friendica.myportal.social/disp…
This is an example of one that I've had, but when I come across them, there is no rhyme or reason as to what they have in common. Has anyone else had this issue?
xy.. likes this.
!Friendica Developers
The post-tag and post-user tables are extremely large at 140GB and 145GB and continue to grow. It seems as if the two tables have lost the ability to delete data. The optimisation via OPTIMIZE TABLE takes a correspondingly long time. I can also reproduce this behaviour on two other, independent nodes, although not to the same extent (fewer users).
Question: is there a kind of clean-up script that I can run from time to time to delete content that has long since expired?
xy.. likes this.
dbclean-expire-limit on your system? If not set, then please set it to 100000.
@Michael 🇺🇦 The SQL logs do not give an error on cancelled ExpirePosts. For 'dbclean-expire-limit' a value 30000 was stored. Is now changed to 100000.
I had followed a discussion here that the value 100000 was set too high and had corrected it. The value you recommended was previously set, but without effect.
Let's see if that changes anything now👍 .
This has happened twice.
Click a comment, report, collapse all post, then report the post to the server owner, then submit, and it takes you to a new report.
But with the box asking the email or url to report.
Would a redirect back to the original page that you reported the comment on, be possible? Or is that lost during the three clicks into the form?
!Friendica Developers Now that the blurry avatars issue fix is confirmed and pushed to dev github.com/friendica/friendica… are there remaining blockers that prevent a RC for being cut?
Am aware of ongoing trouble irt RSS feeds no longer showing content in dev branch. But maybe that is something that could be investigated during RC phase?
Fix blurred images by annando · Pull Request #14873 · friendica/friendica
This could fix #14800GitHub
Question about queues
!Friendica Developers
I noticed I am getting over 240K /inbox requests from lemmy.world each day. My understanding is that if anyone on this server follows a community there that every upvote, downvote, reply, post, in that community will cause an /inbox request to this server. They have many huge communities.
Are these 240K /inbox requests each creating one job each in the queue?
Andy H3 likes this.
/inbox might be processed immediately if the post hierarchy already exists, like it is the case for most activities.
How to wget large files from one server to another
#fediAdmins
@Friendica Support @Friendica Developers
Hi there,
a technical question not directly related to friendica itself.
Trying to transfer an instance from on (shared hosting) server to a debian/ubuntu VPS that's building up.
The problem now is that the DB is around 3.7 gig so I tried to pull the file with wget.
The server responds with:
rootName@VPShosting:/var/storage# wget https://domain.com/folder/DB.sql.bz2
..
Connecting to domain.com (domain.com)|xxx.xxx.xxx.xxx|:443... connected.
HTTP request sent, awaiting response... 413 Request Entity Too Large
ERROR 413: Request Entity Too Large
Can't find a solution on the web.
Google even manages to point to howTo's behind paywall's or subscribe pages, so much about enshittification ..
Did find nginx references but do not have nginx.
Also found for php.ini the following, obviously the values a relative:
max_input_time = 24000
max_execution_time = 24000
upload_max_filesize = 12000M
post_max_size = 24000M
memory_limit = 12000MAs of now changed:
upload_max_filesize = 8000M
post_max_size = 8000mmemory_limit = -1so that's probably not the point.
The error message stays the same.
The input_time and execution_time should kick in in any case while downloading the file.
Ideas how to solve this?
reshared this
I assumed you have ssh accessNope.
[…]
investigating how to do this with sFTP over SSH
I am confused. If you do not have ssh access, sftp will fail.
reshared this
I am confused. If you do not have ssh access, sftp will fail.
two server
- server 1
shared hosting
hosts the file
access over sftp
no access over SSH
access over https, locating the file in a public folder, access with wget, works at least up to 1.9gig - server2
VPS hosting
access over SSH
never intended or tested to access over sFTP till now
needs/requests the 3.7gig file
Probable sFTP option:
rootName@VPShosting:# someSFTPtool sftpUSERNAME@sftpsSharedhostingID/sharedHostingstoragefolder/file.tar.gz /thisVPSserver/localstorage/file.tar.gz
reshared this
I still have problems with this bug and it's mean on the RC 🤕
github.com/friendica/friendica…
Matthias likes this.
