Sublime Text 3 Updated
sublimetext.comI love this editor and it makes me sad to see how the project is stalled.
In 2002 the investors behind Blender launched the "Free Blender" campaign. They asked for 100,000 EUR as a one-time fee for open sourcing it. At the time the project was dying as a proprietary product. The investors got the money and today Blender is a healthy open source project.
I would love to see something similar with Sublime Text. The author seems uninterested in continuing with its development while many users want to see it moving forward. I believe it can raise much more money than Blender at the time.
I'd like to see Jon sign up for something like Patreon: http://www.patreon.com/
As near as I can tell, they would qualify as a creator, the signup process is fast and easy, and with Sublime's massive user community, it would probably give them a nice little revenue stream on the side. Patreon seems like a good way to extract money on a regular basis from people who actively want to give it to you. One of the webcomics I follow daily, Questionable Content, is currently pulling down $9,000+/month from Patreon support alone: http://www.patreon.com/jephjacques (HUGE as webcomic revenue usually goes).
I'd pledge like $10/month just to get regular monthly software updates. Hell, I'd pay $5/month just for regular monthly blog updates on he's been working on, and there'd be absolutely zero downside to his writing a blog update once a month.
Stalled? I can't understand why people has this perception. You're not the only one; SublimeHQ forums are full of people complaining that "Sublime project is dead".
The guy releases a new version almost every month. He doesn't talk much, but keeps chugging along, fixing bugs and releasing incremental functionality every month or two.
True, it is still beta. But so what? ST3 has been incredibly stable since the first version, and all beta versions are a free upgrade to all ST2 licensed users till it is out of beta.
Plus, if you're really missing any functionality, just go ahead and develop a plugin. It's easy, and you don't need to open source the entire editor for that.
I think we should give the author a break. He built one of the best editors out there, and I really hope he can make a living out of this product, in whatever way he can.
What are you talking about? Jon hasn't released anything since the end of 2013...
A lot of ST users are former Textmate users, too. So we've certainly been through this rigamarole before.
Textmate 2 did exactly what Blender did. Went open source, has a healthy developer community, and has had steady updates for a couple of years. They still call it "Alpha" but it is very stable and I use it every day.
Wow, I somehow missed (or forgot?) that TextMate 2 got open sourced. Here is the Github repo: https://github.com/textmate/textmate
True, but before Textmate was open-sourced, there was a long period of frustration due to lack of progress and communication from the developer.
Yes, but Textmate 2, after all this time, has less options and is slower than ST2, much less ST3.
>I love this editor and it makes me sad to see how the project is stalled.
Stalled? It just had an update.
It's also in the 3rd version, under development, which came just a couple of months after the 2nd version had been released. He could just have released ST2 and keep it at that version for 3-4 years.
Instead he immediately started development ST3, which had frenetic development in the first months, and has been completely stable for a year or more (I know, I use it everyday, along with several plugins).
And in the forum he even mentioned ST4 base libs he is preparing a month or so ago.
> And in the forum he even mentioned ST4 base libs he is preparing a month or so ago.
Huh. Do you think there will be an ST4 beta/dev release before there's an ST3 stable release? I wonder if the developer has just decided to label all his releases beta from now on?
> and has been completely stable for a year or more
If this is true, why are the releases still labelled 'dev' releases?
How is anyone supposed to know that it's actually a stable release they should be using? Research it on HN comment threads?
>If this is true, why are the releases still labelled 'dev' releases?
Because it's an arbitrary label.
Or because John wants to hook something additional up or polish something before giving it his "3.0" blessing.
Or because it crashes on Linux and needs to fix that too...
But not because there's any issue with it on OS X, at least not for me, and I'm a heavy (8-10 hours per day) user.
Or because other people in this thread insist that in fact it's not stable, and there are numerous bugs on multiple platforms.
Regardless, if you want people to know that a release is ready for use by people who want a stable release, you label it a final release, right? Presumably the developer(s) do not believe it is ready for wide use. Whether they are being overly conservative or not is I suppose another question.
But it seems silly to suggest that everyone should do research on their own to discover something labelled 'beta' or 'dev' should really be considered the latest stable release.
>Or because other people in this thread insist that in fact it's not stable, and there are numerous bugs on multiple platforms.
I don't know what plugins those people use. They might have thrown in the kitchen sink of unstable plugins. Lots of people using ST3 in the forums without issues. Plus I know my personal experience on my platform of choice.
>Regardless, if you want people to know that a release is ready for use by people who want a stable release, you label it a final release, right?
Well, some projects prefer the perpetual beta designation. Heck, even Google did that for years for stuff like Gmail etc.
the dev releases are hidden on the website. You have to look to find them. The website by default directs you to the very stable 2 version. If you look you can find the stable 3 builds. If you dig deeper you can find the dev builds. Each of these things is clearly labeled and not mixed. Nobody going through the website is going to download a dev build on accident.
