Hello, I would like to make a suggestion for the apache package, I was not too sure which mailing list to put this in, so I picked the general mailing list for the avoidance of doubt :P So the apache package: https://archlinux.org/packages/extra/x86_64/apache/ Currently contains tools which are used by nginx users, and the issue is that is that you can only run one webserver as once, they both are going to try to bind to port 80 and 443 so you got to choose. htpaswd utility is used by nginx users too, and the issue is they have to install the entire apache web server in order to be able to use the single utility. I suggest splitting the package into apache-utils and apache, or, if the maintainer(s) are willing to maintain it, split the utilities into their own separate pacakge and use apache-utils as a group of all the subpackages (htpasswd, htdigest etc). I feel this would be a good change so that nginx users do not need to install the entire apache web server just to use the apache utilities which nginx suggests to use within their documentation (see https://docs.nginx.com/nginx/admin-guide/security-controls/configuring-http-...) As you can see in the guide, Debian, Ubuntu, RHEL, CentOS and Oracle Linux already have seen this issue within their own repositories and split the packages up, why can't arch linux? Thanks for reading this suggestion, -- Polarian GPG signature: 0770E5312238C760 Website: https://polarian.dev JID/XMPP: polarian@polarian.dev
Le 01/02/2023 à 20:10, Polarian a écrit :
Hello,
Hi,
I would like to make a suggestion for the apache package, I was not too sure which mailing list to put this in, so I picked the general mailing list for the avoidance of doubt :P
So the apache package:
https://archlinux.org/packages/extra/x86_64/apache/
Currently contains tools which are used by nginx users, and the issue is that is that you can only run one webserver as once, they both are going to try to bind to port 80 and 443 so you got to choose.
While waiting for someone else to give an answer about splitting the apache package like other distros do, I'd like to point out that the above statement is false. There's no issue having two (or more) webservers installed at once, and no issues having two (or more) of them running at the same time either, as long as they don't run on the same port. While, apache and nginx will indeed both use the 80 or 443 port by default, this is something you can change in the configuration file. So for people that face the situation described in this mail: You can install the apache package alongside nginx (in order to use `htpasswd`) without issue. If you decide to have them both *running *at the same time, just make sure to bind the `listen` parameter in the default apache and nginx configurations files to a different port for each of them.
htpaswd utility is used by nginx users too, and the issue is they have to install the entire apache web server in order to be able to use the single utility.
I suggest splitting the package into apache-utils and apache, or, if the maintainer(s) are willing to maintain it, split the utilities into their own separate pacakge and use apache-utils as a group of all the subpackages (htpasswd, htdigest etc).
I feel this would be a good change so that nginx users do not need to install the entire apache web server just to use the apache utilities which nginx suggests to use within their documentation (see https://docs.nginx.com/nginx/admin-guide/security-controls/configuring-http-...)
As you can see in the guide, Debian, Ubuntu, RHEL, CentOS and Oracle Linux already have seen this issue within their own repositories and split the packages up, why can't arch linux?
As I said earlier, I'll let other people react to this, but I'll still try to give some thoughts/answers: As far as I know, not splitting package /too much/ is actually a choice. Indeed, Arch purposely avoid having too many split packages (hence the lack of `lib` or `utils` packages in Arch compare to Debian for instance) to keep things as simple as possible, according to the Arch philosophy. On a side note, from a purely technical stand point, there isn't a huge difference between installing the apache package and not starting the httpd web-server daemon on Arch and install `apache2-utils` on Debian? I mean, sure the latter only provides utilities and not the web-server itself while the first one provides both, but nothing obligates you to actually start and use the web-server, so I don't see why installing the `apache` package on Arch would be more conflicting regarding nginx than installing the `apache2-utils` package on Debian for people that wants to use `htpasswd`.
Thanks for reading this suggestion,
-- Regards, Robin Candau
Hello,
There's no issue having two (or more) webservers installed at once, and no issues having two (or more) of them running at the same time either, as long as they don't run on the same port. While, apache and nginx will indeed both use the 80 or 443 port by default, this is something you can change in the configuration file.
Ok I guess this is my bad for poorly explaining this, what my point was is that you wouldn't really run apache and nginx together, when you would pick one or the other as a reverse proxy for all your other applications (if you are limited to one IP address and want all your pages accessable on port 443/80 then it is one or the other), that is the point I was trying to make, you would typically pick one and not the other.
So for people that face the situation described in this mail: You can install the apache package alongside nginx (in order to use `htpasswd`) without issue. If you decide to have them both running at the same time, just make sure to bind the `listen` parameter in the default apache and nginx configurations files to a different port for each of them.
I know you can, I have wrote a guide doing exactly this, see https://wiki.archlinux.org/title/Maven#Using_nginx Its simple for them to coincide, you don't Start/enable apache and you can still use the utils with nginx, but the issue is, you don't want to have something installed you will never use!
As far as I know, not splitting package too much is actually a choice. Indeed, Arch purposely avoid having too many split packages (hence the lack of `lib` or `utils` packages in Arch compare to Debian for instance) to keep things as simple as possible, according to the Arch philosophy.
The issue with not splitting things up is that people have redundant binaries installed, which they are never going to use, the one of Arch's other philosophies is that the user is fully in control and can design their system how they like, this breaks that because they are forced to install stuff they will never want to use, just because a small section of the package contains a binary they needed for use with nginx.
