Re: File moved from one package to another - use Conflicts/Replaces or Breaks/Replaces?

[Otto Kekäläinen]

Hell!

On Sun, Apr 4, 2021 at 4:54 AM Andreas Metzler  wrote:
>
> On 2021-04-03 Otto Kekäläinen  wrote:
> > Hello!
>
> > In MariaDB we have over the years moved files around. A file that was first
> > in e.g. mariadb-server-10.3 might have been moved to
> > mariadb-server-core-10.3 and some years later to mariadb-client-core-10.5.
> > The result is a massive debian/control file with a lot of
> > Conflicts/Breaks/Replaces [1]
>
> > While trying to clean up and simplify I started wondering, when should one
> > use Conflicts/Replaces and when Breaks/Replaces?
> [...]
>
> The rough guideline is to use Breaks together with an earlier than (<<,
> <=) condition. Which seems to be what mariadb does.

OK

> I do not think there is a lot to improve here. Did you you see David's
> recent message, <20210331164012.gevwmagm63q2yc54@crossbow>?

Yes, and I solved the original issue already using the
'apt install --with-source ./Packages -s' technique.

Re: Tips for debugging/testing debian/control Depends/Breaks etc changes?

[Otto Kekäläinen]

Hello!

> For me, this whole situation seems wrong though. Why do you have
> versioned package names (mariadb-server-*) when they are all mutually
> exclusive with one another due to all shipping the same binary?

This is how the MySQL packaging has been structured for the past 10
years or so. There was probably some use case that users wanted to
stick to a certain major release by installing e.g. 'mysql-server-5.1'
while installing 'mysql-server' would just auto-upgrade the major
version on dist-upgrade. I don't know for sure, this predates me by
years.

I'd rather not start changing it now due to the potential unknown and
hard to revert consequences it could have to users.

I did however a couple years ago remove the versions from the plugin
package names, so the issue is at least limited to server and client
packages.

For the record, I solved the original issue already using the 'apt
install --with-source ./Packages -s' technique.

Re: Realtek RTL8723DE, RTL8821CE, RTL8822BE and RTL8822CE chipsets

[Agustin Martin]

El dom, 4 abr 2021 a las 12:28, Devops PK Carlisle LLC
() escribió:
>
> This is interesting.
>
> First, I must ask... Are either Ralink or Atheros more quickly adopted
> into Linux support before Realtek? Are they more stable, etc.? (In my
> original post, I noted that I pulled a Realtek based 5ghz dongle from an
> extreme end user system, in no small part due to stability/reliability)?

I do not think so, I guess this mostly depends on how many people are
involved with each driver and how well tested is the code.

In the particular case of rtw88 driver (the one supporting chipsets in
the subject), it is maintained by Realtek folks with additional
contributions from the community. Some chipsets are supported, some
not yet. For the case I know (rtl8821ce), original chipset is
supported, but support for other variants (e.g., rfe 2) was late for
5.11 (Realtek tests took longer than expected), but will be present in
5.12 kernel, requiring new firmware
(https://lkml.org/lkml/2021/2/2/76). I think the best option is to
wait for that kernel and firmware to reach debian-backports. For other
variants of other chipsets things may be fifferent.

Regards,

-- 
Agustin

Bug#986421: ITP: nxtrim -- Optimized trimming of Illumina mate pair reads

[Nilesh Patra]

Package: wnpp
Severity: wishlist
Owner: Nilesh Patra 
X-Debbugs-Cc: debian-devel@lists.debian.org, nil...@debian.org

* Package name: nxtrim
  Version : 0.4.3+dfsg
  Upstream Author : Illumina, Inc.
* URL : https://github.com/sequencing/NxTrim
* License : BSD
  Programming Lang: C++
  Description : Optimized trimming of Illumina mate pair reads
  This package helps rmove Nextera Mate Pair junction adapters
  and categorise reads according to the orientation implied
  by the adapter location.

  I shall maintain this package

debtags.debian.org tag vocabulary now editable by all DDs

[Enrico Zini]

Hello,

 * Debian Developers can add new tags to Debtags

I have now added group Debian to the Salsa repository for the Debtags
tag vocabulary[1]. This means that any Debian Developer can now add tags
to Debtags. People who are not Debian Developers are welcome to submit
merge requests, which any Debian Developer can approve.


 * Vocabulary updates are automatically deployed

Pushing to master triggers a quick CI job that checks the vocabulary for
consistency. If the CI passes, the new tag vocabulary is automatically
deployed to debtags.debian.org.

Hopefully this means that I'm finally removing myself as a bottleneck
for one of the maintenance points to Debtags!


 * I need help with routine dput-ting of tag data

I could still use help from one or more DDs, doing regular uploads of
tags from the site to the archive.

It means running a script that generates a source package, GPG-signing
it, and uploading it with dput. It needs to be done once a month or so,
and I regularly miss doing that. If you're interested, get in touch and
we can go through the script to make it work on your system.


Enrico

[1] https://salsa.debian.org/debtags-team/debtags-vocabulary
-- 
GPG key: 4096R/634F4BD1E7AD5568 2009-05-08 Enrico Zini 


signature.asc
Description: PGP signature

Bug#986435: ITP: libthrift-java -- Java language support for Thrift

[Markus Koschany]

Package: wnpp
Severity: wishlist
Owner: Markus Koschany 
X-Debbugs-Cc: debian-devel@lists.debian.org, 938...@bugs.debian.org, 
debian-j...@lists.debian.org, a...@debian.org

* Package name: libthrift-java
  Version : 0.13.0
  Upstream Author : Apache Software Foundation
* URL : https://thrift.apache.org/
* License : Apache-2.0
  Programming Lang: Java
  Description : Java language support for Thrift

 Thrift is a software framework for the development of reliable and
 performant communication and data serialization. It combines a
 software stack with code generation to build services that operate
 seamlessly across a number of different development languages.
 .
 This package provides the Java language support for Thrift.

libthrift-java is a new build-dependency for apache-jena, the latest
version of mediathekview and possibly cassandra (ITP: #585905). Since
there was no reaction to #938927 for 1,5 years now, I intend to
re-introduce the old libthrift-java package and maintain it within the
Debian Java team.

Re: Realtek RTL8723DE, RTL8821CE, RTL8822BE and RTL8822CE chipsets

[Michael Stone]

On Sat, Apr 03, 2021 at 08:03:23PM +0500, Andrey Rahmatullin wrote:

On Sat, Apr 03, 2021 at 10:37:37AM -0400, Michael Stone wrote:

> > > Not sure what hardware you are talking about but the majority of WiFI
> > > hardware is supported by the mainline kernels, at least after you load
> > > their firmware.
> >
> > I assume you haven't tried very much wifi hardware. Realistically, the state
> > of wifi support is still terrible. The best thing to do is try to buy
> > something known to be supported, but that's relatively difficult for most
> > people because the name on the box generally has nothing to do with the
> > chips inside the box.
> Can you please list some unsupported chips in addition to these specific
> Realtek ones?

It would be easier for you to list ones that you've actually tested and know
work.


OK, whatever.


It's a serious request. In 20 years of trying different wireless devices
I've had just one that's completely reliable under linux (atheros 
qca6174 chipset) even if it's a little slow (only 2x2 mimo, and possibly 
hampered by the laptop's antenna configuration). It's older 802.11ac; I 
haven't ever seen a working 802.11ax. (Not saying it doesn't exist, just 
that I haven't seen it--hence the question!) I've seen lots of reports 
of reliable devices, but they tend to be things that are either ancient, 
or hard to find, or simply don't work as well for me as for others 
(apparently). I've personally used a lot of devices that mostly worked, 
until they hung or started going really slow or in some other way 
behaved much worse than the phone sitting next to the laptop (ironically 
running linux). I've heard good things about the intel adapters, but 
AFAIK they aren't available in USB so if someone didn't happen to get a 
laptop with that solution integrated it's not a simple end-user 
addition.


IMO, wifi is the last major pain point to a fully functional linux 
sytem. Most other things these days you can just buy a something off the 
shelf at the local electronics store and odds are all the pieces will 
work out of the box. But wifi still requires a good bit of research, and 
isn't something a novice will find simple unless they get lucky on their 
components or bought something from a specialty retailer that focuses on 
linux systems.

Bug#986437: ITP: golang-github-jedib0t-go-pretty -- Pretty print tables, lists, progress-bars and text

[Francisco Vilmar Cardoso Ruviaro]

Package: wnpp
Severity: wishlist
Owner: Francisco Vilmar Cardoso Ruviaro 
X-Debbugs-Cc: debian-devel@lists.debian.org, francisco.ruvi...@riseup.net

* Package name: golang-github-jedib0t-go-pretty
  Version : 6.1.0
  Upstream Author : Naveen Mahalingam 

* URL : https://github.com/jedib0t/go-pretty
* License : Expat
  Programming Lang: Go
  Description : Pretty print tables, lists, progress-bars and text

Utilities to prettify console output of tables, lists, progress-bars and text.
.
Formatting  | Description
Table   | Pretty-print tables into ASCII/Unicode strings.
List| Pretty-print lists with multiple levels/indents
  into ASCII/Unicode strings.
Progress| Track the Progress of one or more Tasks
  (like downloading multiple files in parallel).
Text| Utility functions to manipulate text with
  or without ANSI escape sequences.