Please note that this is a dev build of Sublime Text 3. You can only get it by manually downloading or checking for updates using a dev build. The most recent stable build of ST3 is from December of 2013.[1]
Dev builds are alpha-quality and generally unsuitable for everyday use. The current build will change permissions on any file you save.[2]
1. http://www.sublimetext.com/3
2. http://www.sublimetext.com/forum/viewtopic.php?f=2&t=16696&s...
He says he'll fix it in the next build..
Which I assume, will be released in 6 to 36 months. Yay!
Patience is a great virtue!
While we understand sarcasm, we should give credit to Jon.
Developing a great editor such as ST is a task big teams would have a hard time pulling off, let alone a single developer who is, BTW, running all aspects of the company.
So, sit back and shut up, ST2 works great, ST3 works great. You don't need an update every day just for the sake of being updated.
hmm weird, i cannot reproduce the bug on os x. (3dev channel)
touch hello.txt; chmod 777 hello.txt; subl hello.txt; add something and save ; 777 still.
I've almost given up with Sublime Text. I have a bug with the Linux version on xubuntu which causes the application window to shudder when I scroll. It's something to do with smooth scrolling.
I can't register at their forums as it requires a keyword which you can only get by emailing keyword@sublimetext.com - I've tried three times and each time I've been ignored. I've also tried email the other ST employee, but been ignored also.
Very frustrating when I've paid for a licence.
xyzzy3
edit: before anyone gives me shit for doing this it isn't exactly a big secret, it is a pretty poor way of 'securing' the forum. Just stick a captcha on it for gods sake.
Thanks.
I agree - it's the shittest, most user-unfriendly method of securing something against spam I've ever encountered.
There is a setting that can help with the scrolling:
"scroll_speed": 0
I have it in my user preferences (on Debian Sid) and it works well for me.
Unfortunately there is a rise in spammers recently.
Have you emailed sales@sublimetext.com?
It's kind of sad that this project has become so dead that an update to a dev build is news worthy.
Not sure I subscribe to the idea that ST is "dead". The base editor is stable enough for day to day use ("dev" build be damned) and all the extra functionality (i.e. new features) is provided by plugins anyways.
ST could never see another update and would remain a competitive editor for daily use due to its plugin ecosystem.
No ST3 is not stable. I can choose between latest beta (3059) which often crash with 'plugin_host' so I have to restart, because none of the plugins works after that. Or I can choose 3062 which doesn't open sidebar tree sometimes, and have to restart multiple times... I think this update didn't fixed that, will see.
I use the latest betas (actually dev releases) for a year or so, and never had ANY crash ever.
I use several plugins: The Linter package + several linters, the GoSublime, JSON pretty printer, a couple of build setups for different languages, etc.
OS X of course, might be buggier in Linux. Then again I don't think Linux is a priority.
Perhaps it's some plugin you use? I've never had to restart, or have it crash on me.
I use it on OS X also, but I have to restart every day multiple times.
I wouldn't consider ST3 to be stable or usable under Linux since it just wrecks file permissions.
Neither do the Sublime Text developers; even the main release of Sublime Text 3 is listed as beta, and the download link on the front page of their website is still Sublime Text 2.
But do they still update the stable version of ST2? From what I gather "beta" is his way of describing the fact that he's actively developing it
The last release of Sublime Text 2 was over a year ago. And the post I was responding to was claiming Sublime Text 3 was already stable.
My point is that the "released" version has been abandoned. He's not supporting previous versions at all, you pay for access to an actual beta build, but if you happened to buy the full product you're SOL
I'm using the 3059 build and it doesn't change permissions of pre-existing files. And this build is supposed to solve that for the dev builds (I have to check it).
until something else comes along, as with Textmate
dev build is that bad ?
yes
No, it's rock solid. Except if you run Linux it seems.
Is LimeText a viable alternative?
Not yet, the back-end is almost "ready" but the front-end still requires a lot of work.
The article doesn't link directly to the list of changes which are as follows:
Improved quote auto pairing logic
Selected group is now stored in the session
Fixed a crash triggered by Goto Anything cloning views
Windows: Added command line helper, subl.exe
OSX: Added 'New Window' entry to dock menu
Posix: Using correct permissions for newly created files and folders
I don't really understand why these new releases are called 'dev' builds, and there isn't an official non-dev ST3 build yet.
Are most people using ST3 anyway, despite the releases being labelled dev builds? Are they in fact pretty stable? Are most widely used plugins updated for ST3 (and maybe no longer supported for ST2?)
Should I just go ahead and switch ST3 "dev"?
