Please see my comments inline.

On Saturday 31 January 2009 17:35:45 Christian Hammond wrote:

> Thanks for the patches. I'll look into them.
You're welcome.

> > http://reviews.review-board.org/r/729/diff/ ) Probably these keys are
> > intercepted as hotkeys by diffviewer component. I use Firefox-3.0.4 on
> > Linux.
> Sounds like a regression. I was sure we had this fixed. I'll take a look.
> Would you mind filing a bug on this for tracking purposes?


> It's a hard problem to solve, using diffs without the right info, because
> the info is very much needed and having the user input it is extremely
> error-prone and time-consuming.

> Asking for revision info wouldn't even work in all cases. Imagine if someone
> just did a "diff -ur" on some files for a CVS repository and then tried to
> upload it. CVS has per-file revisions, so asking for the right revision on
> each thing would be insanely annoying. 

Well, I completely understand your concerns and, yes, they are quite weighty.
But I'd wish to point out that actually there're two different workflows when 
we talk about submitting changes to a project.

First case is when a regular project developer wishes to make changes on 
day-by-day basis (or at least quite regularly). Certainly this person should 
have actual project sources checked out from a repository. Quite likely he or 
she wishes that review process shouldn't be much harder and longer than a 
simple commit. That's what post-review exists for and probably nothing can be 
drastically changed here, just closer intergration with other SCM-related 
tools. Actually any Web Interface for submitting changes is simply redundant 
and burdened for such persons. Previously I dealt with SmartBear's Code 
Collaborator and can confirm this is true.

But also there's another use-case. And actually when you start creating formal 
releases you'll get a growing number of "mere users" which encountered small 
problems here and there and eventually wish to contribute a line or two to 
fix them. Quite often these users won't have an actual repository; and 
because of Python interpreted nature they even don't need to have "sources", 
just an installed "package" :-)

Probably I was wrong when I mentioned "revisions" in the previous message.
It can be safely assumed that changes such users wish to contribute are to be 
applied "on top" of a chosen branch, no matter what exact revisions are 
involved. BTW choosing a branch from a prepared dropdown list isn't much 
harder and error-prone than specifying '/trunk/reviewboard' in the Base Diff 
Path field. And I still believe that an (optional) ability to upload a 
modified version of a file or two is a viable alternative to making a patch 
in a special format.

Well, it looks like I have another use-case in mind but probably it's more 
specific to distributed SCMs (not just git, they all are quite similar in 
common operations with code). But let it be the subject for another 
message :-)

Yours respectfully,
Alexey Morozov

You received this message because you are subscribed to the Google Groups 
"reviewboard" group.
To post to this group, send email to reviewboard@googlegroups.com
To unsubscribe from this group, send email to 
For more options, visit this group at 

Reply via email to