I mean, sure the latter only provides utilities and not the web-server itself while the first one provides both, but nothing obligates you to actually start and use the web-server, so I don't see why installing the `apache` package on Arch would be more conflicting regarding nginx than installing the `apache2-utils` package on Debian for people that wants to use `htpasswd`.
Again, its not about them conflicting, its about that most people will not use the apache web server with nginx, and thus they only want to use the binaries which the nginx guides use (htpasswd), instead of installing a whole another web server, just to get nginx working? I hope you see the point I am trying to make, -- Polarian GPG signature: 0770E5312238C760 Website: https://polarian.dev JID/XMPP: polarian@polarian.dev
Polarian <polarian@polarian.dev> wrote:
Hello,
I would like to make a suggestion for the apache package, I was not too sure which mailing list to put this in, so I picked the general mailing list for the avoidance of doubt :P
So the apache package:
https://archlinux.org/packages/extra/x86_64/apache/
Currently contains tools which are used by nginx users, and the issue is that is that you can only run one webserver as once, they both are going to try to bind to port 80 and 443 so you got to choose.
htpaswd utility is used by nginx users too, and the issue is they have to install the entire apache web server in order to be able to use the single utility.
Just noting there seem to be many open source htpasswd replacements. Such as https://bash.cyberciti.biz/web-server/htpasswd-replacement-perl-script/ or https://github.com/backbone/htpasswd -- u34
I suggest splitting the package into apache-utils and apache, or, if the maintainer(s) are willing to maintain it, split the utilities into their own separate pacakge and use apache-utils as a group of all the subpackages (htpasswd, htdigest etc).
I feel this would be a good change so that nginx users do not need to install the entire apache web server just to use the apache utilities which nginx suggests to use within their documentation (see https://docs.nginx.com/nginx/admin-guide/security-controls/configuring-http-...)
As you can see in the guide, Debian, Ubuntu, RHEL, CentOS and Oracle Linux already have seen this issue within their own repositories and split the packages up, why can't arch linux?
Thanks for reading this suggestion, -- Polarian GPG signature: 0770E5312238C760 Website: https://polarian.dev JID/XMPP: polarian@polarian.dev
Hello, Thanks for suggestion, but that is besides the point, I am not looking for replacements, I am looking for restructuring of the apache package to allow htpasswd to be installed without the web server, so nginx users can follow the nginx guides. Thanks, -- Polarian GPG signature: 0770E5312238C760 Website: https://polarian.dev JID/XMPP: polarian@polarian.dev
Hi, Am 01.02.23 um 20:10 schrieb Polarian:
Hello,
I would like to make a suggestion for the apache package, I was not too sure which mailing list to put this in, so I picked the general mailing list for the avoidance of doubt :P
So the apache package:
https://archlinux.org/packages/extra/x86_64/apache/
Currently contains tools which are used by nginx users, and the issue is that is that you can only run one webserver as once, they both are going to try to bind to port 80 and 443 so you got to choose.
htpaswd utility is used by nginx users too, and the issue is they have to install the entire apache web server in order to be able to use the single utility.
Nginx is also compatible with passwords generated by openssl so there is no strict need to use htpasswd to generate Apache conformant passwords for Nginx. Use "openssl passwd -6" and put the generated hash into the password file that Nginx should use. Format of the file is the same as for Apache. Each line has username and hash separated by a colon (:). I don't know if this is documented anywhere but at least on my installations it just works. Regards, Uwe
I suggest splitting the package into apache-utils and apache, or, if the maintainer(s) are willing to maintain it, split the utilities into their own separate pacakge and use apache-utils as a group of all the subpackages (htpasswd, htdigest etc).
I feel this would be a good change so that nginx users do not need to install the entire apache web server just to use the apache utilities which nginx suggests to use within their documentation (see https://docs.nginx.com/nginx/admin-guide/security-controls/configuring-http-...)
As you can see in the guide, Debian, Ubuntu, RHEL, CentOS and Oracle Linux already have seen this issue within their own repositories and split the packages up, why can't arch linux?
Thanks for reading this suggestion,
On 2/1/23 15:14, Uwe Sauter wrote:
Use "openssl passwd -6" and put the generated hash into the password file that Nginx should use.
Interesting that -6 works - the dcos say use openssl passwd -apr1 [1] thank you! gene [1] http://nginx.org/en/docs/http/ngx_http_auth_basic_module.html
On Wed, Feb 1, 2023 at 9:04 PM Genes Lists <lists@sapience.com> wrote:
On 2/1/23 15:14, Uwe Sauter wrote:
Use "openssl passwd -6" and put the generated hash into the password
file that Nginx should use.
Interesting that -6 works - the dcos say use openssl passwd -apr1 [1]
thank you!
gene
[1] http://nginx.org/en/docs/http/ngx_http_auth_basic_module.html
arch-general@lists.archlinux.org -- mike c
Hey Mike, May I ask why you send at email to the mailing list, containing the email of the mailing list in the body, and nothing else but your signature? I am just a little confused thats all! Thanks, -- Polarian GPG signature: 0770E5312238C760 Website: https://polarian.dev JID/XMPP: polarian@polarian.dev
Hello, Maybe this is a solution we can add to the Arch Wiki, however its besides the point yet again, I am not looking for alternatives, nginx guides specifically use htpasswd, and the issue is a lot of people follow the nginx guides, unless you want to go through and rewrite all the docs nginx provides to be "Arch compliant", I do not see why the utility binaries cant be pulled and split into a second binary! Thanks, -- Polarian GPG signature: 0770E5312238C760 Website: https://polarian.dev JID/XMPP: polarian@polarian.dev
Hi Polarian,
Maybe this is a solution we can add to the Arch Wiki
What is? For the second time of asking, please quote pertinent parts of the email you are replying to as is normal netiquette as some of us have deleted earlier emails. -- Cheers, Ralph.
