Version history for a program:

Version history for a program:
>2.1
>2.2
>2.3
>...
>2.9
>2.10
>2.11

And the worst part is when the installers are listed as
>2.1
>2.10
>2.11
>2.2
>2.3

Other urls found in this thread:

download.gimp.org/mirror/pub/gimp/v2.8/windows/
download.gimp.org/mirror/pub/gimp/v2.8/windows/?C=M;O=D
semver.org
ftp.mozilla.org/pub/firefox/releases/
twitter.com/AnonBabble

wait i don't understand either thing you're having a fit about.

You are the problem then.

Example here:

download.gimp.org/mirror/pub/gimp/v2.8/windows/

Try to find the latest version on first glance.

what's your issue with the version history? that it's going in chronological order? that it iterates at n.x and not like n.x.y.z? that it doesn't just increment n?

and the second issue, assuming you're complaining that 2.10 shows up before 2.2, is just a sorting problem. you can pass a flag to ls to sort it correctly.

i'm sure i can sympathize with your autism, but i'm not a mind reader, so you just posting a lot of greentext and then stomping around isn't going to get your gripe across.

you being a windows user really says everything.

the blue text usually means it's a link, and you can click on it to sort the packages by things like date.

download.gimp.org/mirror/pub/gimp/v2.8/windows/?C=M;O=D

my ~first glance guess~ is that the package you want is the first .exe file listed there (or if you want the torrent, then the torrent).

... or they could use a proper, non-retarded version numbering scheme so people don't have to play bingo with the ordering to get the results in actual order.

So you are mentally retarded?

what the fuck are you talking about with this bingo reference? you sort by date. older versions shouldn't be getting touched anyway.

if you used even a rudimentary package manager this literally wouldn't be your problem. and yet you've insisted on making it your problem while paradoxically insisting on being a total retard about dealing with it.

Wait, are you having trouble understanding 11>2??

Except that apparently 2>11

and this is evidently causing you to overlook the fact that 16 June > 05 June > 07 April > etc...

1.0
1.01
1.02
1.03
...
1.09
1.10
1.11

>the only correct way of numbering versions

i honestly thought this was about semver. OP, be less vague and niggardly

Wow, that sure was hard.

Year.Month is the best version naming, ubuntu does it right

Wrong m8

Wrong.
Git commit hashes are the best.

The latest version is gimp-2.8.16-setup-6.exe from 16-Jun-2016 you gigantic fucking retard.

gimp-2.8.16-setup-6.exe is not the same as gimp-2.8.16-setup.exe. There are 7 different versions of 2.8.16:
>gimp-2.8.16-setup.exe
>gimp-2.8.16-setup-1.exe
>gimp-2.8.16-setup-2.exe
>gimp-2.8.16-setup-3.exe
>gimp-2.8.16-setup-4.exe
>gimp-2.8.16-setup-5.exe
>gimp-2.8.16-setup-6.exe

This just further proves that this numbering scheme is beyond retarded.

commit hashes don't indicate rank order. ultimately, you need a rank order to make it clear what snapshot people should get.

commit hashes serve an important, but totally different purpose. do you need me to find an explanation of the purpose/value of good versioning practices?

there's nothing wrong with versioning like this, because "version sort" exists, and there's also something similar for stuff like pages in .cbz files (though it's better for those to have leading zeroes so you don't get boned when you use a reader that doesn't support it).

Welcome to Mac OS!
System Software 1
...
System Software 7
Mac OS 7.5
Mac OS 8
Mac OS 9
Mac OS 10 Server
Mac OS 10.0
...
Mac OS 10.9
Mac OS 10.10
Mac OS 10.11
Mac OS 10.12

remember he uses windows, so he doesn't know anything bash related (so the chances of him having been exposed to version sort are low)

> let me just see if that feature is in any of the remotes

develop
master
master_2
release 1.014.13
fix_bugs_branch
terrys-fork
idea
ideaWithExamples
task-091
task-092
task-093
story-001
task/story/001
task/story/001
branch-001-sorryGuysJIRABroke
midnightLizardSquad
Master_2
story-Core_087
story-Core_088
Story-Core087
Story-Croe087
release_0.1.1.15.01.14

2.2, to 2.9 > 2.10 and 2.11
Instead of making it a 2.10 they should just name it 3.0 sine they're incrementing by .1. Or they could just increment it by .05 or less instead. That's the general rule of version history, it's shown in decimal numbers. Or, if you're going with 10+ just call them v2 build10. It has nothing to do with sorting if the site isn't built by morons, it's just how version history has always been shown except by hipster developers.

1.2 > 1.11 , that's how numbers work.

This is the best solution honestly

Semantic versioning you idiot. Version numbers != decimals

It's always been that way, at least on the majority of software. They were always decimals, stop being retarded.

This.

OP is an idiot

ITT: Lincucks so dumb at math they not only *HAVE* to sort files by date for chronological order but even go on to defend this behaviour

YOLD: 2000 and never.

Why the fuck would you ever download gimp

No PS from the past 5 years runs under wine and even if it did I can't afford the monthly fee.

>They were always decimals, stop being retarded.
What? No, they weren't. Versioning schemes aren't universal

Read this semver.org

How the fuck are decimals supposed to work with versions like 1.2.3? Or with something like 1.2.3-alpha-3?

2.11 > 2.2 because 2.11 is the eleventh minor update as compared to 2.2 being the second minor update for the second major update.

>And the worst part is when the installers are listed as
.1
.10
.11
.2
.3
That's nix based operating systems for you.