Tone policing by a member of the community team [Was, Re: Statement regarding Richard Stallman's readmission to the FSF board]

[Steve Langasek]

Hi all,

After some long thought, I believe this message warrants a public response
(and discussion).

The facts are these:

 - an individual who is not part of the Debian community sent me (and other
   people) a private, unsolicited email sharing his views on the current
   topic of the day.
 - I responded, privately, telling him exactly what I thought of him and his
   views.
 - He in turn forwarded my private responses to a public mailing list
   without my consent.

The response by Jean-Philippe, a member of the Debian Community Team, was a
call for de-escalation and civility "by both sides"; i.e., tone policing.

While I am no oppressed minority who is going to be turned away from Debian
as a result of such tone policing, a member of the Community Team tone
policing an ally who is categorically rejecting transphobia sends a very bad
message to trans members of our community.  It shows that responding to
transphobes by communicating using strong language that the Debian community
rejects their views leaves one potentially subject to censure from the
Community Team.

I think the Community Team should do better.

On Fri, Mar 26, 2021 at 09:21:32PM +0100, Jean-Philippe MENGUAL wrote:
> Hi,
> 
> Please descalate, it is an emergency. The RMS debate is very difficult for
> many people I guess, painful or, at least, energy consumer. It often results
> to ghoughts about the compliance of some mails with the code of conduct. But
> those 2 mails are clearly not acceptable from the formal point of view. You
> cannot let yourself dominate by your badest feelings, injuring each other
> etc. I am sure that is completely out of the code of conduct. Be respectful,
> avoiding to say words such as "shit" "fuck" "fascist", and please try
> staying moderated. Debian will be a welcoming community if anyone can
> express his opinons politely and others reply politely, with respect, even
> if he disagrees. So really, descalte! If the thread is so abrasive, just
> don't reply, ignore, I am sure it will not change your daily life in your
> team or in the project. Or wait some hours before doing it. But again, such
> wordings are unacceptable, regardless the topic itself and the ideas. Keep
> in mind you speak publicly and your mails stay in archives forever.
> 
> Thanks in advance
> 
> Regards
> 
> 
> 
> Jean-Philippe MENGUAL
> Debian Developer non uploading
> Community team member
> Accessibility team member
> debian-l10n-french team member
> President of Debian France non-profit organization
> Le 26/03/2021 à 20:22, Michael Shigorin a écrit :
> > On Fri, Mar 26, 2021 at 11:23:21AM -0700, Steve Langasek wrote:
> > > On Fri, Mar 26, 2021 at 09:14:16PM +0300, Michael Shigorin wrote:
> > > > On Fri, Mar 26, 2021 at 09:18:19AM -0700, Steve Langasek wrote:
> > > > > > You absolutely have NO right to speak for all of the community
> > > > ..so do go and apologize for the attempt in public.
> > > 
> > > > > > And I tell you that you're humanophobic by claiming someone
> > > > > > is "transphobic".
> > > > > Fuck off, nobody asked you for your shitty fascist opinion.
> > > 
> > > > It's *you* who's a fascist here.
> > > > Go read the definition and look at what *you* do.
> > 
> > ...e.g., http://dictionary.com/browse/fascism
> > 
> > > > Both of my grandfathers actually fought Nazi and Japanese
> > > > invaders.  Looks like none of yours ever confronted them --
> > > > they would have raised you a human, not a fascist.
> > > 
> > > > But you will learn the hard way.
> > > 
> > > > Good luck.
> > > 
> > > Not very good at fucking off, are you?
> > 
> > Yep.
> > 
> > Jonathan, I hereby demand that the Debian Project gets rid
> > of this manipulative, insultive, divisive and libelous member.
> > He (them? it?) can't even stand by the rules (pro|im)posed.
> > 
> > I'm considering providing financial support to Chris Punches
> > and asking him to not omit suing Steve Langasek by chance:
> > http://web.archive.org/web/20210326090023/https://github.com/rms-open-letter/rms-open-letter.github.io/issues/2250
> > (of course the issue was deleted, mindless as it could be).
> > 
> 

-- 
Steve Langasek   Give me a lever long enough and a Free OS
Debian Developer   to set it on, and I can move the world.
Ubuntu Developer   https://www.debian.org/
slanga...@ubuntu.com vor...@debian.org


signature.asc
Description: PGP signature