[CLOSED] Allow code-folding for selected text in editor

[core-ai-bot]

Issue by thehogfather
Monday Aug 03, 2015 at 21:20 GMT
Originally opened as adobe/brackets#11538

---

This is a small addition to the code folding extension that makes any selected text in the editor foldable (provided the number of lines selected is more than the minimum fold size set in the preference).

The feature can be enabled by setting the preference value "code-folding.makeSelectionFoldable": true

cc@abose

---

thehogfather included the following code: https://github.com/adobe/brackets/pull/11538/commits

[core-ai-bot]

Comment by nethip
Tuesday Aug 04, 2015 at 07:44 GMT

---

@thehogfather Wow! That was quick. Thanks for this PR!.

[core-ai-bot]

Comment by abose
Tuesday Aug 04, 2015 at 10:47 GMT

---

Thanks@thehogfather for the PR.

[core-ai-bot]

Comment by thehogfather
Wednesday Aug 05, 2015 at 23:51 GMT

---

BTW this is still a work in progress. Amongst other things, I need to sort out issues around restoring folds based on selections. Shall update in the next few days. Case 2 below is of particular interest:

1. User selects text, folds it, closes the file. When user reopens the file (without modification) the folded selected text region is restored.
2. User selects text, folds it, closes the file. File is modified outside the context of brackets (e.g., git). User opens the file. What is a reasonable expectation here?

[core-ai-bot]

Comment by abose
Thursday Aug 06, 2015 at 04:28 GMT

---

@thehogfather I have added a review only tag to the title of the PR, please remove it when you feel this pr is ready to be merged!

[core-ai-bot]

Comment by petetnt
Friday Aug 07, 2015 at 05:44 GMT

---

@thehogfather I wrestled with a case similar to 2. with my brackets-persistent-history extension, where I wanted to persist the editing history between project changes and outside changes. My original idea was to invalidate the editing history on the changed part by diffing the files (similar to how you could invalidate / uncollapse changed lines) but in the end I went for a simpler solution where I add another history-step to compensate between the state in the cache and the new changes from outside of Brackets.

However the case here is slightly different so it's hard to say what the reasonable expectation would be: overall I wouldn't want Brackets to uncollapse any folds unless absolutely necessary (ie. the folded content has been modified in a way that it doesn't exist (start/end has been removed/modified) anymore), but then again how would you notice that something has changed inside a collapsed fold? One idea would be showing the change in the collapse marker but that's kinda out there too.

[core-ai-bot]

Comment by thehogfather
Saturday Aug 08, 2015 at 22:27 GMT

---

@petetnt quite!
The case here is also interesting because there really isn't a way to "validate" selection-based folds - any selection is valid - unlike how modifications to a file can render brace-based folds invalid. At the moment what the extension does is to use the starting line of each collapsed region to calculate the 'intended' collapsed region when the file is reopened. If the file has not changed then there is no problem. If the file has changed, then the fold might be invalidated if the start and end lines of the new and cached ranges don't match.

For now I've kept it simple. Selection folds will be restored if and only if they exist in the saved fold preferences and are reported as selection by the ViewStateManager.

It might be worth thinking about rendering selection folds ever-so-slightly differently from other types of folds in the gutter.

[core-ai-bot]

Comment by thehogfather
Wednesday Aug 12, 2015 at 01:10 GMT

---

@abose I think this is ready to merge.

[core-ai-bot]

Comment by nethip
Wednesday Sep 09, 2015 at 14:10 GMT

---

@abose I think we should target this for 1.5

[core-ai-bot]

Comment by abose
Friday Sep 11, 2015 at 13:55 GMT

---

It working well in all cases i tested. External modification is always a grey area, But no data is getting corrupted / opening up the folds on external modification seeems to be reasonable- which is how it is working now mostly.
All unit/integration/main view tests Passing.

Thanks for this feature@thehogfather 👍
Merging.

[core-ai-bot]

Comment by Maetzzen
Friday Oct 16, 2015 at 10:34 GMT

---

Im really happy about this new feature in version 1.5.
But could one tell me if there is a possibility to save this options?

If i restart my brackets, my html files are folded out and i have to manually fold in the stuff i dont want to see again.
There is some content i dont have to see anymore and it would be more clearly if they are folded in.

Greetings Maetzzen

[core-ai-bot]

Comment by ezequias
Friday Oct 16, 2015 at 18:54 GMT

---

Could someone tell me if the collapsed text is correctly iconified on my image (that ugly crossed square)?

[core-ai-bot]

Comment by thehogfather
Friday Oct 16, 2015 at 20:15 GMT

---

@ezequias that is correct and would be and issue to raise with the theming provider you are using.@Maetzzen the fold states should persist when you close and reopen brackets. Can you report what operating system you are running?

[core-ai-bot]

Comment by Maetzzen
Friday Oct 16, 2015 at 20:35 GMT

---

My bracket runs on Windows 8.0

[core-ai-bot]

Comment by ezequias
Friday Oct 16, 2015 at 20:38 GMT

---

Thanks@thehogfather It's Monokai dark soda.

[core-ai-bot]

Comment by thehogfather
Friday Oct 16, 2015 at 22:24 GMT

---

@Maetzzen can you confirm if there is any difference if you select a different file in your workspace before closing brackets. So fold some regions in your html file as you did before then switch to a different file and then close and reopen brackets.

Now check if the folds in your html file are restored or not.