Autocomplete issues
Posted: 08.03.2023 08:25
First of all, thanks, CudaText is amazying, I feel I finally found a text editor I'll be comfortably with on Linux. It's really impressive how customizable and powerful it is, it can also handle big minified files that other editors may struggle with.
Not sure if I should post this here or in "Bugs, issues", as it may be just a question. Sorry if I chosen wrong.
My main concern is the autocomplete. I use autocomplete_autoshow_chars = 1. I installed "Complete From Text". Yet, autocomplete wasn't working for JavaScript files, for instance. So I also installed "JS Tern" and now it kinda works. But with some caveats:
I don't know how to make "Complete From Text" to work along with "JS Tern" in JS files. If I add ",JavaScript" to "lexers=" in plugins.ini, only "Complete From Text" works, not "JS Tern". If I restore default "lexers=", then it's the opposite, only "JS Tern" works.
I'm also not satisfied on how JS Tern works. For instance, it doesn't suggest DOM words like "window", "document" and its properties, or even words like "function", unless they were already used in the context of where the cursor is placed. I'm used to simple autocomplete engines that just offer known words for the language and also words present anywhere in the file. Is there an existing plugin or config for that?
One other thing is that the CudaText autocomplete popup is slow to show up. In my past text editors (Notepad++ on Windows and Kate on Linux), I can type with my normal speed (fast) and still I can see the suggestions popup is revealed while I'm in the middle of the word. With CudaText, on the other hand, I need to pause typing to give time for the autocomplete to be displayed. Is there any config about that to turn it almost-instant like the other editors? I guess the current delay is to display the popup, not to process the matches, because once the popup is open, further word filtering is instant while I type more letters. Actually, it's a little strange that autocomplete popup is displayed as a different window in my task bar instead of just a component of main CudaText window.
P.S.: just saying because I'm not sure if you're aware, but it's possible to enable "Discussions" tab in GitHub repo, it works mostly like this forum and it could prevent things such as the "talk" label in Issues tab. It supports markdown, anyone with a Github account can post...
Edit: it would also be interesting if the suggestion popup can be reduced to fit the suggestion size: if there's only one autocomplete suggestion, it should have the height for just one line. It's width could also match just the enough to fit the word.
Not sure if I should post this here or in "Bugs, issues", as it may be just a question. Sorry if I chosen wrong.
My main concern is the autocomplete. I use autocomplete_autoshow_chars = 1. I installed "Complete From Text". Yet, autocomplete wasn't working for JavaScript files, for instance. So I also installed "JS Tern" and now it kinda works. But with some caveats:
I don't know how to make "Complete From Text" to work along with "JS Tern" in JS files. If I add ",JavaScript" to "lexers=" in plugins.ini, only "Complete From Text" works, not "JS Tern". If I restore default "lexers=", then it's the opposite, only "JS Tern" works.
I'm also not satisfied on how JS Tern works. For instance, it doesn't suggest DOM words like "window", "document" and its properties, or even words like "function", unless they were already used in the context of where the cursor is placed. I'm used to simple autocomplete engines that just offer known words for the language and also words present anywhere in the file. Is there an existing plugin or config for that?
One other thing is that the CudaText autocomplete popup is slow to show up. In my past text editors (Notepad++ on Windows and Kate on Linux), I can type with my normal speed (fast) and still I can see the suggestions popup is revealed while I'm in the middle of the word. With CudaText, on the other hand, I need to pause typing to give time for the autocomplete to be displayed. Is there any config about that to turn it almost-instant like the other editors? I guess the current delay is to display the popup, not to process the matches, because once the popup is open, further word filtering is instant while I type more letters. Actually, it's a little strange that autocomplete popup is displayed as a different window in my task bar instead of just a component of main CudaText window.
P.S.: just saying because I'm not sure if you're aware, but it's possible to enable "Discussions" tab in GitHub repo, it works mostly like this forum and it could prevent things such as the "talk" label in Issues tab. It supports markdown, anyone with a Github account can post...
Edit: it would also be interesting if the suggestion popup can be reduced to fit the suggestion size: if there's only one autocomplete suggestion, it should have the height for just one line. It's width could also match just the enough to fit the word.