What version of OSX are you both using?

I've never seen this issue, but I'm also using the latest TM 2.0.11.

Is this version not available for you?

On Sun, 10 May 2020 at 21:53, Matt Neuburg matt@tidbits.com wrote:

This is the same issue I’ve been complaining about for months. Because of it, I’m still stuck back at 2.0rc10; anything after that is unusable. I don’t understand why everyone else doesn’t feel the same; maybe they all have only short non wrapping lines of code. Files such as you and I use, where a line is a paragraph of text, show the issue. m. -- matt neuburg, phd = http://www.apeth.net/matt/ pantes anthropoi tou eidenai oregontai phusei Programming iOS 13! http://shop.oreilly.com/product/0636920310075.do iOS 13 Fundamentals! http://shop.oreilly.com/product/0636920310068.do RubyFrontier! http://www.apeth.com/RubyFrontierDocs/default.html

On May 10, 2020, at 12:17 PM, Tao Zha zmail@tzha.net wrote:

Hi:

When I use Mac’s recent TextMate version (2.0.6) to edit my latex file (a .tex file), TextMate rolls the screen up randomly away from the cursor so that the screen will not stay in one place. Do you know there is any way to fix this annoying problem? Older versions of TextMate did not have such a problem.

Thanks,

Tao

---

TextMate mailing list TextMate@lists.macromates.com https://lists.macromates.com/listinfo/textmate

---

TextMate mailing list TextMate@lists.macromates.com https://lists.macromates.com/listinfo/textmate

Hey, Matt, I'm curious. How does one make it happen? I mean, how long do the lines need to be to cause the scrolling behavior? I fooled around in 2.0.11 some this afternoon and couldn't make it happen. I'm on Mojave, maybe that's what it is for me.

I'm one of those users who has "short non wrapping lines of code". :)

Got a test case file? Not a real file, just something that causes it.

-- Marc Wilson posguy99@gmail.com

On Sun, May 10, 2020, at 1:53 PM, Matt Neuburg wrote:

This is the same issue I’ve been complaining about for months. Because of it, I’m still stuck back at 2.0rc10; anything after that is unusable. I don’t understand why everyone else doesn’t feel the same; maybe they all have only short non wrapping lines of code. Files such as you and I use, where a line is a paragraph of text, show the issue. m.

matt neuburg, phd = http://www.apeth.net/matt/ pantes anthropoi tou eidenai oregontai phusei Programming iOS 13! http://shop.oreilly.com/product/0636920310075.do iOS 13 Fundamentals! http://shop.oreilly.com/product/0636920310068.do RubyFrontier! http://www.apeth.com/RubyFrontierDocs/default.html

...

On May 10, 2020, at 12:17 PM, Tao Zha zmail@tzha.net wrote: Hi:

When I use Mac’s recent TextMate version (2.0.6) to edit my latex file (a .tex file), TextMate rolls the screen up randomly away from the cursor so that the screen will not stay in one place. Do you know there is any way to fix this annoying problem? Older versions of TextMate did not have such a problem.

Thanks,

Tao

---

TextMate mailing list TextMate@lists.macromates.com https://lists.macromates.com/listinfo/textmate

---

TextMate mailing list TextMate@lists.macromates.com https://lists.macromates.com/listinfo/textmate

Now, that’s weird. Yeah, happens on Mojave. TM 2.0.11.

I did notice...

It doesn’t require a search to trigger it, just something that makes the screen shift positions (you may have said that before). Open the file, hit ⌘↓ to go to the end of the file. It moves to the end of the file, then moves back. Hit ⌘↓ again, it’ll actually go to the end of the file this time.

Hit ⌘↑ to go back to the top. Search for ‘clipping’, it finds it then scrolls back. Keep hitting ⌘G until TM tells you there are no more instances. Hit ⌘↑, then hit ⌘G again. It goes back to the first ‘clipping’ on line 739, but the screen doesn’t scroll back this time. Keep hitting ⌘G. No more odd scrolling.

Are these behaviors because now it knows the end of all the lines?

-- Marc Wilson posguy99@gmail.com

On Sun, May 10, 2020, at 2:22 PM, Matt Neuburg wrote:

Sure, here's one enclosed.

Just view that as Plain Text. Position yourself at the top and find "clipping". On all my machines, we jump backwards instantly after the find. m.

...

On May 10, 2020, at 2:11 PM, Marc Wilson posguy99@gmail.com wrote:

Hey, Matt, I'm curious. How does one make it happen? I mean, how long do the lines need to be to cause the scrolling behavior? I fooled around in 2.0.11 some this afternoon and couldn't make it happen. I'm on Mojave, maybe that's what it is for me.

I'm one of those users who has "short non wrapping lines of code". :)

Got a test case file? Not a real file, just something that causes it.

-- Marc Wilson posguy99@gmail.com

On Sun, May 10, 2020, at 1:53 PM, Matt Neuburg wrote:

...

This is the same issue I’ve been complaining about for months. Because of it, I’m still stuck back at 2.0rc10; anything after that is unusable. I don’t understand why everyone else doesn’t feel the same; maybe they all have only short non wrapping lines of code. Files such as you and I use, where a line is a paragraph of text, show the issue. m.

matt neuburg, phd = http://www.apeth.net/matt/ pantes anthropoi tou eidenai oregontai phusei Programming iOS 13! http://shop.oreilly.com/product/0636920310075.do iOS 13 Fundamentals! http://shop.oreilly.com/product/0636920310068.do RubyFrontier! http://www.apeth.com/RubyFrontierDocs/default.html

...

