RText plugin for Notepad++. More detailed documentation is pending.
sanastasiou / rtextnpp Goto Github PK
View Code? Open in Web Editor NEWLicense: Other
License: Other
Context extraction doesn't work when a line consists only of new line.
To reproduce extract context,
type enter and go to a new empy line
extract context again.
Expected result, new context will be extract.
Observed result, context remains same.
When modifying a line with some attributes, try writing a comma, save the file, and press ctrl+space to autocomplete. After a few retries should crash. It looks like the plugin tries to autocomplete while still reloading the saved data and then crashes.
Try another style from Settings -> Style configurator, the colors are broken. I guess the highlighting colors should be adapted automatically to each style.
Folding markers seem to be randomly missing from part of even small documents. Check if this is an issue of the plugin or scintilla...
When a model is corrected, annotation are being removed, however subsequent line visibility changes bring them back on from the local cache.
Indicators are staying in the document even after the model reloads and has no errors.
Would be nice to normalize background colors.
The annotations are being displayed not directly under the correct line, rather one line more..
It appears the windows can extend over the screen size if zoom scale factor is too big.
Probably the scaling factor is not being taken into account while checking if the window fits in the monitor.
Noticed that no autocompletion options appear when adding a linebreak (and newline) between the child label and opening square bracket (see screenshot). With argument arrays and opening braces this is no problem.
By the way I can't see the request/response in the debug output anymore with v0.7.258, probably because this is a release build or?
The reference links point correctly, but notepad++ jumps to random places if broken lines are used...
Current feature setup branch has this problem. It seems impossible to remove indicators once they are drawn...
Because ctrl+alt is used for reference highlighting...
Could this be the reason you're not able to catch NPPN_DOCORDERCHANGED
? It has the same value as NPPN_READONLYCHANGED
In the Notepad++ source code it is defined as NPPN_FIRST + 17
.
It appears to be a problem with the state machine. Connector is stuck at "busy" while nothing happens.
Also an out of bound exception exists.
Progress updates but description and backend status is showing ready.
Ctrl+C or Ctrl+K currently have no effect on the plugin.
Enable the shortcuts for commenting, taking into account the settings of those shortcuts.
Getting unhandled .net exception when missing "cmd /c" in front of the command in .rtext and trying to load the model:
See the end of this message for details on invoking
just-in-time (JIT) debugging instead of this dialog box.
************** Exception Text **************
System.InvalidOperationException: No process is associated with this object.
at System.Diagnostics.Process.EnsureState(State state)
at System.Diagnostics.Process.get_HasExited()
at RTextNppPlugin.RText.RTextBackendProcess.get_HasExited()
at RTextNppPlugin.RText.ConnectorManager.d__0.MoveNext()
--- End of stack trace from previous location where exception was thrown ---
at System.Runtime.CompilerServices.AsyncMethodBuilderCore.b__0(Object state)
When editing file located on a mapped drive, autocompletion works but after saving load model is not sent (assumed problems with file update event). Same problem when updating .rtext (backend not reloaded).
The auto completion option is not being inserted, cursor stays at same point. Without space after comma the option is correctly inserted.
Just try to update some other plugin while RTextNpp is running. Notepad++ won't shutdown and will hang.
Displaying all annotations simultaneously causes lag, when there are thousands of annotations.
Instead, only the annotations that belong to visible lines should be displayed.
Active document contains no error after model reload. Errors first appear after a buffer activated event.
User cannot type various characters when their language settings isn't english...
Message routing is not correct. Currently there is no way to restore a file state to "not" modified even if the user has undone all edits.
Fix this by routing correct messages to appropriate function of the FileModificationObserver class.
The plugin works in the background to populate errors, even when the relevant setting is disabled.
In order to improve performance when dealing with massive amount of errors, the annotations should be cached while being invisible and then be made visible.
This has the effect that no matter the number of annotations, they appear instantly in the editor.
When multiple child elements (ie. array) are assigned to a role (label), empty autocompletion options are given in the cursor position shown below (expected "CalprmElementPrototype"). Also syntax highlighting is wrong (CalprmElementPrototype colored red like an attribute).
