Package naming question

[Igor Gnatenko]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA256

Hello,

I would like to hear opinion of other packagers about naming. We have
`parallel` utility implemented in Rust which is drop-in replacement for GNU
parallel. I was thinking how to name package and how people would expect it to
be named. So far options are:

* rust-parallel
* parallel-rust
* parallel-rs

I dislike first one because it is not ending up in completion while second and
third are quite good.

Thoughts?
- -- 
- -Igor Gnatenko
> -----BEGIN PGP SIGNATURE-----
>
> iQIzBAEBCAAdFiEEhLFO09aHZVqO+CM6aVcUvRu8X0wFAlolE00ACgkQaVcUvRu8
> X0zcDg/+J4jaboaEmySwsgHiSNDfVqazoxrcZDp0abS8wRO9jSxXCPU0K/MpBgtg
> W3LnJOkEhr4WjRtwoC4MmvZf99sfBo4q6ATfd2hOg0WeYkwP36Fu9lVZUAEP6V0E
> zVNp44UCuxJNU9IxUvop4XVJOTmaF0Be3pGrJ2LRliZkIgJgzVwOk8pkv/eO1iwp
> Y5j2VK7Wjs1Z2/tuL6/u02GqZOpD/0NE145awQUp2PFNwIwYeP70UZ5OjoqdLPNO
> LAWybGPSuoaLN7xT7F3FFvqUBhU6Ypsnc2trd01DyJ81v8ITxeAxZpyMy/jLsoyU
> 0RyP5UJyzI/joGNLHDQimmBnBH3SOcVr68Sx2TIMLxmsyCFaGmv+ehZlTkRWocnM
> QWc0fKOD5FC9W7NY21NOCADG6tuIjj+n8Ncw27gbZ0gexbq/drcq9J7aaiKZ0USf
> HA7iKya6OiuSKI2GAzKGMcBzp3GKkcUl+3NSBOw2bB/fF1NneXJDL4FkQVhvC+Ta
> G9Hmz3MC+f/amozta7y+pPAzswf5yzArc1Lcw/HWx3QG3S/DAlR+fnJntf+Rgz6I
> KRYyasi+RLUizzxBnlOXsi3PM3ppErHXmCfewyyhNup0T+DQsW3Na9/q6nWea/1+
> I2WA3kZEmoYHEZM5d3r4I3vPliz0QXJQjNkpicqCxg+A85CBH44=
> =TEXI
> -----END PGP SIGNATURE-----

[Adam Samalik]

What about an analogy with 'createrepo'? We have:

* createrepo
* createrepo_c

Using this same analogy:

* parallel_rust
* parallel_rs

On Mon, Dec 4, 2017 at 10:20 AM, Igor Gnatenko <
ignatenkobrain(a)fedoraproject.org&gt; wrote:

>  -----BEGIN PGP SIGNED MESSAGE-----
>  Hash: SHA256
>
>  Hello,
>
>  I would like to hear opinion of other packagers about naming. We have
>  `parallel` utility implemented in Rust which is drop-in replacement for GNU
>  parallel. I was thinking how to name package and how people would expect
>  it to
>  be named. So far options are:
>
>  * rust-parallel
>  * parallel-rust
>  * parallel-rs
>
>  I dislike first one because it is not ending up in completion while second
>  and
>  third are quite good.
>
>  Thoughts?
>  - --
>  - -Igor Gnatenko
> >  -----BEGIN PGP SIGNATURE-----
> >
> >  iQIzBAEBCAAdFiEEhLFO09aHZVqO+CM6aVcUvRu8X0wFAlolE00ACgkQaVcUvRu8
> >  X0zcDg/+J4jaboaEmySwsgHiSNDfVqazoxrcZDp0abS8wRO9jSxXCPU0K/MpBgtg
> >  W3LnJOkEhr4WjRtwoC4MmvZf99sfBo4q6ATfd2hOg0WeYkwP36Fu9lVZUAEP6V0E
> >  zVNp44UCuxJNU9IxUvop4XVJOTmaF0Be3pGrJ2LRliZkIgJgzVwOk8pkv/eO1iwp
> >  Y5j2VK7Wjs1Z2/tuL6/u02GqZOpD/0NE145awQUp2PFNwIwYeP70UZ5OjoqdLPNO
> >  LAWybGPSuoaLN7xT7F3FFvqUBhU6Ypsnc2trd01DyJ81v8ITxeAxZpyMy/jLsoyU
> >  0RyP5UJyzI/joGNLHDQimmBnBH3SOcVr68Sx2TIMLxmsyCFaGmv+ehZlTkRWocnM
> >  QWc0fKOD5FC9W7NY21NOCADG6tuIjj+n8Ncw27gbZ0gexbq/drcq9J7aaiKZ0USf
> >  HA7iKya6OiuSKI2GAzKGMcBzp3GKkcUl+3NSBOw2bB/fF1NneXJDL4FkQVhvC+Ta
> >  G9Hmz3MC+f/amozta7y+pPAzswf5yzArc1Lcw/HWx3QG3S/DAlR+fnJntf+Rgz6I
> >  KRYyasi+RLUizzxBnlOXsi3PM3ppErHXmCfewyyhNup0T+DQsW3Na9/q6nWea/1+
> >  I2WA3kZEmoYHEZM5d3r4I3vPliz0QXJQjNkpicqCxg+A85CBH44=
> >  =TEXI
> >  -----END PGP SIGNATURE----- 
>  _______________________________________________
>  devel mailing list -- devel(a)lists.fedoraproject.org
>  To unsubscribe send an email to devel-leave(a)lists.fedoraproject.org
>  


