Bug#930595: RFS: uacme/1.0.15-2 [ITP]
- To: Nicola Di Lieto <nicola.dilieto@gmail.com>
- Cc: 930595@bugs.debian.org
- Subject: Bug#930595: RFS: uacme/1.0.15-2 [ITP]
- From: Adam Borowski <kilobyte@angband.pl>
- Date: Wed, 3 Jul 2019 02:40:30 +0200
- Message-id: <[🔎] 20190703004030.GA9300@angband.pl>
- Reply-to: Adam Borowski <kilobyte@angband.pl>, 930595@bugs.debian.org
- In-reply-to: <20190623081820.xjfgjjorulkjv3lo@einstein.dilieto.eu>
- References: <20190621091341.GA16321@angband.pl> <4bbf384f-35ef-a42b-14cb-5623a2409d5e@gmail.com> <20190621102156.GA17509@angband.pl> <b22e98a4-4195-0ebe-8864-726553eeb5db@gmail.com> <20190622114704.GA23514@angband.pl> <20190622175939.2jgpcr2ycy7ym6md@einstein.dilieto.eu> <20190622190928.GA29825@angband.pl> <20190622221735.doljlisreu73gtqv@einstein.dilieto.eu> <20190622222814.GA1898@angband.pl> <20190623081820.xjfgjjorulkjv3lo@einstein.dilieto.eu> <20190616085200.xlnil3lrkidrf7j5@einstein.dilieto.eu>
(Sorry for slow response time.)
On Sun, Jun 23, 2019 at 10:18:20AM +0200, Nicola Di Lieto wrote:
> https://mentors.debian.net/debian/pool/main/u/uacme/uacme_1.0.16-7.dsc
I haven't actually tested the functionality, but packaging seems almost
ready.
One issue is the watch file being empty (save for comments). If there's
nothing inside, it should be deleted. But, for a Github project that does
sane release, it's too trivial to fill it instead:
.----
version=4
opts=filenamemangle=s/.+\/v(\d\S*)\.tar\.gz/uacme-$1\.tar\.gz/ \
https://github.com/ndilieto/uacme/tags .*/v(\d\S*)\.tar\.gz
`----
The other thing are changelog entries that have never entered the archive.
The changelog is supposed to include only actually released changes -- it's
natural that there are multiple iterations during review, but those are
not interesting to the end-user -- especially these days as we have git
repositories for such minutiae.
Thus, unless there's a good reason, such changelog entries get squashed;
and, since there's no previous release, just a single "Initial release
(Closes: #123456)" is enough. Also, both people and tools tend to get
confused when there's a -7 but no -1.
("Good reasons" include the package being deployed to some actual users,
such as a different distribution or even a company's or a cluster's internal
machines -- in which case, reusing version numbers would cause harm.)
Meow!
--
⢀⣴⠾⠻⢶⣦⠀
⣾⠁⢠⠒⠀⣿⡁ Packager's rule #1: upstream _always_ screws something up. This
⢿⡄⠘⠷⠚⠋⠀ is true especially if you're packaging your own project.
⠈⠳⣄⠀⠀⠀⠀
Reply to: