Follow

@sir has released new results on forgeperf.org/.

A great shout-out to user mappu04 who identified the roadblock that made us look bad in previous run!

The worst-case results for huge repos still a known open issue, potentially resolved by caching in coming gitea release.

NOW OUR CALL FOR CONTRIBUTION: If you agree with us that accessibility is important, and would like to help resolving the accessibility issues tagged in the report: please have a look and consider a pull request!

@sir thank you, but literally all credits to our users ;)

@sandro @sir Codeberg is running gitea as user-facing web frontend indeed.

However, we do not see noteworthy performance-issues on normally sized repos (both gogs+gitea have still known issues for huge repos and can take up to 8sec or so for example to render linux kernel repo views).

The now enabled gzip-content compression should improve the situation for users with seriously constrained bandwidth, those are overweight in the lighthouse measures, thus the huge jump in the metrics.

@codeberg @sir How did they calculate that Matrix? That must be nonsense. No way that Bitbucket has better accessibility than Codeberg. Opening a page there leaves me with a half-empty page and tons of dead (Javascript) links which probably work if I'd enable 5+ 3rd party sites in uMatrix (see # of reqs+size; and no, thanks). On Codeberg, I don't need to enable any Javascript at all – *that's* what I call "accessibility pre-condition" (how shall a blind person figure that on BB – and keep privacy intact?)

@IzzyOnDroid @codeberg these results are generated from lighthouse, and there are details on the methodology and links to the source code on this page.

@IzzyOnDroid @sir Yeah, such automated tests are always a bit problematic, we agree. As far we have seen the tests are using the Chromium lighthouse measure (part of the developer console), details are described on the forgeperf page.

Despite our initial doubts we have to admit tho that the first problem reported (missing content compression) was actually a real issue we could easily fix, cutting down bandwidth significantly.

@codeberg @IzzyOnDroid you didn't like the results until you looked good on them 😉

I am rather confident in the methodology, no one has brought up meaningful problems with it yet. I just see the same surface-level acusations, over and over and over and over...

@sir @IzzyOnDroid Fair. And still sceptical with regard to usability issues.

On the other hand, that the report finally led to discover #176 caused us to reconsider our opinion and this deserves a shout-out! ;)

@codeberg @IzzyOnDroid yeah, the accessibility scores are based on a specific set of criteria which doesn't cover all accessibility concerns and doesn't cover many usability concerns (note that accessibility and usability are different things)

@sir @codeberg Well, what I was playing at is: there's accessibility – and there's accessibility. Both are about ability, but different kinds of: physical ability (and here the stats may have a point, admitted) – and, in this context, freedom of accessibilty with privacy kept. I generally avoid pages that don't work without all kind of JS and "other stuff". – and here BB fails miserably. So basically, the table lacks at least one column 😉

@IzzyOnDroid @codeberg well, all of the accessibility criteria which factor into this score are worthwhile goals: everyone should strive for a score of 100 in that column. Then they need to keep working on accessibility in a more self-directed manner, with feedback from the community. So I think having the stat here is worthwhile, even if it's not the full picture.

@IzzyOnDroid @codeberg and if you want to convince me that javascript is bad then we are perhaps enduring the single most egregious case of "preaching to the choir" ever seen 🙂

@sir @codeberg Now, what about usability? Are there any auto-tests for? Guess that would make BB drop out at the bottom 🤣

@codeberg @sir also for me codeberg is really damn slow. I have no idea why but I think something on my end it broken. It takes around 3k requests and 33s to load the front page with me logged in.

@codeberg @sir my bad. to show content it takes 1.2 minutes. I use umatrix but nothing it blocked, logger is non suspicious and console no errors.

@sandro @sir Can you provide more details about location and connectivity, potentially via email if it contains sensitive details?

@sandro @sir That's more than weird. Usually should not take more than a second.

Can you provide more details?

@laufi @codeberg @sir Fortunately, Lighthouse results aren't anywhere near descriptive as to what is wrong with the site (and it is still possible to get a 100% a11y score and have a completely inaccessible site) - I'd appreciate personal 1to1 guidance for a contribution as I'm not at all familiar with Codeberg's project structure, building process, etc.

@nwng @laufi @sir please don't hesitate to contact us with all questions you encounter!

@codeberg @laufi @sir Oh, I didn't realize you're actually using Gitea. Gitea doesn't have many open issues on accessibility at the moment and those open seem to be abandoned by the maintainers - they're also marked as proposals and not as bugs/breaking issues - what they actually are. The issues too have persisted over a very long time in the software and I reckon that a rewrite of the entire UI with a11y focus would be much easier than to implement this now.

Sign in to participate in the conversation
Mastodon for Tech Folks

The social network of the future: No ads, no corporate surveillance, ethical design, and decentralization! Own your data with Mastodon!