-- 

Adam Šamalík
---------------------------
Software Engineer
Red Hat

[Igor Gnatenko]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA256

On Mon, 2017-12-04 at 10:50 +0100, Adam Samalik wrote:
>  What about an analogy with 'createrepo'? We have:
>  
>  * createrepo
>  * createrepo_c 
That's only because createrepo_c is upstream name.

>  Using this same analogy:
>  
>  * parallel_rust
>  * parallel_rs 
This is really worst user experience.
- -- 
- -Igor Gnatenko
> -----BEGIN PGP SIGNATURE-----
>
> iQIzBAEBCAAdFiEEhLFO09aHZVqO+CM6aVcUvRu8X0wFAlolHQEACgkQaVcUvRu8
> X0yaJQ/+L9WLnv2YtiRmndKs5x08R5pBb9wE6rqMCVKcdig7zvQindrJctBv7+aO
> DfpqGCCsYk54uVsv3vhoHWHoA4IwYm/AvkAyI52b9/wo0AxXvZikS0VwbqPNUuAP
> 5thG/ojTCJF7fpYHQzaoI5SIIPLyHPFksKWsYfsp08tnZ2Y8OpJeBDDE+prP0I73
> +zLD376JpfqYm7JEVUuZqIWNRrks8PlJqxzVA0yc+Py0h5eixA+ehWUDwZrvoHV6
> Vbf8taUKVwOBtH7WlWS+Eq6GDmk5CKDdEoq3Qvrec6vr8LSFKi2mWN6Jipqry8dc
> psokbiJhYG/sBGZgdaZvea4lCHW7isQHKiB4yAT0WvodphNBASfmGkQH6gKyVCL+
> eeToRMbFEX0jEj5Y58qJW5alY7MWGJrgLjTaQd8hovHV0e0Xs3M2PLecIqbt2+a3
> aZnpDa/YvcaA6zl1D6wED7ugumKdznP5iIeNMJU6gClfV48+Ox/ho4U4ABE9plle
> 0RAWHoHjppWvbBdhWINq1w16RI2+HqCCKcnBV9mW3znr01/9F0H7OZ3Wi99YKYs6
> EiftkxYrGE1xRXyTSbk8L4EXiTDZxglBtWJXGUhLekAx9wMlYYb9Vh6cx8GJOyVU
> bcvQfhLzdlh85hKiXM4hRXnKaqtdfpZ4e7U6pM5SK574WP+Osoc=
> =zzi0
> -----END PGP SIGNATURE-----

[Martin Sehnoutka]

Hi,

I'd go with parallel-rs as the "-rs" ending is used often with Rust
projects. e.g. dbus-rs or gtk-rs.