[Forgejo Migration] Migrate GitHub issues to Forgejo · Issue #13391 · friendica/friendica
Part of Forgejo Migration: Migrate GitHub issues to Forgejo Depends on https://codeberg.org/forgejo/forgejo/issues/340 Blocked by https://forum.forgefriends.org/t/about-the-friendly-forge-format-f3...GitHub
like this
Postbox Add-Ons for Friendica
Bookface 1.6 introduces a new feature called "Postboxes" that allows you to create a post with a colorful background like a similar feature on Facebook. It's built into the Bookface scheme, but I also created TWO add-ons for Friendica if a server administrator wants to roll it out for all of their users regardless of what theme they prefer.
Postbox: simply adds a minified stylesheet to the head with the Postbox styling. It doesn’t add any interface for creating a Postbox, but users can still create them manually with BBcode.
Zen Postbox: adds the minified stylesheet to the head, but also adds a tool to the post composer to insert a Postbox or wrap selected text in one.
To manually create a Postbox just wrap your text in the Postbox class BBcode like this:
[class=postbox-redblue]Text that will be wrapped[/class]
Make sure to PREVIEW it before you post it! While Friendica's Postboxes do allow more content than just plain text, certain formatting blocks can't be placed inside them.
If everything looks good, go ahead an POST it. Anyone using Bookface or any server using either the "Postbox" or "Zen Postbox" add-ons will see it like this:
If the "Zen Postbox" add-on is enabled it adds a button for wrapping/inserting a Postbox to the post editor:
Full Documentation of the "Zen Postbox" add-on is available as a GitLab Wiki.
Keep in mind that other servers that don't implement Postbox styling, other platforms (like Mastodon), and third-party apps will not show these colored backgrounds. This feature only works in the Friendica webapp/website.
#bookface #friendica @Friendica Admins @Friendica Developers
like this
reshared this
[class] wrapper since other platforms can't use it anyway. There isn't anything similar in Markdown so far as I know, which might be why the raw BBcode is getting through to Disapora? I'm guessing it's missing the conversion from BBcode [class] to Markdown syntax. Which should be conversion to nothing. I'd check for it myself but I have no clue where the processing for other platforms even is in the Friendica code.reshared this
Ok, I’ve been trying to follow the Friendica code and I may be reading it wrongly, but it looks to me like the BBcode gets converted to HTML for most platforms, but then that HTML gets converted to Markdown for Diaspora (and I think Hubzilla?). But if that’s the case I don’t understand how the BBcode is getting all the way through to those other platforms as shown in the screenshots @Matthias posted?
It looks to me like it should be sending the <span> to them as well not the raw BBcode, because it seems the HTML to Markdown converter by default "preserves HTML tags without Markdown equivalents like <span> and <div>." At least according to the README in /friendica/vendor/league/html-to-markdown/ - which also says [i]"To strip HTML tags that don’t have a Markdown equivalent while preserving the content inside them, set strip_tags…"
So in:
/friendica/src/Content/Text/HTML.php
If I change line 684 from :
$converter = new HtmlConverter([‘hard_break’ => true]);to
$converter = newHtmlConverter([‘hard_break’ => true, ‘strip_tags’ => true]);The Postbox
<span> is removed but whatever’s inside it is not. Is there any particular reason to preserve HTML that has no Markdown equivalents for platforms that want Markdown format?Or am I completely barking up the wrong tree here?
like this
reshared this
#houston .. I mean @Friendica Support and @Friendica Developers, we have a problem!
Look's like bbcode code doesn't get parsed as code on mastodon.
tupambae.org/display/0ac89072-…
troet.cafe/@bitpickup/11420448…
Cătă likes this.
reshared this
Looks first of all like a troet.cafe specific circunstance @bitpickup, what code is troet running?
The combination of two audio files might be problematic in general. Actually it's not possible over here to upload more than one video or audio and the replay of any of the two files displayed with timer might not work.
reshared this
Hi @developers
since my #Friendica server social.yl.ms seems to be unrecoverable, it seems there was a fault in the database that now prevents me to restore the backups I got and I have no motivation to restore it anymore you read this post coming from Pleroma.
I will attempt to update git.friendi.ca in a few minutes, therefor I will shutdown the server, do a backup of the VM and restart the server and start updates in the VM. Then I will reboot the VM.
Finger crossed this will only take a few minutes.
Please could anyone let me know if the message actually shows up in the dev forum?
like this
Friendica Developers reshared this.
Philipp Holzer likes this.
Friendica Developers reshared this.