>Windows 7
>Windows 8
>Windows 8.1
>Windows 10
:^)

They have always worked like decimals, at least on relevant programs. I've never seen 2.9 to become a 2.10 instead of 3.0, 2.9a or 2.9.1, until recently. Which makes me think you're all underage.

There's plenty of relevant programs that go from 2.9 -> 2.10

>Windows 3.11


>until recently
>apache_1.3.11.tar.gz
>2000-01-22

Maybe math just isn't for you.

Can you state which programs do this? If they do, then the devs are either autistic enough to not release any more minor updates after x.9 or they don't follow semver. Not following semver should be a crime.

ftp.mozilla.org/pub/firefox/releases/

Seems these people don't know about nix based operating systems.

>They have always worked like decimals
They have almost never worked like decimals.
The most common past procedure was
..

But . were always a number, and the was just a separate number which pointed to bugfixes and minor patches. In op's case the program is numbered incorrectly or isn't updated often so it only has the major release version plus patch version.

>0.1
>...
>1.0
>10.0
>11.0
>...
>19.0
>2.0
>20.1
>...
>29.0
>3.0
REEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEE

>In op's case the program is numbered incorrectly
There is no incorrect method for versioning, as long as you are consistent in it.

True, but some method are superior in many ways.

>Instead of making it a 2.10 they should just name it 3.0 sine they're incrementing by .1. Or they could just increment it by .05 or less instead. That's the general rule of version history, it's shown in decimal numbers.

No it isn't, retard. The dot isn't a decimal symbol, it's to separate major and minor versions. 2 is the minor version, .10 signifies that it's the 10th update to that version.

This

Let me guess, you one of those people who refer to the first episode of a TV show as "S1E1"

>implying too many minor changes don't make a big change
Also read only programs which don't get updated often use more than 9 numbers after the first dot. And why would there be multiple different methods of versioning? It makes no sense.

Only on the situation.

If your application has open APIs then using a versioning that will clearly display when changes to the API take place is a definite benefit, but if your application is completely closed then it means fucking nothing.

Actually I refer to it as 1.1 if I have to make a specific choice about the labeling.

>>implying too many minor changes don't make a big change
They don't.
They literally don't.

>And why would there be multiple different methods of versioning? It makes no sense.
Because some people are stupid/selfish/ignorant and refuse to be standards compliant.

>implying

respectfully, kys

Stating, actually.

There is no set standard for versioning.

There are a number of standards to choose from, none of them are more correct than others.
Much like you choose a software license based on what you are making, you pick a versioning scheme.

It depends on how you define a minor change, but you still don't have to completely reinvent the program to progress from 1.x to 2.0

No, but you have to do something that you consider to be a major change, like adding a major new feature or doing a considerable code rewrite.
There's no point incrementing the major version just because you have done nine minor releases.

Why would anyone even care about that? It's a version "number", it doesn't matter if they increment it. Unless it's a major update which changed over 20% of the program, then they should always increment the major version.

ok

another

lmao, this thread is why Apple is a billion dollar company. People are too thick to understand why end users don't like to see lists ordered as 1, 10, 11, 12, 2, 3, 4.

>b-b-but you disagree with me! That means you have autism!

people are too thick to understand how to sort through a goddamned list correctly, especially when there's something to sort by last modification date

Those repos are just AppStore for autists

>especially when there's something to sort by last modification date
>rebuild an installer for an older release
>its now out of order in both modification date and filename

>>rebuild an installer for an older release
why the fuck would you do this ever? and if this is the case, you can just look like two entries down and see the one with the larger number

lol, OS X 10.10 after 10.9 ring a bell?

it's the same concept, yeah

>especially when there's something to sort by last modification date
You can't always sort by date, sometimes it just fucks up the shit even more.

>why the fuck would you do this ever?
To add malware to it of course.

in this case, you can

[shitposts]

>>especially when there's something to sort by last modification date

>download all the installers on my computer
>now modification date is the same for all the files

End users should never have to see this list.

This isn't a versioning problem this is a user experience problem. Just link to the fucking latest version on the download page users don't give a shit about versions.

>in this case, you can
I'm not talking about this case you dense fuck, it's the general case.

>downloading all the installers in the first place
asked for it
I am talking about this case, you retarded shithead

My packaging throws an error whenever I do

1.01

It wants 1.1

so I iwant to do 1.15

1.1 > 1.15

Gay huh?

There are times when you may want to rebuild old installers, such as old advertising no longer being relevant, removal of advertising, your installer may download support libraries that need to be updated but you aren't updating the application itself so there's no reason to chance version numbers.

The fact that you need to use the Modified sort rather than the OS just fucking sorting them correctly is just proof of the failing of the OS.

I know. If I made a project I'm going to release it as a UTC timestamp

>not compiling straight from the master branch
stay pleb

this can't be "summer", you can't just get this fucking dumb just for a season

i refuse to believe a person can be this retarded, this must be a troll post

xD

It's s1e01 unless the show has run over 10 seasons then it's s01e01

And how do you know it won't get to 10 seasons? And when it does, you update every instance of s1e01 to s01e01, since it breaks every ordering?

>It's s1e01

No it's not. And what for show with less than 10 episodes it's s1e1? Pleb.

You dont. You reorder after you hit season 10. That's the entire reason you use a naming scheme. Replace s1 with s01. So hard what amaze.

People like you caused the Y2K bug and IPv4 addresses running out and the imminent Y2038 bug.