SciTE v 3.20 Beta available

[Jos]

I have merged Neils SciTE v 3.20 Source with our own repository and have made it available in a separate directory in the Beta site.

SciTE v 3.20

SciLexer.dll v3.20

History of changes

My Changes:

• Support for "(" in filenames on error lines to fix the issue that double clicking these error lines wouldn't work.
• Jump to Line and Column now when double clicking the error from au3Check.
• Fold in Commentblocks fixes for the AU3 lexer.
• Word Highlighting modifications:

  highlight.current.word=1

  highlight.current.word.by.style=1

  highlight.current.word.colour=#00D040

  #~ highlight.current.word.autoselectword=1

  #~ highlight.current.word.wholeword=1

  #~ highlight.current.word.matchcase=1

  #~ highlight.current.word.minlength=2

Appreciate when people can test and let me know if they see issues or if all is working fine.

Cheers,

Jos

Updated to 3.20 6/9/2012.

Edited June 9, 2012 by Jos

[jaberwacky]

Do we just replace the current SciTE 2.X with this one? Nevermind, I imagine so.

First report. Blank lines have fold points.

Edited January 14, 2012 by LaCastiglione

[Jos]

Correct for both files.

[JScript]

I still haven't found problems, but your changes are amazing! Thank you very much!

Regards,

João Carlos.

[wraithdu]

Jos, would you mind posting when you update SciTE.exe / SciLexer.dll in your beta directory so we can be sure we're testing your most updated build? My memory is pretty good, but modified date and time has limited retention

Edited January 15, 2012 by wraithdu

[Jos]

Jos, would you mind posting when you update SciTE.exe / SciLexer.dll in your beta directory so we can be sure we're testing your most updated build? My memory is pretty good, but modified date and time has limited retention

Sure. Somehow I managed to make the same mistake as the last time and missing the larger properties file support which is fixed in the current version available.

There was also a minor change in the Lexer changes I made to support folding in Commentblock which was fixed in both Beta versions of SciLexer.dll.

Jos

[Melba23]

Jos,

#include lines do not seem to be the correct colour with the new v3.02 SciTE.exe and SciLexer.dll:

They are correctly coloured in the v2.28 SciTE version.

From my "SciTEUser.properties" file:

#Special
style.au3.12=fore:#A00FF0,italics,back:#E9FEE0

As you can see, all the other "special" codes like #region seem to take the correct colour.

M23

P.S. And no comments about the colour scheme - I like it!

[Jos]

Have you download the latest version I uploaded around 10PM CT yesterday?

[Melba23]

Jos,

Obviously not! Sorry.

I downloaded as soon as you created the thread and only got around to testing them just now. Is there anything in particular we should look for with this new SciTE version?

M23

[Jos]

The only thing I changed in this version of SciTE 3.02 was the patch i missed to support the bigger Properties files we need, hence the lack of the last keywords from au3.keywords.properties.

As to this new release of SciTE in general: All I have done is to merge our changes into it and haven't done any special testing, so curious to see if people find any nice new stuff in it that is worth mentioning and obviously if all is working for them.

Thanks

Jos

[Valik]

Jos, have a look at the following properties:

error.inline
style.error.0
style.error.1
style.error.2
style.error.3

Might be some useful stuff there.

[Valik]

Here's what I came up with:

error.inline=1
style.error.1=fore:#FF0000,back:#FFFF00
style.error.2=fore:FFFFFF,back:#FF0000

However, there's a bug either with the error.inline feature or with Au3Check. It detects error messages just fine. It draws the line with a red background and white text. It does not detect warnings, though. It draws those in the style.error.0 style instead of style.error.1.

Note: To see this feature in action insert some errors into a script and then run Au3Check. After it is complete double-click on an error line in the output pane and you should see all the errors show up inline with the document as well.

Edit: Wow the feature needs work. My expected behavior is for the errors to show up after running Au3Check. I also expect re-running Au3Check to clear the errors. Doesn't happen. Instead it's possible to get stuck with an orphaned error in the document which won't go away even though the code has been fixed.

Edited January 15, 2012 by Valik

[jaberwacky]

First report. Blank lines have fold points.

This only seems to occur when technology=1. I'm guessing you don't have control over this but I thought I'd mention it if someone else were to experience the same.