Hello Ralph, I thought it was obvious what I was responding to, because I was responding to an entire email, not just a section of the email. But sure I will quote more in the future. Thanks, -- Polarian GPG signature: 0770E5312238C760 Website: https://polarian.dev JID/XMPP: polarian@polarian.dev
On Wed, Feb 01, 2023 at 23:02:38 +0000, Polarian wrote:
Hello,
Maybe this is a solution we can add to the Arch Wiki, however its besides the point yet again, I am not looking for alternatives, nginx guides specifically use htpasswd, and the issue is a lot of people follow the nginx guides, unless you want to go through and rewrite all the docs nginx provides to be "Arch compliant", I do not see why the utility binaries cant be pulled and split into a second binary!
The same argument can be made for bind, as universal DNS tools like `host`, `dig` and `nslookup` are being packaged together with the BIND nameserver. It actually used to be split in "bind" vs "bind-tools", but was unified in https://github.com/archlinux/svntogit-packages/commit/c688d695dc4e82aad9a7ec... Geert
Hello,
The same argument can be made for bind, as universal DNS tools like `host`, `dig` and `nslookup` are being packaged together with the BIND nameserver.
Firstly, each one of these binaries are small enough that it doesn't really matter. Secondly, they all are use for a similar purpose, so it is not the biggest deal. How does this differ? Because apache is a huge http server, and for those who only want to use a script provided by the package, they have to give up additional storage. Nginx users need the use of the scripts in the apache package, and thus splitting it makes a lot more sense.
It actually used to be split in "bind" vs "bind-tools", but was unified in https://github.com/archlinux/svntogit-packages/commit/c688d695dc4e82aad9a7ec...
Why was it unified, staff didn't have enough time to maintain both packages? Thanks, -- Polarian GPG signature: 0770E5312238C760 Website: https://polarian.dev JID/XMPP: polarian@polarian.dev
On Thursday, 2 February 2023 at 08:55 (+0000), Polarian wrote:
Nginx users need the use of the scripts in the apache package, and thus splitting it makes a lot more sense.
I've used nginx as a webserver on several machines in many contexts for a long time, and I've never needed to use scripts in the apache package. Perhaps your use case is not as universal as you're making it out to be. Jaron
Hello,
I've used nginx as a webserver on several machines in many contexts for a long time, and I've never needed to use scripts in the apache package. Perhaps your use case is not as universal as you're making it out to be.
The ht scripts are for BASIC authentication, if you have not needed to do BASIC authentication you would not need these scripts. See: https://docs.nginx.com/nginx/admin-guide/security-controls/configuring-http-... The nginx documentation uses these scripts, so unless you want to rewrite all of the nginx documentation into the Arch Wiki to be "Arch compliant", users got to either install a secondary web server for a use of a small script, or they have to find their own way of doing things without using the nginx documentation! Thanks, -- Polarian GPG signature: 0770E5312238C760 Website: https://polarian.dev JID/XMPP: polarian@polarian.dev
On 2/2/23 04:02, Polarian wrote:
The nginx documentation uses these scripts,
Yes but for completeness, as referenced earlier nginx also says you can use htpasswd or openssl: [1] - encrypted with the crypt() function; can be generated using the “htpasswd” utility from the Apache HTTP Server distribution or the “openssl passwd” command; - hashed with the Apache variant of the MD5-based password algorithm (apr1); can be generated with the same tools; same tools := htpasswd or openssl [1] http://nginx.org/en/docs/http/ngx_http_auth_basic_module.html
On Wed, 2023-02-01 at 19:10 +0000, Polarian wrote:
Hello,
I would like to make a suggestion for the apache package, I was not too sure which mailing list to put this in, so I picked the general mailing list for the avoidance of doubt :P
It would be arch-dev-public. Cheers, Filipe Laíns
On Wed, 2023-02-01 at 21:12 +0000, Filipe Laíns wrote:
On Wed, 2023-02-01 at 19:10 +0000, Polarian wrote:
Hello,
I would like to make a suggestion for the apache package, I was not too sure which mailing list to put this in, so I picked the general mailing list for the avoidance of doubt :P
It would be arch-dev-public.
Cheers, Filipe Laíns
Sorry, I forgot posting to arch-dev-public is limited to the team only. arch- general seems like the best place in this case. Cheers, Filipe Laíns
Ah, Maybe this is another topic which should be discussed, allowing non-TU members to be able to contribute/make suggestions about the official repositories, provided that TUs back and support the changes/suggestions and verified that it is not malicious? Then again, I don't see whats stopping me from changing a package, and then emailing the TU a patch for the change explaining why I have made that specific change. The issue is... official repositories use SVN not git, and thus sending a patch probably wouldn't work in this case, unless it could be converted over to svn after?!?!?! I am just brainstorming ideas here, even though it is slightly off topic, if people feel this is a basis of a new thread to discuss this, let me know. Thanks, -- Polarian GPG signature: 0770E5312238C760 Website: https://polarian.dev JID/XMPP: polarian@polarian.dev
Hello, Thanks for clearing that up, however question is now, should I resubmit this issue seen as the thread is already began the conversation, or do I continue it here and note this for future reference? Thanks, -- Polarian GPG signature: 0770E5312238C760 Website: https://polarian.dev JID/XMPP: polarian@polarian.dev
Polarian,
I suggest splitting the package into apache-utils and apache, or, if the maintainer(s) are willing to maintain it, split the utilities into their own separate pacakge and use apache-utils as a group of all the subpackages (htpasswd, htdigest etc).