!Friendica Developers Hi, how can I read in Friendica the JSON Acccount information (display name, summary, ...)?
For Mastodon I can call
https://kaputtgart.social/api/v1/accounts/lookup?acct=<username>
Also the profile URL
https://mastodon.social/users/<username>
returns this JSON information.
When I call the same Friendica URL
https://friendica.instance/profile/<username>
I get a JSON response where some data is missing or is wrong. ie. the "name" field has the same value as the nickname and not the display name. and the "summary" is missing with the account description.
So how can I access from any Freindica Instance the account display name und summary description for a address in JSON?
Fitik likes this.
https://friendica.world/noscrape/lamf.
Roland Häder likes this.
Roland Häder likes this.
https://friendica.world/api/v1/accounts/lookup?acct=lamf@friendica.world?
Creating several friendica instances and other websites as sub/domains on one VPS with one IP
v.02
@Tutorial
@Friendica Developers and @Friendica Admins.
#howTo change a single #friendica domain #debian #ubuntu 2022.04lts VPS #hosting setup into a single IP multi domain VPS server.
This tutorial builds on the install and/or move friendica to ubuntu 22.04 LTS VPS server howTo with the intention to change the main setup into a domain name-based virtual hosting.
The folder name in the VPS folder structure in /var/www/for the domains can be choosen as you like, for example:
/var/www/domaincom/
All apache tutorials found in a quick search do choose the domain or subdomain name including the respective dot's like this:/var/www/domain.com/
Both options are possible for a correct working environment.
@Ⓜ3️⃣3️⃣ 🌌 pointed out that for automatization scripts it's usefull to keep that consistent and there for to use FQDN (Fully Qualified Domain Name) as part of the path.
Never the less, to keep the concepts of domain URL's and folder names seperate and to prevent the notion that the domain name syntax is mandatory we will stick with simple folder names in this #tutorial.
The naming of the .conf files in /etc/apache2/sites-available/ refer to the actual domain names and their settings. [Even tho these .conf file names also could be chosen probably at will?], as the definitions about the actual domains refereed are given inside those files, we'll stick to the real sub/domain names in this case as it's quite reasonable and logical to do so.
We will address the details for a step by step change from a single VPS single IP domain hosting to a multiple name-based single IP virtual hosting in answers to this topic:
- Inicial VPS domain hosting served at:
/var/www/html/ - Name-based virtual hosting domain VPS where several domains are served by one IP, leaving /html in place if a client request doesn’t match any other sites, like proposed in this digitalOcean tutorial:
/var/www/html/as fall back option/var/www/domaincom/index.html/var/www/subdomaincom/index.html/var/www/domainorg/original friendica installation/var/www/subdomainorg/index.html
Any comments hints and observations are really appreciated!
#admin #fediAdmin
TekNo ⚝ aEvl likes this.
reshared this
<VirtualHost _default_:80>
<VirtualHost _default_:443>
However, please note this will only work with plain HTTP connections, not with HTTPS (HTTP over SSL): name-based virtual hosts rely on knowing what name is being requested, but this information can't be known by Apache until the encrypted SSL connection is established.
serverfault.com/questions/4353…Looks like the conflict as of now is around the SSL settings and port 443.
Apparently it's the one reserved and dedicated for SSL connections and the respective settings and configuration.root@VPShosting:/etc/apache2/sites-enabled# ls -l000-default-le-ssl.conf -> /etc/apache2/sites-available/000-default-le-ssl.conf
cast.domain.org.conf -> ../sites-available/cast.tupambae.org.conf
domain.org.conf -> ../sites-available/tupambae.org.conf
The revision of:
nano /etc/apache2/apache2.conf
didn't result in mayor relevant settings or settings that had been modified.
It basically add's/etc/apache2/conf-enabledand/etc/apache2/sites-enabledas sources for the effective active configuration in place.Looks like averything boils down to
000-default-le-ssl.confwhich was (probably) created when installing certbot let's encrypt SSL in the inicial LAMP instalation.Also in the orginal installation of this server 000-default-le-ssl.conf was mentioned and modified.
It all looks like the content of
000-default-le-ssl.confshould be basically ported entirely to the filedomain.org.confchanging the port of that file to 443 and to than decativate000-default-le-ssl.confby executing:
a2dissite 000-default-le-ssl.confA last question that comes up now is with respect to SSL connections for subdomains of domain.org. If those are already included in the issued let's encrypt SSL certificate or if there have to be made changes to that inicial cert instalation.
Also there will have to be a new SSL cert for eventuell new domain.com sites installed on this VPShosting server.
reshared this
Solved as follows:
The SSL let's encrypt cert is created out of the http file of the given domain or subdomain:/etc/apache2/sites-available/domain.org.conf
standard domain.org.conf
<VirtualHost *:80>
ServerAdmin admin@domain.org
ServerName domain.org
DocumentRoot /var/www/domainorg
# AllowOverride is specifically added for the friendica server
<Directory /var/www/domainorg>
AllowOverride All
</Directory>
# AllowOverride is specifically added for the friendica server
ErrorLog ${APACHE_LOG_DIR}/domainorg_error.log
CustomLog ${APACHE_LOG_DIR}/domainorg_access.log combined
</VirtualHost>
The SSL domain reference file created from the http filename get's added by default -le-ssl to the existing domain.org.conf name. The port for SSL connections is 443 while the standard http connection is over port 80.
domain.org-le-ssl.conf
This is the content of the file created for a multi domain VPS server. This contenet matches the content that was in the original 000-default-le-ssl.conf file of the single domain VPS setup, created by the certbot routine. The SSL certificate is supposed to be updated automatically every year. For now it is not clear if the changes made will be affected negatively by that automatic SSL cert update.<IfModule mod_ssl.c>
<VirtualHost *:443>
ServerAdmin admin@domain.org
ServerName domain.org
DocumentRoot /var/www/domainorg
# AllowOverride is specifically added for the friendica server
<Directory /var/www/domainorg>
AllowOverride All
</Directory>
# AllowOverride is specifically added for the friendica server
ErrorLog ${APACHE_LOG_DIR}/domainorg_error.log
CustomLog ${APACHE_LOG_DIR}/domainorg_access.log combined
ServerName domain.org
Include /etc/letsencrypt/options-ssl-apache.conf
SSLCertificateFile /etc/letsencrypt/live/domain.org/fullchain.pem
SSLCertificateKeyFile /etc/letsencrypt/live/domain.org/privkey.pem
</VirtualHost>
</IfModule>activating and deactivating .conf fles
To make the changes come into effect the new conf file has to be activated, the old one that shouldn't apply anymore should be deactivated and the apache server itself has to be reloaded or to be restarted:rootname@VPShosting:~# a2ensite domain.org
rootname@VPShosting:~# a2ensite domain.org-le-ssl.conf
rootname@VPShosting:~# a2dissite 000-default.conf
rootname@VPShosting:~# a2dissite 000-default-le-ssl.conf
rootname@VPShosting:~# systemctl restart apache2
Reloading the site took a while longer than nomal, indicating that new background tasks where happening.
testing the new folder structure
and if site configuration was actually in place.Renaming the index file of the old /html folder.
Reloading the site did still work so the old /html folder structure wasn't active anymore.
Renaming the entire old /html folder structure taking it out of access completely.
Creating a new /html folder.
Creating a new html. index file in /var/www/html.
rootname@VPShosting:~# mv /var/www/html/index.php /var/www/html/index_old.php
rootname@VPShosting:~# mv /var/www/html/ /var/www/html_old/
rootname@VPShosting:~# mkdir /var/www/html/
rootname@VPShosting:~# nano /var/www/html/index.html
<html>
<title>name-based virtual hosting setup</title>
<h1>Welcome to the /html/ index.html fall back file.</h1>
<p>This is a test file for a name-based virtual hosting setup</p>
</html>Next steps:
Installing a SSL cert for subdomain.org.
Installing a subdomain test site.
Installing a new friendica domain.com site from scratch.
reshared this
Question about how things work
Question for understanding:
When I run this query, I see the following output:
select command, pid, count(*) as processes FROM workerqueue WHERE done = 0 and pid != 0 GROUP BY command ORDER BY processes
DESC;
+------------------------------------+--------+-----------+
| command | pid | processes |
+------------------------------------+--------+-----------+
| UpdateContact | 982072 | 5 |
| APDelivery | 895941 | 2 |
| ContactDiscovery | 895941 | 2 |
| ClearCache | 982947 | 1 |
| addon/bluesky/bluesky_feed.php | 944103 | 1 |
| UpdateScores | 983740 | 1 |
| ProcessQueue | 982642 | 1 |
| UpdateGServer | 981787 | 1 |
| ProcessUnprocessedEntries | 983726 | 1 |
| PullDirectory | 982495 | 1 |
| addon/bluesky/bluesky_timeline.php | 976074 | 1 |
+------------------------------------+--------+-----------+
11 rows in set (0.001 sec)
Are each of these pids one of the "parallel workers" that I configured? I configured quite a bit more than 10 of them. Does this mean I configured excessively many? And are the processes within the number of units of work I specified that can be fetched by each of the parallel workers? Or am I looking at something entirely different?
