Is it possible to check if anything is relying on this not working with negative numbers?

Well, I see a tremendous critism and scepticism concerning modern client side techniques here. The world of frontend tooling has changed tremendously in the last 5 years. And I understand concerns, but nobody has to write JavaScript for certain browsers anymore. Javelin for example tries to solve so many problems that Babel solves with one config entry. But that's okay and I like the architecture here and know we can't change everything at once.

Ok, I don't understand the specifics of css-vars, but if they're better, they're better.

The overwriting of variables with media queries could be configured inside the theming class. So possible breakpoints can be part of a theming API.

And my general approach is: when the client can do something (and that even better), why should the server do it?

@avivey CSS vars are a vital part of modern modularized theming. They are versatile, can be overwritten with media queries or per class/element basis. They are editable in browser dev tools. If you would ask me: this step is pretty important.

In T15736#15726, @avivey wrote:

to have a working dev server when developing js - at the moment I have to ./bin/celerity map after every change to js and css to see my changes...

Are you sure about that? I thought celerity map was only needed when adding a file (in dev mode).

also, adding this file in src/extensions/ will let you have css/js files from src/extensions/rsrc/ loaded automatically:

to have a working dev server when developing js - at the moment I have to ./bin/celerity map after every change to js and css to see my changes...

Looks to me that the existing system is better at providing Themes? It allows real code, inheritance, etc. to set variable values.
That is, I think Extension Themes are better off writing PHP Post-Processor rather then CSS vars.

In T15039#1358, @bekay wrote:

The clone repo modal could really use some copy buttons.

Preferably replacing the Read / Write Button.

I don't know if that is part of this task, but the global typeahead search should hide all disabled users, wikis and repos, shouldn't it? This is nothing you want as a fast suggestion...

I've better triaged my root problem.

So maybe this is just a matter of parsing that commit part "Depends-on:" since everything else seem already implemented to me. Partially related quiz in T15738

Just as a reminder for me: Using a css variable for the height of the tags would be best. I could be overwritten at any point.

Well, my idea would be:

• Updates map

• Fixes some js linting warnings

@aklapper Yeah, that would be great. Sorry, I don't have a working dev instance right now.

@deric: Hi, would you like to arc land your patch, or do you need any help? Thanks in advance!

For the records, adding public function getValueForQuery($value) is needed because otherwise Expected a numeric scalar or null for %d conversion. Query: project.projectDepth >= %d

Would anyone fancy reviewing this? :)

whoops

Make code more readable

In D25525#15322, @valerio.bozzolan wrote:

Maybe simplify if (empty($v)) { to just if ($v) { and swapping the case

In D25512#15325, @valerio.bozzolan wrote:

@sirocyl Does this fix your issue? if yes, I can help in landing

Heck yeah, changing phutil_nonempty_string() to phutil_nonempty_stringlike() there fixed the issue I had mentioned in a comment in T15737: Include information for installing required PHP version in Diviner docs.

@sirocyl Does this fix your issue? if yes, I can help in landing

On it, thank you!

My friend, I think you are affected by this change, and this applies to whatever PHP version

I super-like the result of this in /project/manage/1/, so approved.

I'm on the latest pull, commit f81e821abf2b275d1778d02a018a4d0af7208be8.
A bit off topic here :) but I'll quickly sum up the issue I'm having (and am in the middle of trying things to fix):

• Visiting https://[phorge instance]/auth/config/view/2/, where my Github auth item is, makes nginx dump a blank 502 page, no further info.
• First instinct is to check the /var/log/nginx/error.log, where I get a PHP stack trace concatenated into one line, and an nginx header length error (reformatted for readability. the PHP error is truncated):

Oh, totally possible that I forgot to update it :)

(@speck red light)

Whops, forgot to approve this. Sorry! And thanks.

I'll take your word on it that it's supported, though the current installation guide says otherwise.
Further, I had a few roadbumps on a fresh install to a new server with PHP 8, that I suspect may relate to strlen(NULL) behavior, but I did need to set up PHP 7 to make sure that was or wasn't the case.
(I'm in the process of reproducing those and working out a possible cause or trace. Some manifested as warning banners, one as a 502 on my end in the GitHub auth provider.)

We already support PHP 8...

Agreed. We should not link fonts from outside our own assets (those hosted on an instance of Phorge.) I don't think I suggested that, but I don't disagree.

maybe include "allow extensions to provide themes" here...

We should pack the fonts internally, not use an external source.
We pack all the assets we use, for cases where network access is limited by infosec types or by other reasons.
Just serving few font files from the site would be much simpler, and the extra bandwidth requirements are miniscule anyway.

As far as bold/italic variants are concerned, for fonts that support it, you can preferentially serve the "variable" version and only fall back on specific font variants when that isn't supported, to save on bandwidth in most cases - since the e.g., Inter Variable font file "includes" the bold and italic, condensed and expanded versions, essentially "for free", as they're generated by the font engine/renderer on the fly from the base font.

In T15630#15618, @aklapper wrote:

For consideration of the bigger picture, I'd like to mention https://collinmbarrett.com/block-web-fonts (performance etc) and for example Firefox 118+ blocking font fingerprinting in private windows (yes, private only, but I can vaguely imagine expansion).

Inter and Noto look best to me (1080p laptop screen, Linux, Firefox, no scaling) given all default sizings in Phorge (as a drop-in replacement). Noto, I believe has a lot more Unicode symbol coverage, and a larger file size as a result; but Inter has variable font properties (e.g., more weights than "bold" and "regular", as well as condensed and expanded forms, letter variants for differentiating I/l O/0, etc).

For consideration of the bigger picture, I'd like to mention https://collinmbarrett.com/block-web-fonts (performance etc) and for example Firefox 118+ blocking font fingerprinting in private windows (yes, private only, but I can vaguely imagine expansion).

Broader CSS discussion: T15734

A mermaid-js one could be written based on the old graphviz one:

In T15673#15587, @valerio.bozzolan wrote:

TRUST ME - YOU DON'T WANT TO LISTEN MY ENGLISH :D :D

Btw now the video is here :D :D AGAIN SORRY PHORGE FOR MY MISTAKES <3 <3