Just gotta LOVE git!

Yet another example for git "awesomeness" (yes, compared to svn, of course): PR 254178. To solve this major fuckup, I needed to sync my changes between my builder machine and my testbuilder (running -CURRENT and having poudriere jails for all supported versions). AND I needed to rebase and squash commits. Do that with svn, and have a lot of fun 😈
 
Oh my. Is awk even a programming language? IDK, I never learned it (shame on me!).

About base, if there's a pressing need to have a suitably-licensed replacement for svnlite(1), GIT is (kind of) well-understood, so just reimplement it? Or isn't that what devel/got already did? IDK, never tried that either…

Or maybe, base doesn't even need a full GIT implementation? After all, what the user might need is just getting stuff from a GIT repo. net/gitup would satisfy that need. Then, is it a problem when contributors install a GPL-licensed software from ports? 🤔

The aftermath of the GIT transition probably still needs some discussion. But I rest my case, GIT improved the workflow for contributors A LOT!
 
I don't care too much about being politically correct in UNIX commands... /bin/kill, git blame, typos in /bin/fsck... :p The guy who invented Git in the first place, he had to take a sensitivity training class! For me, things have to work, and logic has to add up to something sensible.
misc/thefuck

Code:
====> Compressing man pages (compress-man)
===>  Installing for thefuck-3.31
===>  Checking if thefuck is already installed
===>   Registering installation for thefuck-3.31
Installing thefuck-3.31...
 
In all honesty, I like the idea of using base utils to fully re-implement Git (That would solve the dependency issues, and clear the way for Git to be included in base).
 
In all honesty, I like the idea of using base utils to fully re-implement Git (That would solve the dependency issues, and clear the way for Git to be included in base).
Still you have to ask whether that's really needed in base. Base needs to be self-contained, no question about that. So there has to be a way to (at least) fetch ports. If portsnap(8) is indeed going away some day, this will leave a hole. But then, net/gitup would fill it.

Reimplementing all the GIT functionality using a really permissive license (so, no GPL crap) would be great, of course. I'm just asking, is it really necessary?
 
Still you have to ask whether that's really needed in base. Base needs to be self-contained, no question about that. So there has to be a way to (at least) fetch ports. If portsnap(8) is indeed going away some day, this will leave a hole. But then, net/gitup would fill it.

Reimplementing all the GIT functionality using a really permissive license (so, no GPL crap) would be great, of course. I'm just asking, is it really necessary?
Well, I'm gonna watch from the sidelines. If somebody can pull something like that off - that would be VERY interesting. :)
 
Still you have to ask whether that's really needed in base. Base needs to be self-contained, no question about that. So there has to be a way to (at least) fetch ports. If portsnap(8) is indeed going away some day, this will leave a hole. But then, net/gitup would fill it.
gitup is good and does the job, but I can still notice that it is much slower than portsnap.
 
is that because gitup pulls in more data than portsnap? 🤔
I think the scanning of ports tree takes more time.

Or actually (I think) portsnap does not scan the ports tree and brings in just the change file, but gitup does the scan and that takes time.
 
Can't tell if you're joking. Yes, Awk is a programming language, and a great one at that. Your loss if you haven't learnt it.
Only partially joking. Yes, I know it is a programming language. And I refuse to learn it. I know enough of these (C, C++, C#, Java, Javascript, Perl, Tcl, (bourne) shell, PHP, 6502 ASM, CBM BASIC, Pascal, Object Pascal, oh well …) to get where I need. Still, awk is obviously a "special purpose" scripting language, so implementing GIT in awk seems like a "because you can" project 😈
 
Only partially joking. Yes, I know it is a programming language. And I refuse to learn it. I know enough of these (C, C++, C#, Java, Javascript, Perl, Tcl, (bourne) shell, PHP, 6502 ASM, oh well …) to get where I need. Still, awk is obviously a "special purpose" scripting language, so implementing GIT in awk seems like a "because you can" project 😈
I haven't really 'learned' awk either... I copy-paste useful awk commands off StackOverflow and KlaraSystems as I see fit. I sometimes do study the snippets just to see how I can adapt it to fit my needs, but I'm not gonna sit down and devote time to properly learning awk so that I can use it off the top of my head for anything. There's other stuff that interests me, and that's where my time goes.
 
And now, I finally updated my ports repo on Github to be a mirror of my local repo on my builder machine 🥳

Back when FreeBSD used SVN, it was "easier" to have the ports I was working on isolated in a git repo, but now, this isn't necessary any more – just publishing my local branch that's continuously rebased!
 
Back
Top