Martin

On 12/04/2017 11:20 AM, Igor Gnatenko wrote:
>  Hello,
>  
>  I would like to hear opinion of other packagers about naming. We have
>  `parallel` utility implemented in Rust which is drop-in replacement for GNU
>  parallel. I was thinking how to name package and how people would expect it to
>  be named. So far options are:
>  
>  * rust-parallel
>  * parallel-rust
>  * parallel-rs
>  
>  I dislike first one because it is not ending up in completion while second and
>  third are quite good.
>  
>  Thoughts?
>  _______________________________________________
>  Rust mailing list -- rust(a)lists.fedoraproject.org
>  To unsubscribe send an email to rust-leave(a)lists.fedoraproject.org
>

[Gerald Henriksen]

On Mon, 04 Dec 2017 10:20:13 +0100, you wrote:

> I would like to hear opinion of other packagers about naming. We have
> `parallel` utility implemented in Rust which is drop-in replacement for GNU
> parallel. I was thinking how to name package and how people would expect it to
> be named. So far options are:
>
> * rust-parallel
> * parallel-rust
> * parallel-rs
>
> I dislike first one because it is not ending up in completion while second and
> third are quite good. 
To me the second and third make it look like the package is part of
the existing parallel package - perhaps rust bindings to parallel -
and not an entirely different program.

In other words, I could see people install parallel-rust because "it
must be part of the parallel package" based on its naming and not
realize it is an entirely different program doing the same thing.

[Zbigniew Jędrzejewski-Szmek]

On Mon, Dec 04, 2017 at 09:19:26AM -0500, Gerald Henriksen wrote:
>  On Mon, 04 Dec 2017 10:20:13 +0100, you wrote:
>  
>  >I would like to hear opinion of other packagers about naming. We have
>  >`parallel` utility implemented in Rust which is drop-in replacement for GNU
>  >parallel. I was thinking how to name package and how people would expect it to
>  >be named. So far options are:
>  >
>  >* rust-parallel
>  >* parallel-rust
>  >* parallel-rs
>  >
>  >I dislike first one because it is not ending up in completion while second and
>  >third are quite good.
>  
>  To me the second and third make it look like the package is part of
>  the existing parallel package - perhaps rust bindings to parallel -
>  and not an entirely different program.
>  
>  In other words, I could see people install parallel-rust because "it
>  must be part of the parallel package" based on its naming and not
>  realize it is an entirely different program doing the same thing. 
Yes. In other words, the problem started when somebody decided to
reimplement a well-known existing package in a different language without
changing the name. It seems like the right solution is to ask the
rust-parallel folks to rename their project instead of squatting on an
existing name.

Zbyszek

[Jun Aruga]

Perl: perl-foo
Python: python-foo
NodeJs (NPM): nodejs-foo
Ruby: rubygem-foo
R: R-foo
PHP: php-foo
Golang:  golang-foo

The prefix pattern "rust-parallel" looks better.

Jun


On Mon, Dec 4, 2017 at 3:45 PM, Zbigniew Jędrzejewski-Szmek <
zbyszek(a)in.waw.pl&gt; wrote:

>  On Mon, Dec 04, 2017 at 09:19:26AM -0500, Gerald Henriksen wrote:
>  > On Mon, 04 Dec 2017 10:20:13 +0100, you wrote:
>  >
>  > >I would like to hear opinion of other packagers about naming. We have
>  > >`parallel` utility implemented in Rust which is drop-in replacement for
>  GNU
>  > >parallel. I was thinking how to name package and how people would
>  expect it to
>  > >be named. So far options are:
>  > >
>  > >* rust-parallel
>  > >* parallel-rust
>  > >* parallel-rs
>  > >
>  > >I dislike first one because it is not ending up in completion while
>  second and
>  > >third are quite good.
>  >
>  > To me the second and third make it look like the package is part of
>  > the existing parallel package - perhaps rust bindings to parallel -
>  > and not an entirely different program.
>  >
>  > In other words, I could see people install parallel-rust because "it
>  > must be part of the parallel package" based on its naming and not
>  > realize it is an entirely different program doing the same thing.
>
>  Yes. In other words, the problem started when somebody decided to
>  reimplement a well-known existing package in a different language without
>  changing the name. It seems like the right solution is to ask the
>  rust-parallel folks to rename their project instead of squatting on an
>  existing name.
>
>  Zbyszek
>  _______________________________________________
>  devel mailing list -- devel(a)lists.fedoraproject.org
>  To unsubscribe send an email to devel-leave(a)lists.fedoraproject.org
>  


