Re: [PATCH] Experimental patch view mode

[Kai <kai willadsen gmail com>]

On 21 March 2011 06:26, Piotr Piastucki <leech miranda gmail com> wrote:
>
>
> On Sun, Mar 20, 2011 at 3:57 AM, Kai <kai willadsen gmail com> wrote:
>>
>> While I'm not certain, this isn't how I've previously envisaged a
>> patch mode, and existing bugs have also put forward a different view.
>> Most of the discussion I've seen around 'patch mode' has been based on
>> the idea of:
>>  source tree + patch = version-control view
>> or directory instead of version-control or similar. See for instance:
>>  https://bugzilla.gnome.org/show_bug.cgi?id=517902
>>  https://bugzilla.gnome.org/show_bug.cgi?id=572565
>>
>> Your take is different, in that it's purely a mode for viewing a patch
>> file, without a source directory as context. I have to admit that I've
>> never wanted to do this, so I don't really see the benefit, but...
>> maybe people would find it useful?
>>
>> I'm also worried because Meld already has three very different view
>> modes, and this is a fourth mode, adding to our already-fragmented UI.
>> The benefit to the other patch-view approaches is that they're just
>> different ways to get at our existing UI.
>
> Yes, your points are valid.
> Working with source directory is on my todo list. The patch provides quick
> preview mode similar to what Kompare offers. Personally, I use it quite
> often and this is actually the only case when I have to use Kompare instead
> of Meld. However, the user should be able to select the source directory in
> some way (dialog/relative to patch location etc.) and if the source
> directory is not selected or the patch cannot be applied quick preview will
> be used as a fallback.

I like the idea much better thinking about it as 'real patch mode'
(i.e., the existing bug requests) with a 'fallback patch mode for when
you don't have source' (i.e., your current patch). This sounds like a
reasonably cohesive concept, though I'm still concerned about the
issues with adding yet-another-UI-mode.

> What are the benefits?
> 1) I can associate Meld with .patch files and preview their content in a
> more pleasant way than just opening them in gedit no matter if I have the
> source directory on my machine or not

Fair enough.

> 2) I have to deal with old patches that cannot be applied without
> modifications due to code base changes. However, the functionality provided
> by the patches is still valid and I need to check what parts of the code can
> be used.

Also seems reasonable. Again, the comparison is with opening the patch
in gedit...

> I fully agree that the number of users of such a functionality may be
> limited and it may not be beneficial to include it in Meld.

...but it also might. :)

I'd agree that your approach is definitely better than opening a patch
in gedit, but I don't know that it's  *massively* better. I'd like to
see if we could come up with concrete examples of ways in which we can
make your cases above compellingly better than just using a text
editor.

cheers,
Kai