Forgejo monthly update - February 2025 #564

Close #556

Preview: https://forgejo.codeberg.page/@pull_564/news & https://forgejo.codeberg.page/@pull_564/2025-02-monthly-update/

• draft the toot & the room announcement
• verify there is no TBD left
• verify all reviews were resolved
• publish
• toot (pin & unpin the previous one): https://floss.social/@forgejo/114190014837558571
• room announcement: https://matrix.to/#/!qjPHwFPdxhpLkXMkyP:matrix.org/$wusXgTb8SLDIjiJ_cn5h76ktaRJSKK4b1JLzLPzg-rk?via=matrix.org&via=tchncs.de&via=envs.net

Room announcement

@room the [monthly update was published](https://forgejo.org/2025-02-monthly-update/) ✨ It is a high level overview of the project activities.
FOSDEM impressions were published and feedback is being collected regarding migrations to Forgejo and commercial usages of Forgejo. code.forgejo.org got hit by a DDoS that was successfully mitigated. A new localization component is now in use and work on federation continued.

Toot

The #Forgejo monthly update was published ✨ It is a high level overview of the project activities.
FOSDEM impressions were published and feedback is being collected regarding migrations to Forgejo and commercial usages of Forgejo. code.forgejo.org got hit by a DDoS that was successfully mitigated. A new localization component is now in use and work on federation continued.
https://forgejo.org/2025-02-monthly-update/

It is the first time when I user forgejo-curl.sh so I'm not sure if and how I could have done this without logging in twice.

since='2025年02月01日T00%3A00%3A00Z'
before='2025年03月01日T00%3A00%3A00Z'
(
	forgejo-curl.sh logout
	forgejo-curl.sh --token {codeberg_org_token} login https://codeberg.org
	
 for repo in codeberg.org/api/v1/repos/forgejo/{design,forgejo,website,docs,tools,governance,sustainability,discussions} ; do
 forgejo-curl.sh api_json "https://${repo}/issues/comments?since=${since}&before=${before}" | jq -r '.[] | .user.login'
 forgejo-curl.sh api_json "https://${repo}/issues?limit=50&since=${since}&before=${before}" | jq -r '.[] | .user.login'
 done
 
 forgejo-curl.sh logout
 forgejo-curl.sh --token {code_forgejo_org_token} login https://code.forgejo.org
 
 for repo in code.forgejo.org/api/v1/repos/forgejo/{runner,end-to-end,act,lxc-helpers,forgejo-curl} code.forgejo.org/api/v1/repos/forgejo-helm/forgejo-helm code.forgejo.org/api/v1/repos/actions/{cascading-pr,setup-forgejo,forgejo-release,ovh-dns-update} ; do
 forgejo-curl.sh api_json "https://${repo}/issues/comments?since=${since}&before=${before}" | jq -r '.[] | .user.login'
 forgejo-curl.sh api_json "https://${repo}/issues?limit=50&since=${since}&before=${before}" | jq -r '.[] | .user.login'
 done
 
 forgejo-curl.sh logout
) | sort -u | grep -v -e '^translate$' -e '^Ghost$' -e '^forgejo-' -e '^thefox$' -e '^end-to-end$' -e '^cascading-pr$' -e '^viceice-bot$'| sed -e 's|^|- https://codeberg.org/|'

Besides putting together the snippets snippets from #556 I've added 'Releases' and 'Federation' sections

At this point the script may need the before parameter set.

It will for sure need the before parameter (I admit that just before creating the PR I realized that I should also update the script and didn't even tried to run it).

Preview ready: https://forgejo.codeberg.page/@pull_564/
https://forgejo.codeberg.page/@pull_564/2025-02-monthly-update/
https://forgejo.codeberg.page/@pull_564/tag/report/6/
https://forgejo.codeberg.page/@pull_564/tag/news/7/

I wanted to prepare the list of contributors for February but I have some questions regarding the script.

