1 |
On Thu, Feb 13, 2014 at 10:30 AM, Kent Fredric <kentfredric@×××××.com> wrote: |
2 |
> |
3 |
> On 14 February 2014 04:28, Anthony G. Basile <blueness@g.o> wrote: |
4 |
>> |
5 |
>> 2) You want your vcs to show the diff in that file and you can make sense |
6 |
>> of that diff. |
7 |
> |
8 |
> |
9 |
> Though how many of them are "well formatted" SVGs, and how many of them are |
10 |
> single-line SVG files without whitespace, such as linefeeds and appropriate |
11 |
> indentation? |
12 |
> |
13 |
> Because diffs are usually not very useful if lots of changes occur on a |
14 |
> single line. |
15 |
|
16 |
If these were really hand-maintained SVG files I could see how it |
17 |
might fit into an scm model. However, these are text files in the |
18 |
same sense that an assembly language listing generated from a C++ file |
19 |
would be. Sure, you could manipulate it by hand, but the reality is |
20 |
that upstream is going to change 47 lines in the source and you're |
21 |
going to upgrade your gcc and as a result 90% of the 3M line listing |
22 |
will change on a small upgrade. Likewise, when upsteam changes their |
23 |
desktop icon it isn't going to result in a 3-line change to the SVG |
24 |
file. |
25 |
|
26 |
I guess the question though is whether they cause harm. If the files |
27 |
are small and don't cause issues with the scm they're stored in, then |
28 |
I don't really see the issue with storing them. It doesn't really |
29 |
matter if they're SVG files or uuencoded whatever (though if our tree |
30 |
is GPL we need to make source available for anything that has it - |
31 |
probably not an issue for SVG unless it is derived from some kind of |
32 |
composite). |
33 |
|
34 |
Rich |