On 31/08/09 08:25, Stonewall Ballard wrote: > It seems that you all aren't using XCode 3.2. Pieter said that it > changes the xibs, hence produces excessively-large patches. I think > I'll have to wait until the origin is updated. Might that happen > anytime soon? Yes, I'm hoping Pieter will push to GitHub soon too ;) About XCode 3.2: I just noticed it is out (though it seems to be targeted at 10.6). I ordered Snow Leopard yesterday and will start using it next week when I get back home.
Diffs of XIBs are often unnecessarily big, regardless of the XCode-version. With some practice one can spot the lines that can be left out of a commit. > I submitted this patch by pasting into a reply panel directly in > Google Groups. I hoped that would work, but apparently not. Is there a > better way to avoid hard-wrapping? That's not a good way to send your patches. You have a few other options 1. Use a mailprogram (Thunderbird, Mail.app, mutt), but make sure that copy-pasting patches doesn't rewrap them. Have a look here for hints on common mail-programs and their ill-behavin': http://repo.or.cz/w/git.git?a=blob_plain;f=Documentation/SubmittingPatches;hb=HEAD 2. Use git send-email like this: git send-email --suppress-cc=self 000* [email protected] You'll have to know your way around setting up a small MTA (think msmtp or sendmail) on your local system. 3. Send patches as attachment (yuk!). This sucks, since we can't review them inline. > Meanwhile, I've put this patch here: > <http://dl.getdropbox.com/u/297034/0001-Add-preferences-for-opening- > things-at-launch.patch.zip> Now, about the patch: It looks fine, one thing one might argue about is the wording of the prefs and the placement, but that's something we can worry about when we get more preference-options. I can't say much about the XIB, but it probably could be a smaller diff. But it applied fine on my bugfixes-branch, so no problem here. One tiny technical detail I was wondering about was: id curPath = [[[NSProcessInfo processInfo] environment] objectForKey:@"PWD"]; This works fine, but shouldn't we use NSString* curPath = ... here? Using a specific type makes it easier for a reviewer to see what kind of object is to be expected, and in this case it's clearly a string, last but not least because [NSURL fileURLWithPath:curPath] expects an NSString* as parameter. Other than that I have no objections to this patch from the usability-standpoint, I even might come to use it myself ;) Greetings, Jojo -- Johannes Gilger <[email protected]> http://heipei.net GPG-Key: 0x42F6DE81 GPG-Fingerprint: BB49 F967 775E BB52 3A81 882C 58EE B178 42F6 DE81