-- 
Jun Aruga jaruga(a)redhat.com
IRC: jaruga, Office: TPB(Technology Park Brno) Building C 1F, Brno, Czech
Republic

[Fedora-packaging] Re: Package naming question

[Jason L Tibbitts III]

> >>>> "JA" == Jun Aruga <jaruga(a)redhat.com&gt;
> writes: 
JA> Perl: perl-foo Python: python-foo NodeJs (NPM): nodejs-foo Ruby:
JA> rubygem-foo R: R-foo PHP: php-foo Golang: golang-foo

Those are all for libraries in the given language.

JA> The prefix pattern "rust-parallel" looks better.

This is not a library written in rust.

https://fedoraproject.org/wiki/Packaging:Guidelines#Library_or_Applicatio...

If you discount the guidelines for multiple parallel-installable
versions of the same package, we don't really have an established naming
convention for alternate versions of a "thing".  Best I can think of are
things like libart_lgpl (which doesn't and maybe never did have a
non-lgpl version), reentrant versions of libraries like libqhull_r
(which has the least useful package %description I've ever seen), though
those aren't generally packaged separately, or, I don't know,
chromium-libs-media-freeworld (which I know isn't a Fedora package).

Honestly I'd just pretend that what you have really _is_ a different
version of the same package, but instead of being version '3', it's
version 'rust'.  And for that we have
https://fedoraproject.org/wiki/Packaging:Naming#Multiple_packages_with_th...
which would suggest "parallel-rust".

For the actual executables, we have a long tradition of prefixing 'k'
for kerberized versions of things, and postfixing version numbers,
sometimes with dashes.  Which really doesn't clarify much of anything.
About all we can say with certainty is that this package can't install
/usr/bin/parallel.

 - J<

[Fedora-packaging] Re: Package naming question

[Jun Aruga]

Jason,

>  If you discount the guidelines for multiple parallel-installable
>  versions of the same package, we don't really have an established naming
>  convention for alternate versions of a "thing". 
Thanks for mentioning it and the detailed explanation.
I did misunderstand it.
I understand what you are talking about.
Yeah, the suffix style "parallel-rust" looks better.

Jun


On Mon, Dec 4, 2017 at 11:10 PM, Jason L Tibbitts III <tibbs(a)math.uh.edu&gt;
wrote:

>  >>>>> "JA" == Jun Aruga
> <jaruga(a)redhat.com&gt; writes:
>
>  JA> Perl: perl-foo Python: python-foo NodeJs (NPM): nodejs-foo Ruby:
>  JA> rubygem-foo R: R-foo PHP: php-foo Golang: golang-foo
>
>  Those are all for libraries in the given language.
>
>  JA> The prefix pattern "rust-parallel" looks better.
>
>  This is not a library written in rust.
>
>  https://fedoraproject.org/wiki/Packaging:Guidelines#
>  Library_or_Application.3F
>
>  If you discount the guidelines for multiple parallel-installable
>  versions of the same package, we don't really have an established naming
>  convention for alternate versions of a "thing".  Best I can think of are
>  things like libart_lgpl (which doesn't and maybe never did have a
>  non-lgpl version), reentrant versions of libraries like libqhull_r
>  (which has the least useful package %description I've ever seen), though
>  those aren't generally packaged separately, or, I don't know,
>  chromium-libs-media-freeworld (which I know isn't a Fedora package).
>
>  Honestly I'd just pretend that what you have really _is_ a different
>  version of the same package, but instead of being version '3', it's
>  version 'rust'.  And for that we have
>  https://fedoraproject.org/wiki/Packaging:Naming#
>  Multiple_packages_with_the_same_base_name
>  which would suggest "parallel-rust".
>
>  For the actual executables, we have a long tradition of prefixing 'k'
>  for kerberized versions of things, and postfixing version numbers,
>  sometimes with dashes.  Which really doesn't clarify much of anything.
>  About all we can say with certainty is that this package can't install
>  /usr/bin/parallel.
>
>   - J<
>  


-- 
Jun Aruga jaruga(a)redhat.com
IRC: jaruga, Office: TPB(Technology Park Brno) Building C 1F, Brno, Czech
Republic