only because i haven't seen anybody mention this: i think i have heard that reasons for not splitting packages that don't come "split" from upstream includes the desire to reduce work for Arch package maintainers, and to reduce possible errors (experienced by users) from doing the splitting. cheers, Greg
Hello,
only because i haven't seen anybody mention this: i think i have heard that reasons for not splitting packages that don't come "split" from upstream includes the desire to reduce work for Arch package maintainers, and to reduce possible errors (experienced by users) from doing the splitting.
This is not a large change however, htpasswd seems to be a standalone tool, thus the apache package could be turned into a multi-package PKGBUILD, splitting of the "support" executables into a secondary package... Thanks, -- Polarian GPG signature: 0770E5312238C760 Website: https://polarian.dev JID/XMPP: polarian@polarian.dev
Hi,
This is not a large change however, htpasswd seems to be a standalone tool, thus the apache package could be turned into a multi-package PKGBUILD, splitting of the "support" executables into a secondary package...
Even "host" and "dig" are standalone tools and you have to install bind9 to get them. If you want split packages, Arch seems not to be the right distro for you. Please no further split packages, alternative init systems etc. discussions again. Regards Bjoern
Hello,
Even "host" and "dig" are standalone tools and you have to install bind9 to get them. If you want split packages, Arch seems not to be the right distro for you. Please no further split packages, alternative init systems etc. discussions again.
That is different however, they are all tools which are relevant to the bind9 package, htpasswd and other utilities are relevant to more packages than just apache, thus should be extracted, so that other packages can properly depend on them. Thanks, -- Polarian GPG signature: 0770E5312238C760 Website: https://polarian.dev JID/XMPP: polarian@polarian.dev
Hi,
That is different however, they are all tools which are relevant to the bind9 package, htpasswd and other utilities are relevant to more packages than just apache, thus should be extracted, so that other packages can properly depend on them.
dig and host can be relevant to any DNS debugging, and as you may know, there are more nameservers out there than bind.
Hello,
dig and host can be relevant to any DNS debugging, and as you may know, there are more nameservers out there than bind.
The package is small enough that this would not matter too much, in the case of apache, you have to install the entire http server in order to use a specific script. Thanks, -- Polarian GPG signature: 0770E5312238C760 Website: https://polarian.dev JID/XMPP: polarian@polarian.dev
The package is small enough that this would not matter too much, in the case of apache, you have to install the entire http server in order to use a specific script.
Where's the difference between "entire http server" and "entire DNS server"?
Hello,
Where's the difference between "entire http server" and "entire DNS server"?
Good point... hmm It seems Arch has conflicting issues, if they split the packages they go against their simplicity, but if they keep the packages together, they go against the configurability of the distribution. It seems like Arch Staff prefer to lean towards the former, because it saves them more time, than the latter which would mean they would have a harder time maintaining the packages. The issue is the package is within the extra repository, not the community, and thus it is maintained by Arch Admins and not TUs, thus it does make sense why they combined them. The only suggestion I can make then if they do not want to maintain it is to provide apache-utils within the AUR which only packages the utilities? I can add it to my arch repository to make it easier to install too, it could be a viable alternative. Thanks, -- Polarian GPG signature: 0770E5312238C760 Website: https://polarian.dev JID/XMPP: polarian@polarian.dev
On Thu, 2023-02-02 at 09:08 +0000, Polarian wrote:
It seems Arch has conflicting issues, if they split the packages they go against their simplicity, but if they keep the packages together, they go against the configurability of the distribution.
Hi, I didn't read the complete thread, hence I might miss something. What is wrong with using NoExtract of /etc/pacman.conf , if you don't want to install the complete content of a package? Regards, Ralf
I didn't read the complete thread, hence I might miss something. What is wrong with using NoExtract of /etc/pacman.conf , if you don't want to install the complete content of a package?
Ah, I forgot about that configuration option, that is a possibility, but that is not "simple" like the Arch philosophy. Personally the best solution I have thought of is implement this into the AUR, see what the userbase thinks, if it is ignored and not used, then obviously people do not mind. If I get time I will do this for both bind/bind-utils and also apache/apache-utils. The issue is I am skirting on the edge of being removed for that, as it technically is packaging existing packages, but in a slightly different format, so it is subjective to the TU reviewing the packages... Its still worth a go don't you think? Thanks, -- Polarian GPG signature: 0770E5312238C760 Website: https://polarian.dev JID/XMPP: polarian@polarian.dev
Hi Polarian, your mails are single-handedly flooding my list mbox - please be mindful of others when interacting with mailing lists, perhaps collate your emails into fewer. On 02.02.23 10:24, Polarian wrote:
The issue is I am skirting on the edge of being removed for that, as it technically is packaging existing packages
Its still worth a go don't you think?
don't knowingly and intentionally break the rules, over silly issues at that. -coderobe
Hello,
your mails are single-handedly flooding my list mbox - please be mindful of others when interacting with mailing lists, perhaps collate your emails into fewer.