[Jos]

This only seems to occur when technology=1. I'm guessing you don't have control over this but I thought I'd mention it if someone else were to experience the same.

Could you give an example snippet to demo this as we do have a couple of situations we have done this deliberately.

[Jos]

Here's what I came up with:

error.inline=1
style.error.1=fore:#FF0000,back:#FFFF00
style.error.2=fore:FFFFFF,back:#FF0000

However, there's a bug either with the error.inline feature or with Au3Check. It detects error messages just fine. It draws the line with a red background and white text. It does not detect warnings, though. It draws those in the style.error.0 style instead of style.error.1.

Note: To see this feature in action insert some errors into a script and then run Au3Check. After it is complete double-click on an error line in the output pane and you should see all the errors show up inline with the document as well.

Edit: Wow the feature needs work. My expected behavior is for the errors to show up after running Au3Check. I also expect re-running Au3Check to clear the errors. Doesn't happen. Instead it's possible to get stuck with an orphaned error in the document which won't go away even though the code has been fixed.

Will have a look at all of this but will not have much time this week.....

[Valik]

I found a "solution" for the orphaned errors issue. Search -> Clear All Bookmarks removes them.

[Valik]

A few more comments after I've just spent some time reviewing SciTE's properties. First a discovery (maybe not new but new to me): The following will draw the lines that show what will be folded in a different color making it easier to visualize what will be folded:

fold.highlight=1
fold.highlight.colour=#00FF00

It requires fold.symbols to be set to either 2 or 3.

Also, I found the settings for the background loading/saving: background.open.size, background.save.size. These take a size in bytes. So the following should disable background loading and saving for anything under 2GB:

background.open.size=2147483647
background.save.size=2147483647

To enable background loading/saving for all files use:

background.open.size=-1
background.save.size=-1

The default is to use it for all files so if Jos didn't override this already that is what's happening.

The technology property causes SciTE to use Direct2D/DirectWrite. It's supposed to look better but I personally didn't like how things looked when I turned it on.

Finally about the inline errors. Those are implemented via something called Annotations. It may be something that can be done via Lua. Also I've discovered that there is a menu command (Tools -> Next Message) that will automatically jump to a message (moves the cursor, boo) and show the inline errors (yay). It may be possible to do something like wrap invoking Au3Check in Lua so that we can do fun things like automatically show the inline errors after running Au3Check as well as automatically clearing orphaned errors by just running Au3Check cleanly. It may also be possible to save/restore the current position so it's a little less annoying.

[jaberwacky]

It happens when you fold and or unfold stuff. Blank lines and other unfoldables will get fold points with technology=1.

Func tidy_view_get_run_diff_prog(Const $this, Const $show)
    Switch $show
        Case '', ' ', 0
            GUICtrlSetState($tidy_select_diff_prog_path, $GUI_DISABLE)
            GUICtrlSetState($tidy_diff_prog_path,       $GUI_DISABLE)
            GUICtrlSetData ($tidy_diff_prog_path,       $show)
        Case Else
            GUICtrlSetState($tidy_select_diff_prog_path, $GUI_ENABLE)
            GUICtrlSetState($tidy_diff_prog_path,       $GUI_ENABLE)
            GUICtrlSetData ($tidy_diff_prog_path,       $show)
    EndSwitch
    Return $this
EndFunc

Func tidy_view_get_add_function_tail_comments(Const $this, Const $add)
    Switch $add
        Case 0
            GUICtrlSetState($tidy_add_endfunc_comment,    $GUI_UNCHECKED)
        Case 1
            GUICtrlSetState($tidy_add_endfunc_comment,    $GUI_CHECKED)
    EndSwitch
    Return $this
EndFunc

[Jos]

It happens when you fold and or unfold stuff. Blank lines and other unfoldables will get fold points with technology=1.

I would say about the same the doctor said when the patient told him it hurts when he presses a certain spot: Stop pressing the particular spot.

technology: On Windows Vista or newer, this can be set to 1 to use the Direct2D and DirectWrite APIs for higher quality antialiased drawing. The default is 0.

Edited January 16, 2012 by Jos

[jaberwacky]

Ah hah, I thought you were being serious about wanting a snippet.