Suzzer start casually tossing around terms like containerization into random conversations.

It wasn't explicitly stated in the articles but I assume the approach could hybridized? Say your monolith has a piece that would benefit from being turned into a microservice and you dedicate the time and resources to turn that into a microservice?

I guess reading the articles again there probably doesn't exist a scenario where that really makes sense. Your entire application either needs the scalability and compartmentalization that the microservice architecture provides or you don't.

Sent from my SM-G900R4 using Tapatalk

It took us about two years to go from a monolithic to a microservices architecture and then a while to dockerize everything. It doesn't sound like it was near as big as Suzzers but we were also a much smaller shop...

It took my old group about 2 years to go from old weblogic/central-server - all environments owned by Operations team to node/scala back end with dev/test hosted in rackspace, automated (partial) CI/CD, etc.

That group is working on containers now and another group has a full blown shared API builder platform where each end point is a microservice with it's own repo - and pretty much anyone can spin one up.

The group I officially belong to however, has been doing POCs in microservices and containers for a year or two with nothing to show for it. Like I said - microservices was literally just a word to my boss, that he got funding for to do POCs, until a month ago.

I'm angling to get back in my old group.

We're 6+ years going on monolith -> microservice and it's still hybrid with a lot of the business logic and shared tables still stuck in the monolith.

My home wifi is driving me nuts. Every 10-15 minutes it just dies. To get it to work I have to turn off wifi on my mac, then turn it back on again. It always starts right up again when I do that. I think waiting X might also work - where X is like a minute or longer.

I called TWC cable - they had me try a different channel, which didn't help. I tried putting the channel on auto. No help. I tried fiddling with some of the other router settings that sounded like they might help - like Band Steering. Didn't help.

Any ideas? Yes I've unplugged and plugged in the modem. The only thing I can think of left is to buy a new router (I bought one, not renting theirs).

Oh yeah - also the fun thing is since fiddling with the router - the 5G network doesn't work 90% of the time. It's either not visible or I can't connect when it is. I even set all the wireless router stuff back to factory default.

Try other devices. If you have to turn off/on again your Mac, the problem is probably your Mac and not your router.

Do you have the dongles to wire directly from the modem to a computer and see if it keeps going out? Could be the router is dying.

I have two MacBook Pros (one for each job). It happens on both of them. I should have mentioned that.

Another dumb question for you git masters. Let's say I just committed a bunch of files and changed my mind. All I want is a big undo button that puts reverts my latest commit, and leaves the changes in an uncommitted state - just like it was before I pushed the commit button.

For some reason every time I try to revert or reverse I lose **** and I don't have files in an uncommitted state like I want.

Yep. just reversed commit. But those changes I committed are lost in the wind. Luckily I have sublime open and can just hit Ctrl-Z to get back to where I was. I found this but it seems painful: http://stackoverflow.com/questions/9...commits-in-git

Once cool think I just found is you can go back to previous commits in Sourcetree and reverse hunks or lines. In that way I was able to get back one of the files I un-committed, which wasn't open in Sublime.

Speaking of sublime I finally moved to VSC after ~6 years of sublime. Its um not bad. And obviously actually under active development and open source and blah blah meow chow.

Yep - biggest thing it doesn't have it Ctrl-Z when something else changes the file you're looking at. I lost a half-day's work on that one.

`git reflog` shows changes to your index (there's probably a more technical way of putting it, I'm not quite a git master) along with the commit hashes. So if you wanted to get back to a previous state after an "oh ****" moment (in this case, losing files you accidentally added then reverted), that would show you the hash of the commit you reverted to so you could check that out again.

Personally, I think using git in a command line is vastly superior to any GUIs. For one, everything you do is logged right there in the terminal window so it's trivially easy to reset to a previous state when all state changes you've made are documented as you commit them.

There's a few things I show git command line guys on SourceTree that they're look "ooh, no I can't do that". I'm a big fan of the diff tool.



Using this I can discard any changes at the hunk or line level. Stuff like console.logs I just put in to test. Also I can discard a whole file.



Mostly just I'm old and don't like having to remember new things. Gimme a context menu to pick from please.

git reset HEAD~1

Nice I always forget about that shortcut. I end up doing git log and copy the second to last sha. Then it's just git reset sha.

Line-by-line commits/reverts is definitely worth using a GUI for. But for everything else I use the command line.

Sweet! Now that I know it works, after some playing around - looks like I can do something similar in Source Tree using reset (I kept trying to reverse, which never works right).



Still not sure the difference between mixed reset and soft reset - which both preserve my file changes and leave them in an uncommitted state.

Stack overflow to the rescue:



http://stackoverflow.com/questions/3...-soft-and-hard

git add -p

git checkout -- filename

Maybe I am a git wizard, I didn't think this stuff was that hard.

I use `gitx` for committing and its way easier to pick the files / hunks / single lines I want to commit than any of the command line tools.

Dude half of my coworkers use notepad++ to update hosts files....so yeah

So what does reverse commit do?

I believe it figures out what changes are needed to make it like a commit never happened and then applies then as a new commit.

So it reverses a commit not by changing history but by making a change that 'negates' the commit being reversed.

Edit: So doing a reset is the same as modifying the history of your repo. That's what makes it dangerous when dealing with changes that have already been pushed to other people.