New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
other: open TODOs #3289
Comments
Rationale for closing stale PRs: It may be a personal thing, but I like the PR list to be clean and up-to-date with non-stale PRs. I hope this does not offend anyone, but maintaining DMS for me is easier when I see all relevant PRs as open (or non-stale WIP). Maintaining the PR and issue list is a lot of work, especially for me, so I'd like to keep it in a way that is comfortable for me. This does not mean your PR contains bad changes / additions! It's just that it has stalled. If you find the time, apply your changes to the most recent version of I spent a lot of time on DMS, and I mean a lot. And I think DMS resonates well with users. But what many people don't see is the amount of work that goes into maintenance. After having maintained DMS for a long time, and keeping track and cleaning the issue tracker, I can confidently say that I feel like this is the right think, even if not everyone agrees. |
Additional TODO to considerMisc
Docs TODO
Example - Local Mail RelayThe example section itself is a candidate for removal, but probably still a worthwhile use-case to document somewhere? (common enough use-case that users run into with various workarounds / misconfigurations) If not removed entirely, these two parts may no longer be relevant:
Alias to redirect mail to external MTA like gmail is fine instead of POP3 AFAIK, but conflicts with using DMS to send mail from an account of the same alias address.
Related (but may not be a v13 goal), is the FAQ Entries
Update for Compose V2 (V1 format deprecated from July 2023):
That would align with the changes in upstream Docker from June 2023. As referenced from their current docs below. This Docker compose docs page has the current notification present:
Final link mentions:
|
* removed FAQ entry about Rancher, see <#3289 (comment)> * update FAQ about special directories, see <#3289 (comment)>
@polarathene could you give me a very rough (short) estimation of what it would take (also time wise) to update our CI so that we can take advantage of the Docker build cache for the ARM64 image as well? |
Sorry wasn't exactly short 😅 Time wise, I am not taking on any new tasks right now (presently on IPv6 docs task with I haven't refreshed on GHA cache driver, so if that looks easier for you, maybe consider that. @georglauterbach it's been a while since I thought about it. I recall two options:
While only one build cache can be exported, multiple can be imported. I think if we are sure we build the image for
You might want to subscribe to this issue to be notified of ARM64 runner support if it ever arrives. But hopefully the caching in CI will be good enough 👍 For current cache approach, I think you could use (and using a platform subdir for the
docker-mailserver/.github/workflows/generic_publish.yml Lines 51 to 52 in bba72da
Then:
cache-from: |
type=local,src=/tmp/.buildx-cache/amd64
type=local,src=/tmp/.buildx-cache/arm64 Just note that other workflows are using the derived key from the build workflow to pull in a cache too. You need to make sure you're properly handling the The build workflow should be fine if it's not doing mixed platform build, then it only needs to update the cache key and |
Thank you very much! ❤️ I will, after much of the above work is done, consider the GHA build cache to finally bring down overall PR test time. |
tl;dr; Providing alternatives 👍
Just for the record. We should do this IMO very conservative. I am fine with providing alternatives like fetchmail --> getmail. My reasoning is, at the moment, there are a lot of setups (mine included), that just works. For example, I've been using fetchmail for some time and didn't run in any issue ever. So from a rational point, there is absolutely no reason, why I should switch to getmail (as long as fetchmail is officially supported and receives security updates if required). Switching could only introduce new issues, that haven't been any before. Never change a running system 😎 |
Agreed 👍🏼 the whole thing was more about switching. Before we actually remove functionality, there are going to be a lot of releases in between I suspect. I wouldn't change a running system either. |
See #3289 These have little effect at the moment, and will only be useful in conjunction with `set -eEu`, but I want to have them in v13.0.0 as this was part of the initial todo issue.
@georglauterbach did you want this issue closed? Or should it be opened and drop the version specific target? |
Good catch. |
@polarathene before we integrate new features like Vector, I'd like to see this issue being cleared. I'd provide the PR for making Rspamd the default in v14. I'd be very nice of we can also take care of the issues that are not yet completed mentioned in your comments :) |
Yes absolutely I agree 👍 I only raised that issue as it was something I had been meaning to better document as a TODO that I could delegate. I don't expect it to be a priority any time soon.
I'll see what I can do, current priorities are:
If there's any other pressing TODO items beyond that, let me know. I'd like to tackle the The LDAP rework will close several long-standing issues, along with the |
@zdenbe I don't know what you're doing with this DMS fork, but you recently contributed a PR there with no context that seems to be commits since That created a reference to many issues / PRs due to your merge commit message using our merge commits and their messages. Please don't do that. |
Perfect! Let me know if I can assist you somewhere (preferably not LDAP 😆) |
If you've got time and any of these interest you, they'd probably be good to tackle:
I didn't get around to checking, but for The sieve and amavis issues are probably a good fit for you. The |
@zdenbe I don't know what you're doing with this DMS fork, but you recently contributed a PR there with no context that seems to be commits since 10.5 to 12.1: evymo#1
That created a reference to many issues / PRs due to your merge commit message using our merge commits and their messages. Please don't do that.
I am sorry. I don't know how it happened… My fault. Nothing to contribute to origin. I tried just update fork…
|
Subject
I would like to contribute to the project
Description
I'd like v13.0.0 to be a release that barely adds new functionality, because I think DMS is already packed with functionality. For the far future, I'd like to remove unnecessary features as much as possible (i.e. OpenDKIM, OpenDMARC, Amavis (plus codecs), Spamassassin, Postgrey, Razor, Pyzor, fts-xapian and the likes). With v13.0.0, the default will change to Rspamd. Later, with upcoming major releases after v13.0.0, we can remove functionality.
The script are already high-class, but they lack Bash's integrity and error checks - this should changes as well. Then there are many TODO tickets in the issue tracker that should be resolved.
1
to0
for OpenDKIM, OpenDMARC, policyd-spf, Amavis & SpamAssassinintroduce(EDIT: could be done, but I lack the time)set -u
andset -eE
to our scripts to make them more robustset -o pipefail
andshopt -s inherit_errexit
to our scripts to make them more robustdeprecateOVERRIDE_HOSTNAME
and introduceDMS_FQDN
The text was updated successfully, but these errors were encountered: