Sublime Forum

Dev Build 3086

#5

Also, adding a word to the dictionary crashes the application entirely.

0 Likes

#6

@wesbos: That could indicate a bad interaction with Package Control. Do you manage the two packages locally (i.e. in a git repo) or does Package Control manage them? Try if disabling package control makesit work. If it does, try adding the two packages to the ignore_vcs_packages setting for PC and report results in a new issue at the PC repo.

0 Likes

#7

I tested the load time with a 37mb sql dump.

Build 3084: 1m19s
Build 3085: 1m03s
Build 3086: 10s

:smiley:

0 Likes

#8

I can’t replicate that. Assuming you’re on OSX, you should be presented with a crash report when this happens - it’d be great if you could email it to me.

0 Likes

#9

[quote=“jps”]

I can’t replicate that. Assuming you’re on OSX, you should be presented with a crash report when this happens - it’d be great if you could email it to me.[/quote]

I’m not experiencing it on OSX Yosemite.

0 Likes

#10

Very grateful for the quick fix for the SublimeSyntax push/set conflict. This image shows one of my favorite things that’s now possible, context-conscious brace coloring. The old arbitrary constraint that braces couldn’t really “participate” in coloring despite context or even core semantic meaning (eg object literal vs block vs binding pattern) always bugged me about tmLanguage. Now I can just makes a whole loop one color, conditional blocks another, etc, instead of just the keywords. The colors are just temporary for testing:

0 Likes

#11

This should be available in all languages. I’d love this in PHP and JavaScript for example.

0 Likes

#12

Hi guys! I downloaded latest ST3 dev build and it said that I need to enter license. Is there any topic about this? I just want to test it and submit bug request if I’ll find something. Thanks

0 Likes

#13

From the sublimetext.com/3dev page (that you should have read) where you downloaded ST:

[quote]These are the in-progress versions of Sublime Text 3, and they are updated more frequently. Feel free to use them, but keep in mind you will be running less tested code, and you’ll be seeing many more update notification prompts.
The current Sublime Text 3 dev build is 3086. Details on new dev build releases are posted on the forum
Sublime Text 3 dev builds are currently available to registered users only.[/quote]

So buy a license or wait for the next beta.

0 Likes

#14

It’s possible to mess up the sidebar view using the API. It can lead to data loss when deleting using context menu in the sidebar. Wrong file can get deleted.
It’s an old problem and it is reported here github.com/bizoo/SortTabs/issues/11 but without very specific details.

I think the root problem is that it’s possible to move ‘temporary’ (single clicked from the sidebar - one without an entry created in open files sidebar group) views using the API.

High level repro steps:

  1. Install github.com/bizoo/SortTabs/
  2. Add some folder to the project with a couple of files, lets say with files ‘1’ and ‘2’
  3. Double click file ‘2’ (order matters because we will be sorting by file name)
  4. Single-click file ‘1’ (to show the preview)
  5. Invoke Ctrl+shift+p -> Sort Tabs Menu -> Sort by file name
  6. Close previewed file ‘1’
  7. Preview file ‘1’ again (step 4)
  8. Again sort by file name (step 5)

Now you will see that sidebar view of open files is broken. The ‘Open files’ header is gone, everything moves up.

0 Likes

#15

JS (ES6) is what I’m doing presently – though I can see how the example screenie doesn’t make that obvious :smile:

0 Likes

#16

Bathos - is your work up on github?

I can’t say how excited I am for your ES6 work!

0 Likes

#17

Seems like variables’ lookup always happens in current syntax, even when variable is defined in embedded one.

There are also some weirdness with backreference. This will end with error saying backref number is wrong:

strings:
- match: '("''])'
  push:
  - meta_scope: string.quoted
  - match: '\1'
    pop: true
  - match: (?<!\\)\n
    pop: true

Though if you remove second match it will work.

0 Likes

#18

Really impressed with the parsing speedups in the recent few builds, Jon. Great job, keep it up!

0 Likes

#19

Where can we detect these new scopes to update our themes to take advantage of this?

I just tried ScopeHunter and it doesn’t seem to work anymore.

0 Likes

#20

It works. Try restarting or re-install it.

0 Likes

#21

I recommend ScopeAlways, personally. But I wonder from the way you phrased the question if you’re expecting these scopes to simply be available with the new build? If so, no, this isn’t the case. That screenshot is from me working on a new syntax definition. Presumably others are already doing the same thing for other languages, so in the near future there will probably be a lot of new syntax definitions to choose from that do provide these refinements.

FWIW, the scopes I used for braces there follow this pattern:

punctuation.definition.block.conditional.begin
punctuation.definition.block.conditional.end
punctuation.definition.expression.conditional.begin
punctuation.definition.expression.conditional.end
punctuation.definition.block.loop.begin
punctuation.definition.block.loop.end
(etc)

The “punctuation.definition.xxx.[begin|end]” pattern is already common in tmLanguage so I figured it made sense this way, but I’m open to other (existing?) ideas for naming these newly “disambiguatable” sorts of things.

0 Likes

#22

OK, thanks. At want point will this yield benefits for those of us who simply write themes?

0 Likes

#23

Thanks for the report, will fix in the next build.

[quote=“lastsecondsave”]
There are also some weirdness with backreference. This will end with error saying backref number is wrong:

strings:
- match: '("''])'
  push:
  - meta_scope: string.quoted
  - match: '\1'
    pop: true
  - match: (?<!\\)\n
    pop: true

Though if you remove second match it will work.[/quote]

Currently backref substitution is only performed for the last pop rule in a context. There’s no fundamental reason for this, it’s just a legacy of the tmLanguage support, which only ever allowed for a single pop rule. I’m not currently planning on addressing this until sregex handles backrefs itself.

0 Likes

#24

Well, I have no idea about anyone else, but if you’re interested in theming Javascript / Ecmascript, feel free to take a look at what I have so far:

github.com/bathos/Ecmascript-Sublime

Quite close to done with the definition itself now actually, though I’m sure there will be a lot of fixes and refinements to make once I get deeper into testing it.

EDIT

The syntax itself is done, save for the ongoing bug screening and refinements.

Here’s a horrific example of what it can do using randomized colors:

0 Likes