1 |
>>>>> On Tue, 18 Sep 2012, Brian Harring wrote: |
2 |
|
3 |
>> > from diffball (under current EAPIs) |
4 |
>> |
5 |
>> > """ |
6 |
>> > RDEPEND=">=sys-libs/zlib-1.1.4 |
7 |
>> > >=app-arch/bzip2-1.0.2 |
8 |
>> > app-arch/xz-utils" |
9 |
>> > DEPEND="${RDEPEND} |
10 |
>> > virtual/pkgconfig" |
11 |
>> > """ |
12 |
>> |
13 |
>> > becomes the following under the proposal: |
14 |
>> |
15 |
>> > """ |
16 |
>> > DEPENDENCIES=">=sys-libs/zlib-1.1.4 |
17 |
>> > >=app-arch/bzip2-1.0.2 |
18 |
>> > app-arch/xz-utils" |
19 |
>> > dep:build? ( virtual/pkgconfig )" |
20 |
>> > """ |
21 |
>> |
22 |
>> Which is longer than the original. ;-) |
23 |
|
24 |
> I see 5 lines in the first version, and 4 in the second. I also see |
25 |
> either someone who counted wrong, or basing that statement purely on |
26 |
> byte count (which is frankly arguing to argue on your part). |
27 |
|
28 |
Can we agree that both counting of lines and characters is silly? ;-) |
29 |
My point was that the new syntax isn't significantly more compact than |
30 |
the present one. In one case there is another variable assignment, |
31 |
in the other case you need an additional "dep:build? ( |
32 |
virtual/pkgconfig )" group. |
33 |
|
34 |
Readability is more important, and there I still don't buy the |
35 |
argument that the new syntax is better, and that any gain would |
36 |
outweigh the cost of changing. After all, the existing variables for |
37 |
dependency specification won't disappear, so devs would have to |
38 |
remember both. |
39 |
|
40 |
Ulrich |