First of all, why is the time from since timestamp added as 'T00:00:02+00:00'?
I mean, is there a reason for those 2 seconds?

Secondly, concerning the issues list, it seems that it is paginated and by default it uses a page size of 30; so I'm afraid that not all the issues are returned if there was a lot of activity during a given month.

helm chart had 4 bugfix releases, two for v11 and two for v7

I added a sentence about Forgejo Helm chart releases within the 'Releases' section.

Regarding the script for generating the list of contributors, I would do it like this:

since='2025年02月01日T00%3A00%3A00%2B00%3A00'
before='2025年03月01日T00%3A00%3A00%2B00%3A00'
(
 for repo in codeberg.org/api/v1/repos/forgejo/{design,forgejo,website,docs,tools,governance,sustainability,discussions} code.forgejo.org/api/v1/repos/forgejo/{runner,end-to-end,act,lxc-helpers,forgejo-curl} code.forgejo.org/api/v1/repos/forgejo-helm/forgejo-helm code.forgejo.org/api/v1/repos/actions/{cascading-pr,setup-forgejo,forgejo-release,ovh-dns-update} ; do
 forgejo-curl.sh --verbose --debug api_json "https://${repo}/issues/comments?since=${since}&before=${before}" | jq -r '.[] | .user.login'
 forgejo-curl.sh --verbose --debug api_json "https://${repo}/issues?since=${since}&before=${before}" | jq -r '.[] | .user.login'
 done
) | sort -u | grep -v -e '^translate$' -e '^Ghost$' -e '^forgejo-' -e '^thefox$' -e '^end-to-end$' -e '^cascading-pr$' -e '^viceice-bot$'| sed -e 's|^|- https://codeberg.org/|'

So between 2025年02月01日T00:00:00+00:00 and 2025年03月01日T00:00:00+00:00
I would also try with Z offset instead of +00:00 in order to simplify it a little bit.

However, I have the following concerns:

1. Concerning the issues list, as also mentioned 2 days ago, it seems that it is paginated and by default it uses a page size of 30 (meaning that it might happen to not return all of them).
2. The since and before parameters are mapped to UpdatedAfterUnix respectively UpdatedBeforeUnix search options. This means that if one created an issue few months ago (and did not had any other activity meanwhile) and someone adds a comment to that issue during the period of interest, the author will still end up in the contributors list. Right?!

Moreover, don't know if I am doing something wrong or some limitations were added after the DDoS issues, but when I run the script I get 401 errors for repositories from code.forgejo.org.

If someone else wants to just prepare the list without answering all my questions I don't mind.

Moreover, don't know if I am doing something wrong or some limitations were added after the DDoS issues, but when I run the script I get 401 errors for repositories from code.forgejo.org

You could get 403 or 429 but 401 suggests a token is not right. You can use curl instead of forgejo-curl.sh api_json. There is no added value there.

Concerning the issues list, as also mentioned 2 days ago, it seems that it is paginated and by default it uses a page size of 30 (meaning that it might happen to not return all of them).

Correct. This shell snippet is not very good.

The since and before parameters are mapped to UpdatedAfterUnix respectively UpdatedBeforeUnix search options. This means that if one created an issue few months ago (and did not had any other activity meanwhile) and someone adds a comment to that issue during the period of interest, the author will still end up in the contributors list. Right?!

Right. It would be better to further filter on the creation date, maybe with jq?

You could get 403 or 429 but 401 suggests a token is not right.

The token was indeed made for codeberg.org. So before seeing you comment I already tried to logout and login again with a code.forgejo.org token; after this 401 was returned for codeberg.org repositories. Therefore I don't know how similar scripts (I started with the one from the previous monthly update PR) worked for both sites, using a single token. I can of course workaround this by first getting the list from codeberg.org, logout and login with the second token and then also get the list from code.forgejo.org.