Thanks!
Can I simply rename the username in the userd table to allow it to be used again?
!Friendica Developers
I have someone who deleted their account on my server a while ago and regretted switching to a different instance and wants to return. But their username is now in the userd table and I assume this is the reason they are being told the name is not available.
Because I don't know all the table relationships, I don't want to just go in using an UPDATE statement and rename the username entry in userd to something else, so the name is available again without first checking with a developer.
Can I simply change the username in the userd table, so a name can be reused?
Thanks.
Jerry on Friendica likes this.
ERROR worker Uncaught exception in worker method execution
File Line Function UID Process ID
Worker.php 570 execFunction ba8cbf 171319
Data
class
TypeError
message
Friendica\Core\Storage\Type\Filesystem::put(): Argument #1 ($data) must be of type string, null given, called in /var/www/html//src/Model/Photo.php on line 446
code
0
file
/var/www/html/friendica/src/Core/Storage/Type/Filesystem.php:120
trace
#0 /var/www/html/friendica/src/Model/Photo.php(446): Friendica\Core\Storage\Type\Filesystem->put()
#1 /var/www/html/friendica/src/Model/Photo.php(637): Friendica\Model\Photo::store()
#2 /var/www/html/friendica/src/Model/Contact.php(2423): Friendica\Model\Photo::importProfilePhoto()
#3 /var/www/html/friendica/src/Model/Contact.php(2942): Friendica\Model\Contact::updateAvatar()
#4 /var/www/html/friendica/src/Model/Contact.php(2722): Friendica\Model\Contact::updateFromProbeArray()
#5 /var/www/html/friendica/src/Worker/UpdateContact.php(34): Friendica\Model\Contact::updateFromProbe()
#6 [internal function]: Friendica\Worker\UpdateContact::execute()
#7 /var/www/html/friendica/src/Core/Worker.php(568): call_user_func_array()
#8 /var/www/html/friendica/src/Core/Worker.php(376): Friendica\Core\Worker::execFunction()
#9 /var/www/html/friendica/src/Core/Worker.php(111): Friendica\Core\Worker::execute()
#10 /var/www/html/friendica/src/Console/Worker.php(91): Friendica\Core\Worker::processQueue()
#11 /var/www/html/friendica/vendor/asika/simple-console/src/Console.php(108): Friendica\Console\Worker->doExecute()
#12 /var/www/html/friendica/src/Core/Console.php(171): Asika\SimpleConsole\Console->execute()
#13 /var/www/html/friendica/vendor/asika/simple-console/src/Console.php(108): Friendica\Core\Console->doExecute()
#14 /var/www/html/friendica/src/App.php(234): Asika\SimpleConsole\Console->execute()
#15 /var/www/html/friendica/bin/console.php(22): Friendica\App->processConsole()
#16 {main}
previous
worker_id
8ed9294
worker_cmd
UpdateContactTekNo ⚝ aEvl doesn't like this.
reshared this
Fitik likes this.
reshared this
pataphysician likes this.
reshared this
reshared this
Friendica is an ideal platform to combine several systems. Means: You could create a Wordpress blog and could enable the option to automatically reshare all blog entries via your Friendica account. Same would be possible with - for example - Videos from Peertube.
People just have to follow your Friendica account and will always get the whole content.
like this
reshared this
#bugResearch @Friendica Developers
Platform Info: 2023.05 - stable
Friendica Version: 2023.05 - outdated2023.05 - stable]Friendica Version: 2023.05 - stable
2023.05 - stable]Friendica Version: 2023.05 - dev
#tusky
I didn't have a look at gitHub
Short bug description:
Repeated sending of one short comment from the @Tusky APP of an answer to a post send to the helpers forum.Details of the buggy problem:
Even tho the problem of multiple answers that ended up looking like spamming the federation itself was mitigated by the owner of the originating initial post to helpers and by the profileowner of the multiple answers, the deletion didn't federate thru out the system.¿Possible to reproduce, if so how:
we prefer not to do so.Expected result:
¿what should be the correct behavior?Interesting question.
