Re: New security/py-pysha3

2019-09-24 Thread clematis
> On Mon, Sep 23, 2019 at 11:05:09PM +0100, Stuart Henderson wrote:
> > On 2019/09/23 16:17, Kurt Mosiejczuk wrote:
> > > Did we really want to name it py-pysha3? I figure py-sha3 is what we would
> > > usually do in that case. (See pytest -> py-test, pydot -> py-dot, others)
> 
> > We have it both ways, personally I prefer it this way (portgen does
> > this also) but I don't mind much either way.

I did notice we weren't necessarily always consistent with the naming.
I chose this way for a couple of reasons. (I am not saying this is the
right way). 

1/ Be consistent with other platform and faciliate the search. The
module name is 'pydot'. People wondering if it's available in ports
will search for 'pydot'.
You will not find it, not with pkg_info -Q, not with make search
key=pydot, not on openports.se (Package Name). (But you will find
'pygal' for example). Potentially you would have noticed some py-* so
will search for py-pydot, or py-pygal. But will again find only the second
one.

2/ Prevent collision. To come back to this pysha3 ports, a 'sha3' [1]
module does exist. What would be his package name? py-sha3.
Would be a problem if we name 'pysha3' this way.

I understand this could feel awkward for the modules py-py*

But long story short, PKGNAME = py-${DISTNAME} makes more sense to me than 
PKGNAME= ${DISTNAME:S/py/py-/}.

But this is just my opinion. If there's another approach I would be
happy to stick with it in the future and stay consistent.


[1] https://pypi.org/project/sha3/
-- 
clematis (0x7e96fd2400fe7b59)



Re: New security/py-pysha3

2019-09-23 Thread Stuart Henderson
On 2019/09/23 16:17, Kurt Mosiejczuk wrote:
> On Mon, Sep 23, 2019 at 08:47:24PM +0100, Stuart Henderson wrote:
> > On 2019/09/23 17:52, clematis wrote:
> 
> > Slightly tweaked version attached. s/sha/SHA/ in COMMENT, and add the
> > missing WANTLIB/LIB_DEPENDS. This one is OK sthen@ if somebody would like
> > to import it.
> 
> Did we really want to name it py-pysha3? I figure py-sha3 is what we would
> usually do in that case. (See pytest -> py-test, pydot -> py-dot, others)
> 
> Otherwise I'm fine with the port (and can import).
> 
> --Kurt
> 

We have it both ways, personally I prefer it this way (portgen does
this also) but I don't mind much either way.

But I've just noticed, this is a backport and not needed for py3.6+,
so we should kill the flavour.



Re: New security/py-pysha3

2019-09-23 Thread Stuart Henderson
On 2019/09/23 17:52, clematis wrote:
> On Mon, Sep 23, 2019 at 02:36:45PM +0100, Stuart Henderson wrote:
> > that is kmos@
> Oops wrong Kurt. You're right, this was kmos@. Sorry. 
> 
> > : COMMENT = python SHA-3 implementation
> > 
> > as a proper noun Python should start with caps.
> Ok, I've reformulated it to take this into account and to not start with
> an uppercase. 
>  
> > nitpicking, but you have =, please
> > zap the extra spaces and just use tabs there
> Done.
> 
> > MODPY_SETUPTOOLS already adds to dependencies as necessary
> > (RUN_DEPENDS and BUILD_DEPENDS; TEST_DEPENDS isn't needed for something
> > which is already in BUILD_DEPENDS because you can't test until you've
> > built it).
> Understood. So I've removed both RUN and TEST _DEPENDS to only keep
> MODEPY_SETUPTOOLS.
> 
> Build and install OK on amd64. 
> I've re-run the regression tests for both flavors and all 50 completed
> OK.
> 
> See attached for a new tarball including these changes.
>  
> Thanks,
> -- 
> clematis (0x7e96fd2400fe7b59)


Slightly tweaked version attached. s/sha/SHA/ in COMMENT, and add the
missing WANTLIB/LIB_DEPENDS. This one is OK sthen@ if somebody would like
to import it.




py-pysha3.tgz
Description: application/tar-gz


Re: New security/py-pysha3

2019-09-23 Thread clematis
On Mon, Sep 23, 2019 at 02:36:45PM +0100, Stuart Henderson wrote:
> that is kmos@
Oops wrong Kurt. You're right, this was kmos@. Sorry. 

> : COMMENT =   python SHA-3 implementation
> 
> as a proper noun Python should start with caps.
Ok, I've reformulated it to take this into account and to not start with
an uppercase. 
 
> nitpicking, but you have =, please
> zap the extra spaces and just use tabs there
Done.

> MODPY_SETUPTOOLS already adds to dependencies as necessary
> (RUN_DEPENDS and BUILD_DEPENDS; TEST_DEPENDS isn't needed for something
> which is already in BUILD_DEPENDS because you can't test until you've
> built it).
Understood. So I've removed both RUN and TEST _DEPENDS to only keep
MODEPY_SETUPTOOLS.

Build and install OK on amd64. 
I've re-run the regression tests for both flavors and all 50 completed
OK.

See attached for a new tarball including these changes.
 
Thanks,
-- 
clematis (0x7e96fd2400fe7b59)


security_py-pysha3.tar.gz
Description: application/tar-gz


Re: New security/py-pysha3

2019-09-23 Thread Stuart Henderson
On 2019/09/23 15:21, clematis wrote:
> Hi Team,
> 
> kurt@ pointed out that I've missed the pysha3 dependencies working on
> devel/py-merkletools. 

that is kmos@

> Attached is a first submission for security/py-pysha3.
> 
> From previous conversations I understood devel/ was getting busy so
> security/ sounded like the best fit.
> All regression tests (50 tests) completed OK on amd64 for *both* flavors. 
> This does satisfy the devel/py-merkletools dep but I am still running
> further testing on that specific one with both flavors. I will update
> the other email thread once I am done.

: COMMENT = python SHA-3 implementation

as a proper noun Python should start with caps.

: MAINTAINER =  Clem Atis 

nitpicking, but you have =, please
zap the extra spaces and just use tabs there

: RUN_DEPENDS = devel/py-setuptools${MODPY_FLAVOR}
: TEST_DEPENDS =devel/py-setuptools${MODPY_FLAVOR}
: 
: MODPY_PI =Yes
: MODPY_SETUPTOOLS =Yes

MODPY_SETUPTOOLS already adds to dependencies as necessary
(RUN_DEPENDS and BUILD_DEPENDS; TEST_DEPENDS isn't needed for something
which is already in BUILD_DEPENDS because you can't test until you've
built it).



New security/py-pysha3

2019-09-23 Thread clematis
Hi Team,

kurt@ pointed out that I've missed the pysha3 dependencies working on
devel/py-merkletools. 

Attached is a first submission for security/py-pysha3.

>From previous conversations I understood devel/ was getting busy so
security/ sounded like the best fit.
All regression tests (50 tests) completed OK on amd64 for *both* flavors. 
This does satisfy the devel/py-merkletools dep but I am still running
further testing on that specific one with both flavors. I will update
the other email thread once I am done.

Comments?

Thanks,
-- 
clematis (0x7e96fd2400fe7b59)


security_py-pysha3.tar.gz
Description: application/tar-gz