Let's visually mark this to attract people.

I've carefully read the diff :) Thanks

Totally agree. Thanks for any edit in the stub.

Have you digged Internet Archive to fix this? :)

In T15121#3880, @valerio.bozzolan wrote:

I don't think that the author should ever allow to re-license its work. [...] Normal volunteers do not accept that.

Go pohrge!

Clean up my local testing mess

Most phurl.io short URIs point to old upstream tasks. The two related patches make the code at least point directly to these tasks (to still be found on https://web.archive.org/ if secure.phabricator.com ever went down) as there's no direct content equivalent on phorge.it.

@valerio.bozzolan Are there any prehistoric traces that this extension has been successfully used in the past? :P

I tried with only merge this commit https://phabricator.wikimedia.org/rPHABae334e4589a0413c1eb8a9e8d993e553bfeff478
but still not work.

It's a Phacility server, managed at https://secure.phabricator.com/phurl/

I agree, badges seems like the right way to go. They are an underutilized feature really, IMO.

@littleggghost So it turns out that in order to enable this feature, there was a patch to Phabricator core. That patch isn't part of the wikimedia extension, and hasn't been merged into phorge.

I've looked down some of the links of this request, and it looks like the definition of "new user" should be install-specific, and open to interpretation.

When I search in

resources/celerity/map.php

I run this cmd in this docker container, with output:

root@f0d9837493f0:/var/www/phorge/phorge# ./bin/celerity map
Rebuilding 1 resource source(s).
Rebuilding resource source "phabricator" (CelerityPhabricatorResources)...
Found 113 binary resources.
Found 399 text resources.
Found 11 packages.
Writing map "resources/celerity/map.php".
Done.

Makes sense, thanks everyone. Let's decline for the time being.

@littleggghost one thing you might want to try is updating the celerity map. You may be missing the custom css rules to style the deadline appropriately. I’m not familiar enough with the phorge docker image to be able to tell you exactly how that is accomplished. With non-containerized phorge instances you would just run bin/celerity map from within the phorge root directory.

I changed my mind again. No need for edges (short story is, getting the "color" property from the edge is annoying).

In T15743#15855, @valerio.bozzolan wrote:

• https://github.com/xemlock/php-latex (PHP server side)

Interestingly, I just discovered Wikimedia has a pretty cool project to render math on the server side, by transpiling some js into php:

I went down a rabbit hole while trying to add unit tests for this. Our core remarkup unit tests are pretty thorough but they only seem to cover the "core" remarkup engine and syntax. When I added a new test case for this it would pass but always caused a different test case to fail in weird ways. I haven't yet figured out what part of the unit test or engine is stateful such that my unit test alters the state in a way that breaks a later test. If I change the order that tests run in, it just causes another different test case to fail.

Fix lint again

fix lint.

Added contrasting color calculation

There is also MathJax, a subset of LaTex that only covers the mathematical notation without the page layout functionality.

Unrelated note, but interestingly the contributor agreement chooser (https://contributoragreements.org/) is broken. I've already sent an email to their support.

I see that it depends on a NodeJS external script (server-side).

Interesting, I agree. Note that also #bbb is valid but I see that 6 is minimum characters.

Interesting. What's the use case you have in mind?

This is not yet tested.

fix lint error.

This would be useful for discussions about design, specifically I wished for this feature while reviewing D25491: Improve contrast of Links in Dark Mode

Yeah, I don't like the diviner page.
For other included 3rd parties, we often put a license file right next to the asset:

• https://we.phorge.it/source/phorge/browse/master/webroot/rsrc/externals/font/lato/LICENSE.txt
• https://we.phorge.it/source/phorge/browse/master/webroot/rsrc/externals/d3/LICENSE
• https://we.phorge.it/source/phorge/browse/master/externals/mimemailparser/LICENSE
• and, https://we.phorge.it/source/phorge/browse/master/externals/octicons/LICENSE which is not next to the actual content.

In D25491#15484, @20after4 wrote:

In D25491#15482, @Iniquity wrote:

Hello! Are you sure changing the links color is a good idea? Perhaps we should just make the background darker?

I actually think a darker background and lighter link color would be good. Maybe not quite as light as #accadf, perhaps #46B2FF or so

Volkan 🌋 ural

In D25079#15507, @valerio.bozzolan wrote:

https://we.phorge.it/T15178#4880

I'm also unsure and I would agree if somebody suggests to omit the Diviner page. Relevant discussion also from @avivey

I guess I'm fine with adding that mp3 file.
I am personally not convinced by the creation of src/docs/flavor/project_credits.diviner: I've seen way too many "let's create some list, with best intentions, and at some point unfortunately forget about updating it" examples in FOSS projects. Every added line in software projects creates additional maintenance costs (and in some cases bitrot); human resources are in theory unlimited and in reality rather sparse.
Credits could go into the git commit message, maybe?

It's the anniversary of this super-important killer feature 💌

In T15719#15304, @Tgr wrote:

Given how it's displayed on the Gmail UI, IMO unsubscribing from all emails or at least all emails of the given type (e.g. all Maniphest emails if it's a Maniphest notification) is the only thing that makes sense. If I mark an email as spam, and then use the unsubscribe option instead, I would not expect to continue getting identical email from all the other tasks.
No idea how (or if) other email clients handle the header, though.

This was both harder and easier then I expected. I got to crash the IDE and hard-code a bunch of stuff, but this actually works. Didn't even start on the Edges implementation yet - I'm directly using the Flags table.

See discussion and reasons in https://secure.phabricator.com/T11833

In D25491#15482, @Iniquity wrote:

Hello! Are you sure changing the links color is a good idea? Perhaps we should just make the background darker?

The only progress I've made is to think through the implementation but I haven't actually started coding on it. It is sort of high on my wish list so I may get to it soon.

The word solution would solve it from the expecting side, you are right. Here in my company the search is used frequently and only as a plain text search. So we have to escape the special regex signs time and again. Just would be interested if this is seen as an issue by others who uses the feature.

We can probably add a checkbox, and some nicer handling for the exception. Or possibly, just adding the word "regex" might solve this from the "user expectation" side.

Thanks for the clarification. I still think we should tackle the exception.

The word "pattern" was meant to imply "regex pattern".
Specifically - the ability to search for a regex is much more useful then a plain-text only search (and costs us basically nothing).

Thanks, it works