Actually the owner of the answer should have access to delete the comments globally, while the originator of the post should maybe be able to delete the answers delivered by the helpers forum.
Screenshots:
No screen shots needed. Refenrence links and conversations more into deep are in the reshare link.
Information for dev and server team:
php.error.log
LEAVE THIS AS IS, the server team will use this report if possible/necessary
Thx for explaining more into detail @JB Carroll!Look's like that's right here and now a good #bugReport for the @Friendica Developers and/or @Tusky.
Adding details:
The kernel panic by #tusky was probably because of the neat beauty of the UI and was mentioned by JB who apparently runs dev version 2025.02 on ubuntu 2020.lts at the time.
(btw 2024.04lts still is somehow a mess, looks like never update to stable but only to old stable for a serious working environment)The original post was send actually from friendica 2023.05 and forum.friendi.ca is on stable 2024.12.
As of now no idea if @Steffen K9 🐰 ( oh there he is!) is available. He might be the one to delete on the helpers profile itself.Over here the comments were deleted in the moment, so in any case, as the original link points to over here, there is also some harm reduction and some moderation capability already in the setup.
Actually the biggest problem seems to be that the answers do still show up on mastodon.
Looks like for mastodon the owner is the helpers page and not you or the profile over here.So mastodonian @mʕ•ﻌ•ʔm bitPickup, are you gone apologize for being rude or do you choose:
"Go straight to the @Pillory!
Don't pass by the service station, don't catch likes or visite any other notification first."@Jeff Jarvis Jarvis @Eva Wolfangel Wolfangel @Christian Stöcker @taz @taz @Digitalcourage @Cory Doctorow
Friendica Developers reshared this.
Get real client IPs in dockerized php-fpm Friendica deployments
!Friendica Support !Friendica Developers
Hello there, this message will probably not reach Friendi.ca groups for whatever reason... anyway, here is a PR about that thing... I really want to see real remote client's public IP address in Friendica container app image, instead of docker internal/private IP address (meaningless).
(this time cc @Tobias ... if you are admin on Friendi.ca could you tell me what's wrong with my messages delivery please ?)
reshared this
reshared this
reshared this
I will abandon publishing it to groups I’m bored of it 😮💨
reshared this
!Friendica Developers !Friendica Support
all of a sudden my notifier tasks is stuck on the little timer icon, but in the logs it shows done, and I have no errors, im not sure if this a bug, a issue or just a fluke, Friendica 'Interrupted Fern' 2025.02-dev - 1576 what ever is the problem is also preventing my instance/profile from fully federating, I cant get any new posts of my profile in any rss/atom feed reader at all, this just started and I am clueless as to what happened and where to look as there are no errors logged anywhere, and workerqueue shows jobs as done when they are completed, so I am really at a loss, and would like to fix this asap.
here are my current worker settings per admin/site/worker
Maximum load average
20
Maximum system load before delivery and poll processes are deferred - default 20.
Minimal memory
0
Minimal free memory in MB for the worker. Needs access to /proc/meminfo - default 0 (deactivated).
Maximum number of parallel workers
20
On shared hosters set this to 5. On larger systems, values of 20 are great. Default value is 10.
Maximum load for workers
15
Maximum load that causes a cooldown before each worker function call.
Enable fast-lane
When enabled, the fastlane mechanism starts an additional worker if processes with higher priority are blocked by processes of lower priority.
Decoupled receiver
Decouple incoming ActivityPub posts by processing them in the background via a worker process. Only enable this on fast systems.
Worker defer limit
15
Per default the systems tries delivering for 15 times before dropping it.
Worker fetch limit
25
Number of worker tasks that are fetched in a single query. Higher values should increase the performance, too high values will mostly likely decrease it. Only change it, when you know how to measure the performance of your system.
I have dug through every file and the entire database and can not find what is causing this, I can see new posts from other people on other federated instances, I can see my posts, but they are not showing up in my feed reader, indicating they are either not completed tasks or not getting federated, I have been all over this, for hours today and no errors no logs showing any issues at all, yet I clearly have an issue, as you can see the Notifier Task is showing Pending instead of being Checked as done, Help would be greatly appreciated.