Ok... for the moment I used the approach with 2 tokens (one for codeberg.org and the other for code.forgejo.org) in order to prepare the list of February contributors.

@earl-warren wrote in #564 (comment):

Right. It would be better to further filter on the creation date, maybe with jq?

I'm not sure this will help; for the moment I will just ignore this and maybe take a closer look tomorrow or some other day.
Maybe it would also be better to continue this discussion in a separate item.

@floss4good wrote in #564 (comment):

You could get 403 or 429 but 401 suggests a token is not right.

The token was indeed made for codeberg.org. So before seeing you comment I already tried to logout and login again with a code.forgejo.org token; after this 401 was returned for codeberg.org repositories. Therefore I don't know how similar scripts (I started with the one from the previous monthly update PR) worked for both sites, using a single token. I can of course workaround this by first getting the list from codeberg.org, logout and login with the second token and then also get the list from code.forgejo.org.

I was using the script without being logged in: there is no need for a token. This is why it worked for both sites.

Regarding API pagination, you can apply limit=50 to the URLs to get slightly more without implementing pagination in the script.

Here's a patch that includes Weblate-only contributors: 0ko/forgejo-website@fa559376cd

Note that due to different sorting some entries are moved around, causing the - lines. Feel free to pipe it back to the other sorter before applying.

I was using the script without being logged in: there is no need for a token. This is why it worked for both sites.

The reason for which I logged in is that when I tried without it, I got an error (and from the logs I can see it tries to read the header from ~/.forgejo-curl/header-token):

user@host:~> forgejo-curl.sh --verbose --debug api_json "https://codeberg.org/api/v1/repos/forgejo/forgejo/issues"
+ PS4='${BASH_SOURCE[0]}:$LINENO: ${FUNCNAME[0]}: '
/usr/local/bin/forgejo-curl.sh:272: main: true
/usr/local/bin/forgejo-curl.sh:273: main: case "1ドル" in
/usr/local/bin/forgejo-curl.sh:323: main: shift
/usr/local/bin/forgejo-curl.sh:324: main: api_json https://codeberg.org/api/v1/repos/forgejo/forgejo/issues
/usr/local/bin/forgejo-curl.sh:63: api_json: api -H Content-Type:application/json https://codeberg.org/api/v1/repos/forgejo/forgejo/issues
/usr/local/bin/forgejo-curl.sh:59: api: client -H @/home/user/.forgejo-curl/header-token -H Content-Type:application/json https://codeberg.org/api/v1/repos/forgejo/forgejo/issues
/usr/local/bin/forgejo-curl.sh:102: client: log_verbose curl --cookie /home/user/.forgejo-curl/cookies -f -sS -H @/home/user/.forgejo-curl/header-token -H Content-Type:application/json https://codeberg.org/api/v1/repos/forgejo/forgejo/issues
/usr/local/bin/forgejo-curl.sh:32: log_verbose: true
/usr/local/bin/forgejo-curl.sh:33: log_verbose: log curl --cookie /home/user/.forgejo-curl/cookies -f -sS -H @/home/user/.forgejo-curl/header-token -H Content-Type:application/json https://codeberg.org/api/v1/repos/forgejo/forgejo/issues
/usr/local/bin/forgejo-curl.sh:24: log: echo curl --cookie /home/user/.forgejo-curl/cookies -f -sS -H @/home/user/.forgejo-curl/header-token -H Content-Type:application/json https://codeberg.org/api/v1/repos/forgejo/forgejo/issues
curl --cookie /home/user/.forgejo-curl/cookies -f -sS -H @/home/user/.forgejo-curl/header-token -H Content-Type:application/json https://codeberg.org/api/v1/repos/forgejo/forgejo/issues
/usr/local/bin/forgejo-curl.sh:103: client: curl --cookie /home/user/.forgejo-curl/cookies -f -sS -H @/home/user/.forgejo-curl/header-token -H Content-Type:application/json https://codeberg.org/api/v1/repos/forgejo/forgejo/issues
curl: option -H: error encountered when reading a file
curl: try 'curl --help' or 'curl --manual' for more information
/usr/local/bin/forgejo-curl.sh:104: client: fatal_error
/usr/local/bin/forgejo-curl.sh:42: fatal_error: log_error
/usr/local/bin/forgejo-curl.sh:28: log_error: log
/usr/local/bin/forgejo-curl.sh:24: log: echo
/usr/local/bin/forgejo-curl.sh:43: fatal_error: true
/usr/local/bin/forgejo-curl.sh:44: fatal_error: exit 1
user@host:~> 