AUTOSAR {
ARPackage P1 {
InternalBehavior InternalBehavior1, component: /P1/SWC1 {
sharedCalprm: [
|
CalprmElementPrototype sharedCal1, type: /P1/UInt8
CalprmElementPrototype sharedCal2, type: /P1/UInt8
]
}
}
}
request:
Info: 3:46 PM:41: [15:46:41] DEBUG request: {"context"=>["AUTOSAR {", "ARPackage P1 {", "InternalBehavior InternalBehavior1, component: /P1/SWC1 {", " sharedCalprm: [\t "], "column"=>25, "type"=>"request", "command"=>"content_complete", "invocation_id"=>45}
response:
Info: 3:46 PM:41: [15:46:41] DEBUG response: {"type"=>"response", "invocation_id"=>45, "options"=>[]}
If auto completion is requested where only a single option exists, then it appears as if the user cannot type anything.
This feature should only be active when the user explicitly requests auto completion with Ctrl+Space.
Brackets are treated as errors...
Autocomplete gives just "AUTOSAR" after a backslash (should give options for the command in the previous line).
Plugin gives a new element as autocompletion result (because request column is wrong), even though attribute value is expected:
AUTOSAR {
ARPackage P1 {
InternalBehavior InternalBehavior1, component: /P1/SWC1 {
RunnableEntity run8 {
WaitPoint wp1, trigger: |
}
}
}
request:
Info: 3:19 PM:34: [15:19:34] DEBUG request: {"context"=>["AUTOSAR {", "ARPackage P1 {", "InternalBehavior InternalBehavior1, component: /P1/SWC1 {", "RunnableEntity run8 {", "WaitPoint wp1, trigger:", ""], "column"=>1, "type"=>"request", "command"=>"content_complete", "invocation_id"=>38}
When a cloned document is open in both views, buffer activated events are interfering with both of the views. Last view seems to "win", thus visibility info ( lines ) are shared between both views, which messes up the drawn indicators...
When .rtext is modified after the backend runs, the backend will restart but the .rtext file will not be re-read. Editor restart required for the new changes.
To reproduce, just try to run the plugin at a pc without SDK installed.
Even when no errors are present, jumping to a reference link causes margin to be drawn.
When the mouse hovers quickly over a lot of elements, the number of request sent to the backend will need to be debounced. Otherwise editor will hang.
When the user is typing, sometimes although the list if filtered, the best match is not being brought into view. This has as a result, among others, that the description tooltip appears out of place.
Consequently pressing Ctrl+Space at the end of a line where the last available option has been entered and where this option is a literal, the last available option is re-entered.
Solution do not insert anything after a label is something is already inserted. Instead just show the completion option to the user.
The editor doesn't move the text into view if line wraps are active..
Document appears without syntax highlighting and folding.
Purge caching when auto completion is invoked on another line...
Some times, highlighting is hanging - context is not updating, trigger buttons seem to be constantly pressed, even if released.. etc..
The console text should fit the actual size of the console, thus when this size is enlarged the lines should also rearrange themselves.
Currenlty the tooltip is placed on the right of a border of an option item.
The problem with this approach is, that if the option is too large to fit in the maximum length of the auto completion window, the tooltip appears out of place in the "virtual" edge of the border.
To fix this, always place the left edge of the tooltip to the right edge of the completion list.
Margins appear to be persistent, regardless of whether they are activated or not.
Reference link window below shifted down and right at some cases...
If an automate file is opened, the keyboard events are marked as handled and not routed, even if another windows e.g. Options has got Focus.
Fix this by checking if the editor has focus...
A declarative, efficient, and flexible JavaScript library for building user interfaces.
๐ Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
An Open Source Machine Learning Framework for Everyone
The Web framework for perfectionists with deadlines.
A PHP framework for web artisans
Bring data to life with SVG, Canvas and HTML. ๐๐๐
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
Some thing interesting about web. New door for the world.
A server is a program made to process requests and deliver data to clients.
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
Some thing interesting about visualization, use data art
Some thing interesting about game, make everyone happy.
We are working to build community through open source technology. NB: members must have two-factor auth.
Open source projects and samples from Microsoft.
Google โค๏ธ Open Source for everyone.
Alibaba Open Source for everyone
Data-Driven Documents codes.
China tencent open source team.