Sorry, however I do not know what I meant to do about this, I compress as many comments into a single email as possible, it is not like I send tons of emails, unfortunately I can not anticipate what people are going to reply with until they do, and then I reply back. I am aware I am very active on the mailing list, you are far from the first person to bring this up, but I am not too sure what I meant to do. I am not spamming or abusing the mailing list, I am simply asking questions, and getting support, I have been recommended to join the IRC channels so I might do that instead, but I drastically prefer communication over email and that is why I am so active on the mailing list.
don't knowingly and intentionally break the rules, over silly issues at that.
I think you misunderstood, it is not intentionally breaking the rules, because it is a region which is subjective, yes it is repackaging already packaged software, but its packaged in a different format, thus it should not be breaking the rules, but it is subjective to whatever TU who reviews the package at that moment in time. I assume from your point of view this breaks the "do not have duplicate packages" guideline, but when I ask myself "is the package different enough to warrant a new package?" my brain says yes. By the way I do not know all TUs by name, and if you are not using your @archlinux.org email address I might miss the fact you are a TU, if I have accidentally questioned a TUs authority I apologise, TUs have went through an application process and know what they are talking about, I am not saying non-TUs don't, just they haven't been verified :) (I feel like I am digging a hole for myself which this awful articulation, I hope you get the point I am trying to make, I question to ensure I have the right information, not to be an a**hole) Again sorry for flooding your inbox, not sure how I can fix this issue though, I might try moving my smaller questions over to IRC, but that is about it... I do not know what else to do... Thanks, -- Polarian GPG signature: 0770E5312238C760 Website: https://polarian.dev JID/XMPP: polarian@polarian.dev
Hi Polarian,
your mails are single-handedly flooding my list mbox - please be mindful of others when interacting with mailing lists, perhaps collate your emails into fewer.
Sorry, however I do not know what I meant to do about this
I'll try to help.
I compress as many comments into a single email as possible, it is not like I send tons of emails
Relatively, you do.
unfortunately I can not anticipate what people are going to reply with until they do, and then I reply back.
Yes, you reply back. That's part of the problem. You've told us you have autism. I know next to nothing about it but presume it's influencing you behaviour here. What I see: - You email with a suggestion for improvement. - Someone tells you why not. - You, very promptly, reply with why you disagree with them. So far, so good. - Someone else gives you their opinion. They may also disagree triggering... - Another, very prompt, reply from you tacking their arguments, which may not duplicate the first replier's but could overlap with them. By the time you've had a few people try to tell you why they dislike the idea, you've replied quite a few times and your points are becoming repetitive because their arguments are overlapping quite a bit by now. If I then read the list after twelve hours away, I see a big thread with you as the main voice and not a lot of progress being made. How to fix it: - Don't reply quickly. Don't even write the reply but not send it. - Wait until more replies may have come in. I'd suggest at least a couple of hours. - Do they collectively alter your opinion? - If so, pick one which made the most impact on you and tell us your new stance. - Or pick the one which seems to most intelligently argue against your case and reply. - By picking the hardest one to rebut, you will think more about your case and the list's many readers will get higher signal. - If a second one makes a completely different argument against your case then reply to it but only if you think their case has merit. - It could be the majority of the list's readers can quickly see for themselves that the reply to you isn't worth much and will not want to see the two of you battle it back and forth. - Be content not to reply to everyone who replied to you. - Quote just the pertinent bits of their email and reply underneath each. - This gives context to readers, avoids you having to frame what you say with extra words, stops you accidentally misrepresenting them, and keeps their thoughts in your focus as you write rather than it being out of sight. - Part-way through, ask ‘Is this worth it?’. It's taking up your time. You could be doing something else; there's an opportunity cost. It will take the time of the list's readers to process it. I abandon quite a few emails once that initial ‘There's something wrong on the Internet’ has ebbed. - When you've finished, re-read your email looking for text to clarify and, better still, text to cut. Burn your time to save your readers' time by making your case clearer and more succinct. - Much back and forth can be triggered on both sides by a lack of clarity: ‘Be precise in your speech’, as someone wrote. -- Cheers, Ralph.
On Sat, 4 Feb 2023 at 09:24, Ralph Corderoy <ralph@inputplus.co.uk> wrote:
Hi Polarian,
your mails are single-handedly flooding my list mbox - please be mindful of others when interacting with mailing lists, perhaps collate your emails into fewer.
Sorry, however I do not know what I meant to do about this
I'll try to help.
I compress as many comments into a single email as possible, it is not like I send tons of emails
Relatively, you do.
unfortunately I can not anticipate what people are going to reply with until they do, and then I reply back.
Yes, you reply back. That's part of the problem.
You've told us you have autism. I know next to nothing about it but presume it's influencing you behaviour here. What I see:
- You email with a suggestion for improvement. - Someone tells you why not. - You, very promptly, reply with why you disagree with them.
So far, so good.
- Someone else gives you their opinion. They may also disagree triggering... - Another, very prompt, reply from you tacking their arguments, which may not duplicate the first replier's but could overlap with them.
By the time you've had a few people try to tell you why they dislike the idea, you've replied quite a few times and your points are becoming repetitive because their arguments are overlapping quite a bit by now.
If I then read the list after twelve hours away, I see a big thread with you as the main voice and not a lot of progress being made.
