t know about programming languages it would be very dangerous if it would try to deduce semantic meaning from change sets. Therefore it can only do relatively simple merges (i.e. if a file changed in two completely different areas).

The common solution is to use small commits. Don’t keep big changes piling up. Commit often. Update often.