All multiple years of 'dev' releases gets you is confusion, if this is what most people are using and most plugin developers are targetting, i wish the developer would just call it a release.
There are actually non-dev builds of ST3, and those are somewhat stable and what most people use.
Dev builds:
http://www.sublimetext.com/3dev
'Stable' builds
I've been using ST3 for some time off and on, and never had any issue. I just download it from http://www.sublimetext.com/3
Sublimetext 3 beta is really stable and fast as hell. I don't know what black magic they use under osx but it is even smoother than http://chocolatapp.com which is a native cocoa based editor with nice builtin code completions for many languages.
Agreed. That's why we all love Sublime Text and why we are so pissed off that we can find the pulse.
The magic is called OpenGL and hardware accelerated graphics.
Anyone have a clue as to what were supposed to be able to do with this "Windows: Added command line helper, subl.exe". I'm guessing it's mainly for a shorter exe name to add to your path when launching sublime from the command line, but maybe I'm missing something?
It's just to launch the editor from the command-line.
IT's nothing new - sublime text 2 (at least on OS X) had the "subl" command so you could launch an edit session from the cli.. I assumed the rest had the same.
It's certainly not a new feature or anything to celebrate.
It is new on Windows.
not so much, you could always do sublime somedirectory and it will open it. It appears the novelty is the --help option and more importantly the --wait option, so you could include sublime in a script workflow (as a git commit editor ?)
If it works like the OSX subl command, it integrates with git nicely so it can pause git commands to allow you edit commit messages in sublime and resume the command once you save. If you try to do this with standard exe in windows, the git command will run start to finish and you will not be able to edit the commit messages in sublime.
The reason it exists is for things like git, so they can launch the editor aimed at a commit message, then block until the editor closes and then continue.
It's also useful if you want to open the editor aimed at a file in the current directory without re-navigating there in the gui.
People talk about the "bus factor", and how OSS would have been a better option.
But far more often I've been beaten from the:
1) "Yeah, core devs moved on now, and nobody cares to maintain this OSS software".
2) "Yeah, core devs decided to rewrite everything from scratch and change APIs and ABIs".
Because that doesn't happen outside of the OSS world?
Don't know, haven't seen it as often.
Money and existing customers depending on you and you depending on them provides a good motive to not do (1) and to avoid (2).
Also a lot of no-OSS software, especially of the desktop variety, is unpolished and un-finished, usually a moving target, so (1 -- the owners moving on and nobody picking it up) is more damaging for its user compared to some proprietary software being abandoned which he can still use for ages as is.
For example Microsoft is legendary for avoiding (2) at almost all costs, to the point were 20 year old Windows 98 programs still run as is on a modern OS. Heck, people still use VB6 which was superceded like 13 years ago... If you pick a software with a good following, chances are they will continue to support it.
hardly a noteworthy update.
Sad to see there is still no word on an ARM build.
At this rate, it's best you forget it and revisit the idea in the next few years.
Glad to see the project is still alive :)
It's hardly still alive. These aren't even cosmetic changes, honestly!
Why does it need cosmetic changes? There are a few actual bugs to sort out first.
Agreed, my point was that it's a dead project as new updates don't even have cosmetic changes. I bought a license, I really like it, but just like Intype and E on Windows and Kod and TextMate on Mac, the fate of Sublime Text is pretty clear.
but textmate is alive and new builds are released quite often! https://github.com/textmate/textmate/
True, TextMate's agony is at a lower pace than Sublime Text, but it's still not having the metabolic rate of Atom and the likes.
Wow, another update (3064)!
I read up on this editor a little bit, but the thing that gets me is: what's wrong with emacs? I don't see anything this editor does that emacs can't do.
Just my 2 cents...
Sublime feels more lightweight and faster.
It has sensible defaults, and there is not much need to configure stuff.
It has a really "standard" interface (e.g. Tabs, Ctrl+PageUp/PageDn to switch between them, Ctrl+XCV for Clipboard, on Mac it uses the corresponding modifier keys instead). Emacs is rather idiosyncratic, to put it mildly.
It is prettier, and it's easier to install pretty themes (although not perfect yet), and while that may sound petty, it is important if you stare at it the whole day. I found you can make emacs look nice, but there was always some things that annoyed me.
Most importantly, antialiased fonts didn't work on some platforms (I'm sure you can make it work everywhere with some effort though). With sublime, you get native looking, readable fonts on every platform.
You have extensibility, but not so much that you can easily break the editor. The LISP environment in emacs is not beneficial for me, but rather a source of problems. Though, I would like a tiny bit more configurability in sublime - mainly the ability to place icons in the sidebar.