How to fix it:
- Don't reply quickly. Don't even write the reply but not send it. - Wait until more replies may have come in. I'd suggest at least a couple of hours.
- Do they collectively alter your opinion? - If so, pick one which made the most impact on you and tell us your new stance.
- Or pick the one which seems to most intelligently argue against your case and reply. - By picking the hardest one to rebut, you will think more about your case and the list's many readers will get higher signal.
- If a second one makes a completely different argument against your case then reply to it but only if you think their case has merit. - It could be the majority of the list's readers can quickly see for themselves that the reply to you isn't worth much and will not want to see the two of you battle it back and forth.
- Be content not to reply to everyone who replied to you. - Quote just the pertinent bits of their email and reply underneath each. - This gives context to readers, avoids you having to frame what you say with extra words, stops you accidentally misrepresenting them, and keeps their thoughts in your focus as you write rather than it being out of sight.
- Part-way through, ask ‘Is this worth it?’. It's taking up your time. You could be doing something else; there's an opportunity cost. It will take the time of the list's readers to process it. I abandon quite a few emails once that initial ‘There's something wrong on the Internet’ has ebbed.
- When you've finished, re-read your email looking for text to clarify and, better still, text to cut. Burn your time to save your readers' time by making your case clearer and more succinct. - Much back and forth can be triggered on both sides by a lack of clarity: ‘Be precise in your speech’, as someone wrote.
-- Cheers, Ralph.
Do we really want to start micromanaging the replies here? In all seriousness, may I suggest people are getting upset about non-issues? They are just text-based emails, a few hundred bytes each time. If you really feel overwhelmed, you can filter out by originator. I do not think that we need to single out the person with autism and tell them to walk on eggshells, do we?
Ok, I have enough of people slandering me, and then other people stepping in to defend. This mailing list is not to slate people, neurodivergent or not, about how they speak, reply, etc. It is for Arch related issues, and right now you are the noise within the mailing list. Yes, I am active within the Arch community, you will see me a lot on the Arch Wiki, and sometimes on the AUR, I do not post aimlessly. If you have an issue with me, bring it up with me offlist, or even better report me to an administrator, because quite frankly I do not think it is going to go very far either way. You are complaining for the sake of complaining, this is a mailing list, I get hundreds of emails a day from mailing lists and I do not go emailing the people causing it telling them to shut up... Don't like active mailing lists, unsubscribe, but I am fed up of the constant complaints about me within this thread! I hope I have made myself clear! -- Polarian GPG signature: 0770E5312238C760 Website: https://polarian.dev JID/XMPP: polarian@polarian.dev
Hi Polarian,
It is for Arch related issues, and right now you are the noise within the mailing list.
Who is? You didn't quote anyone. Assuming it's me, you kept asking the list what you were doing wrong because you were puzzled why others complained. I tried to help, because you apparently have a distinct lack of self awareness. It was not intended to be offensive and I'm sorry it could be taken that way.
I do not post aimlessly.
Straw man.
If you have an issue with me, bring it up with me offlist
You asked on-list, repeatedly.
You are complaining for the sake of complaining
You are wrong. I took too much of my precious time trying to help.
I hope I have made myself clear!
Perfectly. I will not reply again. -- Cheers, Ralph.
Raph, it was not directed at you, it was directed as you, as far as I am aware you have only posted constructive criticism. I did not direct it at anyone for a reason, because I do not want to maliciously insult someone, my comment was directed at everyone in the mailing list who are intentionally, and maliciously insulting. Thanks, -- Polarian GPG signature: 0770E5312238C760 Website: https://polarian.dev JID/XMPP: polarian@polarian.dev
Hi Polarian,
Raph, it was not directed at you, it was directed as you, as far as I am aware you have only posted constructive criticism.
I assumed wrong, sorry. Thanks for putting me right. -- Cheers, Ralph.
On 4. Feb 2023, at 14:54, Andy Pieters <arch-general@andypieters.me.uk> wrote:
I do not think that we need to single out the person with autism and tell them to walk on eggshells, do we?
Strong agreement. I think, there was much more noise because of people complaining about noise than because of maybe one or two emails that might have been compressed into one :)
Hi, FWIW In fact, Polarian has sent an average of 2.3 (small sized) emails/day to this mailing list for the last 30 days [1]. It's apparently a (denial of service) mailbombing attack. Hard to believe the server wasn't overloaded. ;) Regards, Ralf [1] "The following statistics are from the past 30 days[...] Polarian 68 posts" - https://lists.archlinux.org/archives/list/arch-general@lists.archlinux.org/
Hello, Good to know I am a mailbomb, maybe it is true, I am a robot :P Email servers can handle massive numbers of throughput, therefore no... it wouldn't be overwhelmed. -- Polarian GPG signature: 0770E5312238C760 Website: https://polarian.dev JID/XMPP: polarian@polarian.dev
your mails are single-handedly flooding my list mbox - please be mindful of others when interacting with mailing lists, perhaps collate your emails into fewer.
Man, you could criticize Polarian for some things, but complaining that you get a lot of mail on a mailing list is the most absurd thing I've read in a long time. Without any acrimony, if you don't like that people expose their ideas and answer to what they receive, unsubscribe because you haven't understood what this tool is for. It is in fact these comments that make people afraid of the community, to ask and discuss. And then we complain that there is no "new blood" or that there is no "generational replacement". Greetings.