I have even tired dropping back to stable, and that still did not work, so I went back to Friendica 'Interrupted Fern' 2025.02-dev - 1576 I really wish I could figure out what is going on, sure would like to have this working as its really my only form of communication with the world
So I updated my php to 8.4.2, but that has not made any difference, and I cant go backwards, as I have something else on the same server that cant go down in versions, I have dug more and more, and have found nothing that is causing this issue, having some kind of idea where to look would be helpful, considering I have done everything imaginable to solve the Notifier Tasks Pending problem of hangging and have yet to find a solution, so I have tried to think outside the box and find a solution that does not include debugging, checking error logs, confirming config files, and assuming the database is correct and healthy, find another way to address this issue, and I am hitting a brick wall.
so now I just did git reset --hard origin/develop, and the problem is still there. 🙁
I have even created a test user on my instance, and verified it is instance wide.
reshared this
UPDATE `inbox-status` SET `success` = '2025-01-31 08:53:46', `gsid` = 129, `archive` = 0 WHERE (`url` = 'https://mastodon.sdf.org/inbox')MySQL server has gone away
TekNo ⚝ aEvl likes this.
reshared this
reshared this
@Friendica Support @Friendica Developers
#friendica #featureRequest
Hi there,
is there a way to change locally the profile picture displayed of a contact?
Haven't seen any option to that regard in the contact page.
I'm asking in part specifically because of the #eyeContact issues some people have. It's about the fact that they have real problems facing full face pictures that kinda "look straight at you", actually a known psychological effect not only in primates.
(Just stare consistently at a dog and you'll see. Well actually better you don't, unless you know what you are doing)
At the same time that's probably not the only reason a avatar could be unwanted for someone. So, optional I could think of or a standard feature that disallows the display of a avatar or a feature that allows to replace the avatar picture with something chosen by the user.
🤔
reshared this
github.com/friendica/friendica…
github.com/friendica/friendica…
github.com/friendica/friendica…
github.com/friendica/friendica…
reshared this
Roadmap for Friendica 2025.02
Hey there,
my current idea for the release of 2025.02 would be the following.
- Tuesday after FOSDEM I will switch my personal node to the current development branch. Soon after I would like to branch off an RC branch. Except anyone of you is aware of a serious bug that should be fixed before the testing phase for the next release is started.
- Giving two week of RC phase, this would aim for the weekend of the 22./23. February for the release of the next stable version of Friendica.
Has anybody of you objections against the plan?
like this
Issue is github.com/friendica/friendica…
Remote avatars blurred or missing · Issue #14800 · friendica/friendica
[x ] I have searched open and closed issues for duplicates Bug Description On current develop version avatars / profile pictures of remote users are blurred or completely missing. This might be dep...GitHub
Der Marek likes this.
Steffen K9 🐰 likes this.
Steffen K9 🐰 likes this.
Save to folder function broken or my Browser?
@Friendica Developers @Friendica Support
Does "Save to folder" function work for you on the current stable and dev?
When ever I enter a folder and confirm, the browser (Firefox) freezes until I refresh it. The browser asks on refresh if I really want to leave the site since it is still processing.
After the manual refresh, the site is correctly saved.
Background: I was diving in the Code to adjust the color for the "save to folder" dialog, which is actually colorbox. Are we interested to make it a modal like the compose window?
I already fixed the color appearance, but am reluctant to create a PR on a may be faulty feature in general.
PS: I am so grateful for the dev container for friendica, it simplifies the dev process at leat for ui tremendously! ❤
xy.. likes this.
reshared this
@Der Marek The feature works for me, I've fixed the awkward UX recently on develop: github.com/friendica/friendica…
Improving the display would be welcome, by using the regular modal window instead of the the colorbox for example.
like this
reshared this
Ahh okay, thank your very much! I pulled your PR and now the save works like a charm.
On the other hand, the background color is not fixed for frio dark / black.
I created an issue on github, elaborating my assumption for the cause for better discussion:
Hypolite Petovan likes this.
reshared this
anubis2814
in reply to anubis2814 • • •OldKid ⁂
in reply to anubis2814 • • •@anubis2814 If you want to remove the share from a post you shared, simply click the share button again to cancel the share.
If you only delete it locally, the share will still be there and you will continue to receive notifications. Deleting it locally only means that you delete it from the display in your timeline. It is still present on your instance and so you will continue to receive notifications.
anubis2814
in reply to OldKid ⁂ • • •