retroflexzy

A significant part of my job, unfortunately, is helping people fix their workspaces when Perforce (p4) goes bad, or creating guardrails and wrappers to stop Perforce doing bad things.

In fairness, p4 predates most of the VCSes we consider "modern", so I empathize with a lot of the underlying architecture decisions. However, it has and continues to utterly fail at improving at a reasonable pace.

For example:

  - p4 tracks file metadata of client workspaces on the server (sync'ed locally, opened for edit, file revision, etc) and uses this as the basis to avoid doing unneeded work.  If this becomes desync'ed, a reconcile or force sync must be used.  A reconcile can take hours, potentially days; it tries do detect file moves by default, so likely at least O(c^n) for some c>1. I have never personally seen a default reconcile operation _complete_ over any modestly large game code base, and in practice, people accumulate a litany of workarounds and scripts to fix this for themselves.

  - Scripting p4 is a nightmare. Documentation is poor, schemas do not exist, and all the language-specific libraries are just thin wrappers over its C++ API.

  - By default, p4 "helps" you with text files by "correcting" line endings on sync or even converting between encodings. This works until you have a mixed-OS environment, and discover a part of the pipechain that _must_ have a certain style.  There are various levers to pull to make this better, but I've yet to find something fool proof.

  - By default, p4 keeps flies read-only, only unlocking them when explicitly marked as being edited.  This means, to avoid having to do this manually, every tool you use needs to be p4-aware.  Or, you can turn this off, and choose to contend reconcile instead. (See above)

  - Branching a modest game project, with, say, Unreal source code, can take hours.  And this is the quick version where you ask the server to simply create new metadata, with no file transfer to a client.

  - p4 is licensed by the user-account. Every user entity in p4 not intended exclusively for performing backups and maintenance operations counts toward this, including users required to integrate with other services.  Plus, often times, these integration users must have admin access to be useful. The security posture is horrific.

Back in my youth, after the Internet became common but before Wikipedia, I tried to discover the answer to this and came away disappointed again and again. Every article I could find simply stated "because light scattering", and barely much more.

How does scattering work? Why does light scatter? _What does scattering even mean in the context of light?_

After C:, it really is just allocated in order.

Between CD/DVD drives, writers, Zip Drives, and extra hard drives, it wasn't unusual for a workstation to naturally end up with G: or H:, before mapped network storage became common.

> Ansible is a complete mess. You're better off managing things with Puppet or Salt, as that gives you an actual declarative mechanism

We thought this, too, when choosing Salt over Ansible, but that was a complete disaster.

Ansible is definitely designed to operate at a lower abstraction level, but modules that behave like desired state declarations actually work very well. And creating your own modules turned out to be at least an order of magnitude easier than in Salt.

We do use Ansible to manage containers via podman-systemd, but slightly hampered by Ubuntu not shipping with podman 5. It's... fine?

Our mixed Windows, Linux VM and Linux bare metal deployment scenario is likely fairly niche, but Ansible is really the only tenable solution.

It was "fun" discovering this the hard way a number of years ago when active US Android user count for a game we were supporting dropped 15% essentially overnight. The TCP stack in the client only did IPv4.

The challenge, ironically, was convincing management that adding IPv6 was the thing worth trying. After almost a week of getting nowhere (and almost 2 weeks of outage), I forced the issue by saying "Look, I'm doing this. I need one engineer for 2 days. If it doesn't work, then it doesn't work."

He got the change implemented in 2 hours. QA OKed it the next day. The topic never came up again.

I once attempted to use XSLT to transform SOAP requests generated by our system so the providers' implementation would accept them. This included having to sufficiently grok XSD, WSDL el at to figure out what part of the chain is broken.

At the end of the (very long) process, I just hard-coded the reference request XML given by the particularly problematic endpoints, put some regex replacements behind it, and called it a day.

Or, at least, not giving a fair take on its particular approach to open source.

We both self-host and pay for the service. There is ample engagement from the development team with the larger community. There are also a myriad of open source projects without the same licence restrictions that Sentry-the-company publishes or maintains which make up key functionalities in Sentry-the-product.

There was a period in World Rally Championship history when the top drivers would manipulate the starting order for the following day's stages by intentionally slowing down before the end of the stage. It was bizarre to watch teams intentionally give up 10+ second margins when stage wins can come down to half-second gaps.

There is a cursor rendering fix in xf86-video-radeonhd (or perhaps -radeon) that flips a single bit.

It took the group several years to narrow in on.