Regarding API pagination, you can apply limit=50 to the URLs to get slightly more without implementing pagination in the script.

To be honest a list of 174 contributors is already quite long (for a blog post) from my point of view; however, if I use limit=50 only 3 extra lines are added (so I included them).

Here's a patch that includes Weblate-only contributors: fa559376cd

Done.

To be honest a list of 174 contributors is already quite long...

I agree. It is significantly more than before which makes me wonder if there are not too many?

Maybe we can add them as a comma separated list of usernames with links to their profile page:
[someuser](https://codeberg.org/someuser), [anotherone](https://codeberg.org/anotherone), ...

I wonder if all of those actually contributed something in the time period. 170 gives me the feeling that some of them are included even though they did not contribute a comment / issue / pull request. But maybe I did not pay enough attention and there was a lot of people in February. 🤔

I wonder if all of those actually contributed something in the time period.

I think it's not that easy to check. However, after a quick look at the results from the /repos/{owner}/{repo}/issues/comments endpoint for /forgejo/forgejo repository I would say that the list is pretty genuine. And only for this repository a list of 118 comment authors is returned (the number goes up to 148 when retrieving the comments from /forgejo/{design,forgejo,website,docs,tools,governance,sustainability,discussions}).

EDIT: Correct numbers: 127→118 / 163→148

Maybe we can add them as a comma separated list [...]

Ok... I see 2 thumbs up for this proposal.
Should I proceed with the change or wait for more reviewers?!
Anyone else having better suggestions?

I think it is ready to be published. Let me know if I should push the button 😇

It needs an update of the publication date (19 March instead of 17 March).

@floss4good wrote in #564 (comment):

I wonder if all of those actually contributed something in the time period.

I think it's not that easy to check. However, after a quick look at the results from the /repos/{owner}/{repo}/issues/comments endpoint for /forgejo/forgejo repository I would say that the list is pretty genuine. And only for this repository a list of 127 comment authors is returned (the number goes up to 163 when retrieving the comments from /forgejo/{design,forgejo,website,docs,tools,governance,sustainability,discussions}).

https://codeberg.org/caesar?tab=activity they are listed although there has been no activity in a while 🤔

@floss4good I think the list of contributors being slightly off is not a blocker. It can be updated later. Let me know if this is good to merge.

First of all, I've made a silly mistake and I want to clarify this; the numbers from this comment were also taking into account the /repos/{owner}/{repo}/issues endpoint; I have updated the above comment (127→118 / 163→148).

Secondly, concerning caesar, it is included at least because of forgejo/docs#51 (created by him 2 years ago and edited by someone else last month); so my second concern listed in this comment is now proved.

As for the merge, I think you and other reviewers should know better than me if the PR is ready (since for me it's the first time I do this). Personally I don't have any change to add (besides the contributors list, which could take some time to validate and I prefer to do something more productive with that time).

This is excellent work 👍 Let's merge in a few hours unless someone objects then.

@mahlzahn will you publish the floss.social announcement?

Room announcement posted at https://matrix.to/#/!qjPHwFPdxhpLkXMkyP:matrix.org/$wusXgTb8SLDIjiJ_cn5h76ktaRJSKK4b1JLzLPzg-rk?via=matrix.org&via=tchncs.de&via=envs.net (and description updated).