@Móż I assume that the code is in the time machine, waiting for when "they" undo their undo "decision".

But look on the bright side: it's a negative bug day :-)

Except for those bugs where the code wasn't chopped correctly!

@andy256 our git spiderweb is a thing of beauty and/or terror, depending on your feelings about branching and merging. I expect it would drive an SVN user to drink. It has everything from the Raspberry Pi fork to a bunch of experimental "boss brainfart" false starts, as well as the usual "five parallel development branches" because I only rarely have one new thing under development. Oh, and a bunch of rust and SQL code.

Plus there's a separate scratch directory full of little programs that test various things. Including one that tries to reproduce a problem another zlog user is having where his threads sometimes hang until a log rotate (which I can't reproduce). Etc

I did a bit of a "works on my machine" rejection to an edit, someone wanted the https images turned to http. I flagged as "damage" so I could leave a comment (can't do that for 'no improvement')

hm?

By WOMM I mean "I tried three browsers on Windows, two on Centos"... I realise that both options work for me, but I don't think dropping https is worth doing one page at a time. It should be site-wide, one way or the other.

@Batman there's a pending edit. It changes the images from https to http.

its working fine on chrome and safari on mac

@Batman hence my saying it's no improvement. It fixes one page, for one user. Hopefully.

its likely some extension they're running

@Batman which even more suggests that the solution is to fix their end.

@andy256 I was encouraged to look at git branching models... I thought my approach was a standard branch-and-pull strategy by now I'm confused about what other people do. My "merges" are 99% actually pulls, and only days like this one do I pull a branch wholesale, and the mega-commit that comes out will amount to "these 10 affected files... they are completely different from what went in".

Coz I'm deleting half the changes from branch A and applying them to branch B, deleting anything that conflicts from branch B. When that's done I will pull the result into the test branch, and after that it's all the usual process.

@Móż Now is not the time to play with a new Branch / Merge model.

Repository structures are inherently tricky to grasp. Since you're flying well with a model, stick with it.

Work on a new model as a specific project, with clear evaluation criteria.

@Móż I can still see the slow blinking look I got from a user (Boss) who complained about sth not working on his browser. I told him I'd tested his reported scenario on 6 versions of IE, 4 releases of FF, the current release of Safari, and it worked on them all; would he like to demonstrate the problem to me? "Ugh, no I'll try it again". Damn right you will.

@andy256 it was more "what jargon should I use to describe it", coz it would take another developer using my codebase to even get me to think about changing. Well, or someone like you saying "this model, because your model has these issues".

@andy256 I am currently still laughing about the car dealer who apparently sincerely told my boss that the only way to make his new iPhone work with an Audi was to buy the latest model of the car he has (a luxury sedan). I suggested getting a Tesla S instead. But he has decided to upgrade to Android. In shocking news, it pairs with the car using bluetooth with no problems, and it all works. Importing was also simple.

Voice commands are fine, he is just wibbling about the touch interface being different in places.

Just "buy a new car" as a solution? WTF?

@Criggie Ppl do it all the time. Genetic survival.

heh probably not purely because the older siblings are crook.

not intentionally, anyway

Wow. Nine different people have been doing review lately, and the review queues are all empty.

First time I've seen that since SE changed the IU to show it.

Nice