Fwd: [Catalog-sig] Trove categories for Buildout recipes and extensions
Hi,
I am forwarding a mail from Catalog-sig regarding two new trove categories
for Buildout recipes and extensions. Those who publish Buildout recipes and
extensions to PyPI, please use the new trove classifiers in addition to
the existing "Framework :: Buildout".
If you are developing a recipe, use this classifier:
classifiers=[
...
"Framework :: Buildout",
"Framework :: Buildout :: Recipe",
...
],
If you are developing an extension, use this classifier:
classifiers=[
...
"Framework :: Buildout",
"Framework :: Buildout :: Extension",
...
],
If the distribution contains both extension and recipe, use:
classifiers=[
...
"Framework :: Buildout",
"Framework :: Buildout :: Extension",
"Framework :: Buildout :: Recipe",
...
],
Regards,
Baiju M
---------- Forwarded message ----------
From: Richard Jones
It would be nice if we can add two sub categories:
Framework :: Buildout :: Recipe Framework :: Buildout :: Extension
There being no objections I've added these classifiers. Richard
On 2011-1-26 04:27, Baiju M wrote:
Hi, I am forwarding a mail from Catalog-sig regarding two new trove categories for Buildout recipes and extensions. Those who publish Buildout recipes and extensions to PyPI, please use the new trove classifiers in addition to the existing "Framework :: Buildout".
If you are developing a recipe, use this classifier:
classifiers=[ ... "Framework :: Buildout", "Framework :: Buildout :: Recipe",
Doesn't "Framework :: Buildout :: Recipe" already imply "Framework ::
Buildout" ?
Wichert.
--
Wichert Akkerman
If you are developing a recipe, use this classifier:
classifiers=[ ... "Framework :: Buildout", "Framework :: Buildout :: Recipe",
Doesn't "Framework :: Buildout :: Recipe" already imply "Framework :: Buildout" ?
If it indeed could have meant "Framework :: Buildout :: Extension" instead, and if "Framework :: Buildout :: Extension" and "Framework :: Buildout :: Recipe" are indeed different things, then: no. Regards, Martin
On 1/26/11 10:40 , "Martin v. Löwis" wrote:
If you are developing a recipe, use this classifier:
classifiers=[ ... "Framework :: Buildout", "Framework :: Buildout :: Recipe",
Doesn't "Framework :: Buildout :: Recipe" already imply "Framework :: Buildout" ?
If it indeed could have meant "Framework :: Buildout :: Extension" instead, and if "Framework :: Buildout :: Extension" and "Framework :: Buildout :: Recipe" are indeed different things, then: no.
That's not what I meant exactly. What I meant that is you declare a Trove classifier "A :: B :: C" aren't you implicitly declaring "A :: B" and "A" as well, so you don't need to specify them separately? Wichert.
Doesn't "Framework :: Buildout :: Recipe" already imply "Framework :: Buildout" ?
If it indeed could have meant "Framework :: Buildout :: Extension" instead, and if "Framework :: Buildout :: Extension" and "Framework :: Buildout :: Recipe" are indeed different things, then: no.
That's not what I meant exactly. What I meant that is you declare a Trove classifier "A :: B :: C" aren't you implicitly declaring "A :: B" and "A" as well, so you don't need to specify them separately?
Oops, I read it reverse, sorry. It unfortunately depends on who is interpreting the data. For compatibility, I would err on the safe side and just include the old classifier as well. But you are right; it would be redundant. Regards, Martin
Doesn't "Framework :: Buildout :: Recipe" already imply "Framework :: Buildout" ? [...] That's not what I meant exactly. What I meant that is you declare a Trove classifier "A :: B :: C" aren't you implicitly declaring "A :: B" and "A" as well, so you don't need to specify them separately?
You’re not implicitly declaring “A” because it is not a classifier: They all have at least two parts. In this example: Topic :: Software Development :: Version Control :: CVS these implied classifiers may be useful in your application: Topic :: Software Development :: Version Control Topic :: Software Development however, “Topic” alone is not a classifier, and is not very useful (“this distribution defines one topic”). Maybe we should provide a module in distutils2 to make useful operations with Trove classifiers strings. Regards
participants (4)
-
"Martin v. Löwis" -
Baiju M -
Wichert Akkerman -
Éric Araujo