On 03.02.23 10:11, ogarcia@moire.org wrote:
Man, you could criticize Polarian for some things, but complaining that you get a lot of mail on a mailing list is the most absurd thing I've read in a long time.
Without any acrimony, if you don't like that people expose their ideas and answer to what they receive, unsubscribe because you haven't understood what this tool is for.
It is in fact these comments that make people afraid of the community, to ask and discuss. And then we complain that there is no "new blood" or that there is no "generational replacement".
Greetings. Thanks a lot for stating this, I couldn't agree more. Its a good thing if people state their ideas, even if they diverge from what is commonly standard and therefore take a couple of more emails to explain.
Hello,
Thanks a lot for stating this, I couldn't agree more. Its a good thing if people state their ideas, even if they diverge from what is commonly standard and therefore take a couple of more emails to explain.
Add a few more emails on for bad articulation in the first place, and having to reiterate my point at least 3-4 times before people actually understand :P Thanks, -- Polarian GPG signature: 0770E5312238C760 Website: https://polarian.dev JID/XMPP: polarian@polarian.dev
Hello,
Man, you could criticize Polarian for some things, but complaining that you get a lot of mail on a mailing list is the most absurd thing I've read in a long time.
Wow, did not expect someone to defend me xD
Without any acrimony, if you don't like that people expose their ideas and answer to what they receive, unsubscribe because you haven't understood what this tool is for.
You do make a good point, I am not the only one posting to this mailing list, there are topics I get spammed about and not interested about, you just gotta mute the thread, simple!
It is in fact these comments that make people afraid of the community, to ask and discuss. And then we complain that there is no "new blood" or that there is no "generational replacement".
I can not tell if this is an insult towards me (being a "newbie") or whether this meant something else... Thanks, -- Polarian GPG signature: 0770E5312238C760 Website: https://polarian.dev JID/XMPP: polarian@polarian.dev
I can not tell if this is an insult towards me (being a "newbie") or whether this meant something else...
No, not at all, I simply see my reflection in you. Many times those of us who have been in this "business" for many years complain that there is no generational replacement, that people are not interested in following in our footsteps and that the day we leave it will become a wasteland, but then when someone arrives they are scared away because they do not agree with our ideas or our way of acting, and that is a hypocritical attitude. There is nothing wrong with being a newbie either, we've all gone through that point (I'm still going through a lot of things myself). And if, being a newbie, you live in fear of "messing up" you will neither integrate nor learn because you will throw in the towel before you start. This is a mailing list to discuss our favorite distribution. No one should be afraid to do so freely, regardless of how crazy their ideas may sound initially. As long as the rules are respected, feel free. And if someone doesn't like this it's very easy, mute the thread or simply unsubscribe. And well, I'm not saying anything else because I'm going completely off topic and that's not respecting the rules ;-) Greetings
On Fri, 2023-02-03 at 09:11 +0000, ogarcia@moire.org wrote:
if you don't like that people expose their ideas and answer to what they receive, unsubscribe
Hi, the suggestion to leave the mailing list is just as wrong as whining about too many probably unnecessary e-mails. It's very easy to filter emails, both server-side and with an email client. Anyway, the thread leads to nothing. If I were the OP I would either build my own local package or use NoExtract of /etc/pacman.conf instead of wasting time fighting windmills. Regards, Ralf
Hi ogarcia,
your mails are single-handedly flooding my list mbox - please be mindful of others when interacting with mailing lists, perhaps collate your emails into fewer.
you could criticize Polarian for some things, but complaining that you get a lot of mail on a mailing list is the most absurd thing I've read in a long time.
That wasn't his complaint. His complaint was valid. ‘attacking a straw man’ — https://en.wikipedia.org/wiki/Straw_man -- Cheers, Ralph.
On Thu, Feb 02, 2023 at 09:08:58AM +0000, Polarian wrote:
Hello,
Where's the difference between "entire http server" and "entire DNS server"?
Good point... hmm
It seems Arch has conflicting issues, if they split the packages they go against their simplicity, but if they keep the packages together, they go against the configurability of the distribution.
It seems like Arch Staff prefer to lean towards the former, because it saves them more time, than the latter which would mean they would have a harder time maintaining the packages.
You seem to misunderstanding the priority and intention of the principles in the first place:
Arch Linux defines simplicity as without unnecessary additions or modifications.
Splitting the upstream package is arguably a mostly unnecessary modification. Furthermore, user-centricity is often misinterpreted as being focussed on the experience of the end user. This is explicitly not the case:
The distribution is intended to fill the needs of those contributing to it, rather than trying to appeal to as many users as possible.
(Both inline quotes taken from https://wiki.archlinux.org/title/Arch_Linux ) So a whole lot of noise for saving at most 6MB on disk, seems pretty unnecessary. Kind regards, Khorne
On Thursday, 2 February 2023 at 09:04 (+0000), Polarian wrote:
dig and host can be relevant to any DNS debugging, and as you may know, there are more nameservers out there than bind.
The package is small enough that this would not matter too much, in the case of apache, you have to install the entire http server in order to use a specific script.
I install bind for the use of dig. pacman -Si bind
Download Size : 1973.12 KiB Installed Size : 11734.10 KiB
pacman -Si apache
Download Size : 1726.92 KiB Installed Size : 6627.10 KiB
Apache actually has a smaller footprint. Jaron
Hello,
I install bind for the use of dig.
pacman -Si bind
Download Size : 1973.12 KiB Installed Size : 11734.10 KiB
pacman -Si apache
Download Size : 1726.92 KiB Installed Size : 6627.10 KiB
Apache actually has a smaller footprint.
