[Laszlo-dev] [Laszlo-checkins] r11453 - openlaszlo/trunk/WEB-INF/lps/lfc/controllers

[Lou Iorio]

On Oct 16, 2008, at 4:36 PM, P T Withington wrote:

> Well that too, which is annoying, but I am concerned about helping  
> us humans, not whether svn can be efficient or not.
>
> You are right, that if we re-fill text, svn will do a really poor  
> job of integrating changes from separate branches.  With code, we  
> are not re-filling though, just re-indenting, so svn should not be  
> too perturbed.
>
> [Personally, when I source-control marked-up text, where whitespace  
> is not significant, my rule of thumb is to 'hard wrap' new text at  
> 75 chars per line, and not re-wrap old text (if necessary, split a  
> line and/or insert a line, but don't re-fill).  It's somewhat  
> annoying that our existing doc was apparently written in an editor  
> that gave the illusion your text was nicely wrapped, when it  
> actually was all one long line...]

I understand now. I also hard wrap all new stuff (but at 80  
characters, as the Apple ][ wanted), but as you say, this:

<section>
	<title>Zombie Woof</title>
	<para>I am the Zombie Woof, I'm that creature all the ladies been  
talking about.</para>
</section>

is easier to read than this:

<section><title>Zombie Woof</title><para>I am the Zombie Woof, I'm  
that creature all the ladies been talking about.</para></section>

I'm sure our existing doc was computer generated, and I'm sure that  
whoever wrote it (especially if he still works here) was well- 
intentioned.

Your other suggestion (but I'm sure nobody cares about it as it  
pertains to doc) is to do a one-time re-fill with no content changes.

>
>
> On 2008-10-16, at 16:15EDT, Lou Iorio wrote:
>
>> It's very possible that I misunderstand the issue here, but isn't  
>> the problem the way svn tracks
>> changes? Our DocBook stuff is a total mess; nearly impossible for  
>> humans to read, and easy to fix,
>> but I was under the impression that I couldn't fix it because svn  
>> tracks changes line by line.
>>
>>
>> On Oct 16, 2008, at 4:05 PM, P T Withington wrote:
>>
>>> Agreed.  But the 'diff' that gets sent with a review is not as  
>>> smart.  Maybe we should just tell the diff to ignore whitespace  
>>> too.  Oh, I think we did not originally because there was a lot of  
>>> Python in the code base.
>>>
>>> On 2008-10-16, at 14:36EDT, André Bargull wrote:
>>>
>>>> Hmm, I'm using WinMerge to inspect my diffs and WinMerge has got  
>>>> an option to ignore all whitespace changes. This makes it pretty  
>>>> easy for me to focus on real changes.
>>>>
>>>>
>>>>> On 2008-10-15, at 04:37EDT, bargull at openlaszlo.org <http://www.openlaszlo.org/mailman/listinfo/laszlo-dev 
>>>>> > wrote:
>>>>>
>>>>> >/ (And did some work for the all-time favourite LPP-2623.)
>>>>> /
>>>>> When we did the 'class conversion' Phil intentionally did not  
>>>>> re- indent the code to make it easier for reviewers to see what  
>>>>> changes  had been made.
>>>>>
>>>>> I was thinking it would be great to go through the whole code  
>>>>> base and  re-indent (with no algorithmic changes) at some point.
>>>>>
>>>>
>>>
>>>
>>
>