On May 10, 2020, at 12:17 PM, Tao Zha zmail@tzha.net wrote: Hi:

When I use Mac’s recent TextMate version (2.0.6) to edit my latex file (a .tex file), TextMate rolls the screen up randomly away from the cursor so that the screen will not stay in one place. Do you know there is any way to fix this annoying problem? Older versions of TextMate did not have such a problem.

Thanks,

Tao

---

TextMate mailing list TextMate@lists.macromates.com https://lists.macromates.com/listinfo/textmate

---

TextMate mailing list TextMate@lists.macromates.com https://lists.macromates.com/listinfo/textmate

---

TextMate mailing list TextMate@lists.macromates.com https://lists.macromates.com/listinfo/textmate

-- matt neuburg, phd = http://www.apeth.net/matt/ pantes anthropoi tou eidenai oregontai phusei Programming iOS 13! http://shop.oreilly.com/product/0636920310075.do iOS 13 Fundamentals! http://shop.oreilly.com/product/0636920310068.do RubyFrontier! http://www.apeth.com/RubyFrontierDocs/default.html

---

TextMate mailing list TextMate@lists.macromates.com https://lists.macromates.com/listinfo/textmate

*Attachments:*

• TestFile.txt

On 11 May 2020, at 3:53, Matt Neuburg wrote:

This is the same issue I’ve been complaining about for months. Because of it, I’m still stuck back at 2.0rc10; anything after that is unusable. I don’t understand why everyone else doesn’t feel the same; maybe they all have only short non wrapping lines of code. Files such as you and I use, where a line is a paragraph of text, show the issue. m.

I think I have acknowledged this, but it really is a design flaw that requires a new layout engine. I know you say it has gotten worse after 2.0rc10, but I really can’t think of what has changed: Sometimes system API change behavior when linking with different SDK versions, although in this case, it would be weird, but it wouldn’t be the first time something like this has happened.

I have started work on new editor/layout engine, and work is going fairly well, already much improved speed compared to current release of TextMate (no issue with large files or long lines) and better unicode support, but there is quite a threshold to pass before it can replace the current shipping version, so I hope for some patience.

Support for LSP and related features is also closely related to this work and a high priority.

On 11 May 2020, at 04:04, Allan Odgaard mailinglist@textmate.org wrote:

I think I have acknowledged this, but it really is a design flaw that requires a new layout engine. I know you say it has gotten worse after 2.0rc10, but I really can’t think of what has changed: Sometimes system API change behavior when linking with different SDK versions

Really? Isn’t all code dynamically linked, which would mean that actual implementation which is executed at runtime won’t change when linking with a new SDK. Unless they have some macro or other way redirect an existing function to a new implementation.

I have started work on new editor/layout engine, and work is going fairly well, already much improved speed compared to current release of TextMate (no issue with large files or long lines) and better unicode support

That’s great to hear.

Support for LSP and related features is also closely related to this work and a high priority.

Awesome, I’ve been wanting to have LSP integration for a long time.

On 16 May 2020, at 03:31, Allan Odgaard mailinglist@textmate.org wrote: From Apple: “The system detects how an app was built by looking at the “SDK” entry in the app’s Mach-O header. When you relink your app against the latest SDK, you might notice different behaviors, some of which might cause incompatibilities.”.

Ha! Sneaky.

I tried to find a working LSP implementation that would provide “intelligence” for some of my projects, and the results were somewhat disappointing.

Hmm, that’s unfortunate. It depends on which language those projects are written in, but for TextMate which is implemented in C++ and Objective-C++ there’s clangd [1]. I have not used it myself though so I don’t know how good it is. There’s also SourceKit-LSP [2], which is an LSP server for Swift and C-based languages. It uses clangd under the hood for the C-based languages. I don’t know if it’s any better than plain clangd.

For testing the client you could try the LSP server that Visual Studio Code uses for C# or TypeScript. I would expect those to be top notch, since they’re heavily prompted by Microsoft.

langserver.org http://langserver.org/ has a list of LSP server implementations as well [3].

But too early to give any ETA, as I currently add items to my to-do faster than I can check them off (as mentioned, this isn’t just about LSP, there are lots of other things that needs to be improved in TextMate, and most of it really requires major overhauls).

Yeah, I can imagine.

Not sure if you want/need it but there’s a Swift library, SwiftLSPClient [4], for communicating with LSP servers. It’s used by the Chime text editor [5], a Go editor for macOS.

Perhaps it’s too soon to start discussing, but clangd supports some useful extensions to the LSP protocol [6]. I’m mostly thinking of the UTF-8 offsets. Code completion scores would be nice to have as well.

[1] https://clangd.llvm.org https://clangd.llvm.org/ [2] https://github.com/apple/sourcekit-lsp https://github.com/apple/sourcekit-lsp [3] https://langserver.org/#implementations-server [4] https://github.com/ChimeHQ/SwiftLSPClient https://github.com/ChimeHQ/SwiftLSPClient [5] https://www.chimehq.com https://www.chimehq.com/ [6] https://clangd.llvm.org/extensions.html