1 |
On Fri, 2 Dec 2016 13:38:19 -0600 |
2 |
William Hubbs <williamh@g.o> wrote: |
3 |
|
4 |
> On Fri, Dec 02, 2016 at 07:47:01PM +0100, Michał Górny wrote: |
5 |
> > On Fri, 2 Dec 2016 13:02:48 -0500 |
6 |
> > Mike Gilbert <floppym@g.o> wrote: |
7 |
> > |
8 |
> > > The devmanual states: |
9 |
> > > |
10 |
> > > The name section should contain only lowercase non-accented |
11 |
> > > letters, the digits 0-9, hyphens, underscores and plus |
12 |
> > > characters. Uppercase characters are strongly discouraged, but |
13 |
> > > technically valid. |
14 |
> > > |
15 |
> > > https://devmanual.gentoo.org/ebuild-writing/file-format/index.html |
16 |
> > > |
17 |
> > > |
18 |
> > > Why are uppercase characters strongly discouraged? |
19 |
> > > |
20 |
> > > Wouldn't it make sense to follow upstream's naming convention? |
21 |
> > |
22 |
> > I'd say keeping things lowercase makes sense for end user packages. |
23 |
> > For pure dependencies with consistent conventions (e.g. perl), it |
24 |
> > makes sense to keep upstream's naming. |
25 |
> |
26 |
> I'm not advocating renaming this, but I found an example of this when |
27 |
> looking to package something: |
28 |
> |
29 |
> dev-python/configargparse is called ConfigArgParse upstream. |
30 |
> If we had named it dev-python/ConfigArgParse, we wouldn't need to set |
31 |
> MY_PN, MY_P or S in our ebuild, and I wouldn't have had to check the |
32 |
> package to see if it was the same as the package I need to depend on. |
33 |
> |
34 |
> William |
35 |
> |
36 |
|
37 |
It gets worse than that. I recently added several pkg to the tree that |
38 |
had slightly different names for the github repo than they publishedin pypi. |
39 |
I in turn ended up naming it slightly different to fit our |
40 |
category/pkg system. So, the end result is not one common name, but |
41 |
three. For some, I tried to stick with the github repo name, despite |
42 |
getting the tarball from pypi due to the inconsistent github sha's. |
43 |
|
44 |
-- |
45 |
Brian Dolbec <dolsen> |