Re: [Flickrfs-users] Bug fixes and my enhancement checked in.
Brought to you by:
manishrjain
From: <rdm...@bi...> - 2006-02-09 02:55:56
|
On Thu, 9 Feb 2006 at 10:03, Manish Rai Jain wrote: > On 2/9/06, rdm...@bi... <rdm...@bi...> wrote: > Gr8 that you finished off the ctime/mtime functionality. And I think it wd > be more useful if ctime is 'time photo is taken' extracted from photos Exif > data. While utime wd be the time when photo is uploaded. If by 'utime' you mean what is in the code self.atime, then yes I agree that makes maximum use of the available data. So date_taken=ctime, last_update=mtime, and date_uploaded=atime. >> If it's OK with you, Manish, I'd next like to do something that I >> actually enjoy, though many programmers don't: refactoring. I can see >> several places where the code could be tightened up and simplified, >> now that you've roughed out how it all works. It will mean some fairly >> radical changes to the code, but in the end I think it will be easier >> to maintain and enhance (or I wouldn't suggest I do it!). > > Ok David. No probs with me, but I got some concerns. Lot of changes may tend > to break the code. To make sure that the changes can be tracked, and the > code is still working, it wd be better to do all the changes followed by > thorough testing locally, and then put it upstream in 1 go, instead of doing > cvs update after each small change. The way I do this is I have a test suite. I only check in code when the test suite passes. The only problem is that my test suite doesn't test 100% of the functionality. In particular I'm not currently testing upload. So I guess I'd better figure out a way to do that. Come to think of it I had ideas about that when I first started writing the tests but had to give up because of some of the bugs I've now fixed, which means I should be able to resurrect that test code... Anyway, the way I'd like to do this is to make incremental changes, checking in as often as I can, where "I can" is defined as the full test suite runs without error. Does that address your concerns? (That methodology, by the way, is what I followed in doing the bug fixes, although in that case I was making more and more of the tests pass with each commit. But I always made sure nothing new broke before I did a check-in.) > Also, there are some functionalities and testing I wanted to complete: How about if I throw this list into a 'TODO' file in CVS? Presently I'll take a look at tackling them. --David |