Well it seems this discussion is not going to get anywhere, if they joined the bind package and bind-utils together, then they are not going to split apache into apache and apache-utils. As I have suggested in my previous email, as the packages are not within the community repository, TUs will not be able to maintain them, and thus the manpower is a lot lower, thus the only choice is to implement this in the AUR. Thanks, -- Polarian GPG signature: 0770E5312238C760 Website: https://polarian.dev JID/XMPP: polarian@polarian.dev
Hi,
Well it seems this discussion is not going to get anywhere, if they joined the bind package and bind-utils together, then they are not going to split apache into apache and apache-utils.
As I have suggested in my previous email, as the packages are not within the community repository, TUs will not be able to maintain them, and thus the manpower is a lot lower, thus the only choice is to implement this in the AUR.
I really don't get the point why this is a problem. I assume Arch isn't installed on 80MB IDE disks where 5MB more usage matters. More and more I get the impression that you construct a problem which you have, but you think it's a problem for the whole userbase.
I install bind for the use of dig.
You can use `drill`[1] instead of `dig` and avoid installing anything (as long as you have OpenSSH installed, which, dare I say it, is a must). It may seem a little strange at first, but it's essentially the same thing. [1]: https://linux.die.net/man/1/drill
On Friday, 3 February 2023 at 09:03 (+0000), ogarcia@moire.org wrote:
I install bind for the use of dig.
You can use `drill`[1] instead of `dig` and avoid installing anything (as long as you have OpenSSH installed, which, dare I say it, is a must). It may seem a little strange at first, but it's essentially the same thing.
Thanks, drill seems like a nice replacement. Though, these days I find myself using `resolvectl query` for most simple tests. I don't suppose you have a similar replacement for my other dependency on the bind package, dnssec-signzone? Jaron
Thanks, drill seems like a nice replacement. Though, these days I find myself using `resolvectl query` for most simple tests.
I don't suppose you have a similar replacement for my other dependency on the bind package, dnssec-signzone?
As a command I personally do not know of any replacement. But if you simply use `dnssec-signzone` to check the DNSSEC status of a domain you can pull online tools like https://dnssec-analyzer.verisignlabs.com/ (not that it is the best, but it is something).
On Friday, 3 February 2023 at 10:03 (+0000), ogarcia@moire.org wrote:
I don't suppose you have a similar replacement for my other dependency on the bind package, dnssec-signzone?
As a command I personally do not know of any replacement. But if you simply use `dnssec-signzone` to check the DNSSEC status of a domain you can pull online tools like https://dnssec-analyzer.verisignlabs.com/ (not that it is the best, but it is something).
No, I need to actually locally sign zonefiles for my own DNSSEC. Then even `resolvectl query` (part of base!) can check their DNSSEC status. Jaron
No, I need to actually locally sign zonefiles for my own DNSSEC. Then even `resolvectl query` (part of base!) can check their DNSSEC status.
No, I don't know anything. It is not an operation that I do either, so I cannot contribute much in this case :-(
On 2/3/23 05:03, ogarcia@moire.org wrote:
I don't suppose you have a similar replacement for my other dependency on the bind package, dnssec-signzone?
As a command I personally do not know of any replacement. But if you simply use `dnssec-signzone` to check the DNSSEC status of a domain you can pull online tools like https://dnssec-analyzer.verisignlabs.com/ (not that it is the best, but it is something).
I use dnssec to sign domains and I am using ldns to do the leg work (key creating, zone signing). ldns may have what you want. I even wrote a little python tool to help manage deployment - it even automates key rollovers - its in AUR and github [1] gene [1] https://github.com/gene-git/dns_tools https://aur.archlinux.org/packages/dns_tools
Hello, I would like to highlight the fact that this thread has split into two conversations, one over apache and the other over DNS... I assume this has went completely off topic... surely this should be moved to a new thread which is new relevant per mailing list guidelines. Thanks, -- Polarian GPG signature: 0770E5312238C760 Website: https://polarian.dev JID/XMPP: polarian@polarian.dev
On Friday, 3 February 2023 at 06:34 (-0500), Genes Lists wrote:
On 2/3/23 05:03, ogarcia@moire.org wrote:
I don't suppose you have a similar replacement for my other dependency on the bind package, dnssec-signzone?
As a command I personally do not know of any replacement. But if you simply use `dnssec-signzone` to check the DNSSEC status of a domain you can pull online tools like https://dnssec-analyzer.verisignlabs.com/ (not that it is the best, but it is something).
I use dnssec to sign domains and I am using ldns to do the leg work (key creating, zone signing). ldns may have what you want.
Thanks! I'll check out ldns and roll it's signing utility into my own hacky script for automatically resigning and deploying DNSSEC/TLSA ;) Then pacman -Rs bind... Jaron
participants (19)
-
Andy Pieters
-
Bjoern Franke
-
Filipe Laíns
-
Geert Hendrickx
-
Genes Lists
-
Greg Minshall
-
Jaron Kent-Dobias
-
Khorne
-
LuKaRo
-
Lukas Rose
-
Mara Broda
-
Mike Cloaked
-
ogarcia@moire.org
-
Polarian
-
Ralf Mardorf
-
Ralph Corderoy
-
Robin Candau
-
u34@net9.ga
-
Uwe Sauter