Last but not least, it has the killer feature of multiple selections (Ctrl+D). This allows you to do many cool editing maneuvers that you would use special emacs commands, key combos, regexes etc. (or in vi: movement combinations), but with just one simple key combo + cursor keys + shift, and most important, interactively. You don't have to think "I want to select this and that, but not that" before you press the buttons, you can "just do it".
Of course, emacs can do some things sublime can't. Extreme extensibility is one, but that's not important to me. More important is that it can run in a terminal (e.g. over ssh). It also has the ability to use different fonts and to embed images in its editor (useful for LaTeX).
It's not that there is anything emacs can't do that sublime can. Sublime is just more pleasant to use in my opinion.
----
IMHO there is no justification nowadays for most apps to not work instantaneously, given how fast computers are. If you have to initialize stuff, do it at install time, not at startup time. I want to click the button and have the gui immediately there. Interestingly, emacs pioneered this. IIRC, it has a function to dump its memory to disc, and to just load the memory image at startup.
To understand what's wrong with emacs pretend that you are a new user that doesn't know emacs, but is familiar with editors in browsers, and other places ordinary people use.
When you start to use emacs everything will be wrong, the few shortcuts you know such as ctrl-c/v/x/z/f will do strange things, scrolling won't move text smoothly like other programs but will jump several lines always keeping first line exactly aligned, so you feel lost all the time. When you select something and scroll emacs will destroy the selection and keep cursor in viewport (i am still not sure if this is a feature, or just a glaring bug that no one cares to fix).
Now you try sublime and it is beautiful! It comes with beautiful theme, beautiful chrome-like tabs (they even scroll!), fantastic smooth scrolling, pretty minimap that helps you see whole document at once, built in fuzzy search, and easily configurable plugins.
Why would you even try learning emacs where you constantly feel lost, instead of sublime which makes you feel powerful?
> I don't see anything this editor does that emacs can't do.
in short it's not only important to do something, but do it with style;)
Ok, as I am an emacs user, I somewhat see your point. But the counterpoint is that most hackers spend a good part of their day on text editors and ide's, so it's not unreasonable to spend some time learning it for greater productivity gains later. Yes, there is a greater learning curve with emacs, but it seems to be more extensible than any other editors out there. The ctrl-c/v/x/z/f shortcuts can easily be remapped, the minimap module is available, as well as multiple cursors. That's the great thign about emacs, it can be extended to do just about anything.
It reminds me of a quote from the great Marvin Minsky: “A computer is like a violin. You can imagine a novice trying first a phonograph and then a violin. The latter, he says, sounds terrible. That is the argument we have heard from our humanists and most of our computer scientists. Computer programs are good, they say, for particular purposes, but they aren’t flexible. Neither is a violin, or a typewriter, until you learn how to use it.”
I agree that extensibility is a great thing, that's why i spent some time learning emacs. But unfortunately no amount of elisp could fix the bugs in core: jumpy scrolling, and keeping cursor always on screen. And more importantly after some time it turned out that i used all that extensibility mostly to get on par with other editors, so to me emacs felt more like a construction kit for a violin rather than an actual violin.
Hopefully new generation of browser based editors like cloud9 or atom will provide extenibility similar to emacs, but will start from more sensible place.
It's not what it can't do or how many things more Emacs can do, it's how each goes about them.
Not being tied into a BS-pseudo-Lisp runtime is an essential for me. And not having been designed with 1990 display technology in mind too.
I still don't get it. What's wrong with using a lisp as the underlying architecture? This should be an advantage.
Not for me.
1) I don't consider Lisp great for text work.
2) I don't consider Emacs Lisp a great Lisp either.
3) Even more fundamentaly, I don't think the kind of hooks Emacs offers for extensibility, with the modes etc, are to my liking, especially in how the interoporate with the visual representation.
4) And I'd rather, personally, use Python for my plugins.
One of the problems is that it's a terrible Lisp and runtime. If it was at least a Scheme, there would be less complaining. Also, Lisps aren't popular, as you surely know.
The display technology assumes a text terminal, just like vim. There are many things you can't do with that which are easy with a proper GUI.
Actually the Lisp runtime of GNU Emacs is not terrible. It's just that there are a few better in the Lisp world. Some of the things in Emacs Lisp are there by design.
> The display technology assumes a text terminal,
GNU Emacs and Xemacs have a lot of support for non-text-terminal displays.
Why is this being downvoted? As an emacs user, I'm trying to figure out what's the advantage to using sublime.
Number one advantage: it is not emacs.
emacs? The thing that gets me about you emacs users is, what's wrong with vim?
it's easier to extend the editor with emacs-lisp than vimscript is the main advantage. But yes, vim isn't bad as a text editor either. I use both.
vim? I don't understand why manually punching cards isn't good enough for you.