[Box Backup-dev] Reviewing code -- help!

[Martin Ebourne]

On Wed, 2005-12-07 at 11:12 +0000, Ben Summers wrote:
> I sat down to start reviewing the code, and I'm now a bit confused.  
> I'm not entirely sure which bits are what, and what should be reviewed.
> 
> I would help me enormously if the contributors could post a list of  
> svn diff commands which will show the proposed changes in the most  
> easily reviewed way.

In order I think we have agreed to merge:

chromi's optimised diffing:

svn diff http://bbdev.fluffy.co.uk/svn/box/trunk http://bbdev.fluffy.co.uk/svn/box/chromi/diffopt

My autoconf:

svn diff http://bbdev.fluffy.co.uk/svn/box/trunk http://bbdev.fluffy.co.uk/svn/box/martin/autoconf

I am planning on reviewing Jonathan's fixes and then merging his branch
then my autoconf into trunk this afternoon, unless someone objects.

Then all my fixes and Johnathan's will be in, leaving win32 and maybe
Gary's stuff.

As I understand it the win32 branch to be merged is:

svn diff http://bbdev.fluffy.co.uk/svn/box/trunk http://bbdev.fluffy.co.uk/svn/box/chris/win32/11-code-cleanups/

I reviewed this one yesterday, and feel there's still a few things that
could be improved before merging it. However, I was generally impressed
with it, seems to be getting pretty close.

The biggest issue I had with the win32 patch was a whole load of #ifdefs
for (s)printf. This is really an issue with the box code - we really
need to convert from printf to streams because printf is a portability
disaster. I'm not sure at which stage it would be best to do this, maybe
after the win32 merge.

There were a few other minor points, but I don't have the list to hand.

After this the outstanding patches I'm aware of are:

Gary's various fixes:
- maxdiffingtime
- keepalive
- decrypting filenames for compare

Nick's non-win32 changes:
- adding -i to bbackupquery
- change to BackupClientRestore's handling when a file already exists

I don't believe any of those really exist on a branch in svn in a way
suitable for merging.

Apologies to anyone I've mis-attributed or forgotten anything. And
please correct any mistakes I've made.

> BTW, am I right in thinking that the aim to get all the Win32 stuff  
> merged into trunk in a way which doesn't break POSIX, but may or may  
> not work properly under Win32. Then fix Win32?

This is one for Chris, but I'd be happy with that plan.

Cheers,

Martin.