Hide codelens and bug step decorations when document is dirty

[csordasmarton]

If there are unpersisted changes in the current document we will hide the provided codelens items and bug step decorations.

[Discookie]

• Stepping into a dirty file is very unintuitive this way. The user clicks next step, switches from eg. a clean main.cpp to a dirty main.h, and no repr. steps are shown anywhere except the sidebar.
  There should be a popup or some other indicator that the steps are not shown because of the dirty flag. (Some extensions do this via a warning on line 0 column 0.)

• When the user presses accidentally types a character and then undo's it, CodeLens reappears, but the highlights do not.

• I'm not sold on the idea of hiding the highlights in the currently active file without a file switch. As long as there's no file switch or metadata reload, I think the diagnostics render fine.

There's also the consistency issue of a step's file being changed, so eg. changing main.h doesn't refresh the reports in main.cpp, and main.cpp's steps still render in an invalid way. That might be better suited for a separate PR though.

[Discookie]

As discussed in a meeting, because CodeLens doesn't move properly with the code on Linux, hiding the decorations in the current file is fine.

Once we find a solution for #35, the decorations won't need to be hidden in the currently edited file anymore.

Otherwise LGTM!

[csordasmarton]

If there are unpersisted changes in the current document we will hide the provided codelens items and bug step decorations.

[Discookie]

• Stepping into a dirty file is very unintuitive this way. The user clicks next step, switches from eg. a clean main.cpp to a dirty main.h, and no repr. steps are shown anywhere except the sidebar.
  There should be a popup or some other indicator that the steps are not shown because of the dirty flag. (Some extensions do this via a warning on line 0 column 0.)

• When the user presses accidentally types a character and then undo's it, CodeLens reappears, but the highlights do not.

• I'm not sold on the idea of hiding the highlights in the currently active file without a file switch. As long as there's no file switch or metadata reload, I think the diagnostics render fine.

There's also the consistency issue of a step's file being changed, so eg. changing main.h doesn't refresh the reports in main.cpp, and main.cpp's steps still render in an invalid way. That might be better suited for a separate PR though.

[Discookie]

As discussed in a meeting, because CodeLens doesn't move properly with the code on Linux, hiding the decorations in the current file is fine.

Once we find a solution for #35, the decorations won't need to be hidden in the currently edited file anymore.

Otherwise LGTM!

[csordasmarton]

If there are unpersisted changes in the current document we will hide the provided codelens items and bug step decorations.

[Discookie]

• Stepping into a dirty file is very unintuitive this way. The user clicks next step, switches from eg. a clean main.cpp to a dirty main.h, and no repr. steps are shown anywhere except the sidebar.
  There should be a popup or some other indicator that the steps are not shown because of the dirty flag. (Some extensions do this via a warning on line 0 column 0.)

• When the user presses accidentally types a character and then undo's it, CodeLens reappears, but the highlights do not.

• I'm not sold on the idea of hiding the highlights in the currently active file without a file switch. As long as there's no file switch or metadata reload, I think the diagnostics render fine.

There's also the consistency issue of a step's file being changed, so eg. changing main.h doesn't refresh the reports in main.cpp, and main.cpp's steps still render in an invalid way. That might be better suited for a separate PR though.

[Discookie]

As discussed in a meeting, because CodeLens doesn't move properly with the code on Linux, hiding the decorations in the current file is fine.

Once we find a solution for #35, the decorations won't need to be hidden in the currently edited file anymore.

Otherwise LGTM!

[csordasmarton]

If there are unpersisted changes in the current document we will hide the provided codelens items and bug step decorations.

[Discookie]

• Stepping into a dirty file is very unintuitive this way. The user clicks next step, switches from eg. a clean main.cpp to a dirty main.h, and no repr. steps are shown anywhere except the sidebar.
  There should be a popup or some other indicator that the steps are not shown because of the dirty flag. (Some extensions do this via a warning on line 0 column 0.)

• When the user presses accidentally types a character and then undo's it, CodeLens reappears, but the highlights do not.

• I'm not sold on the idea of hiding the highlights in the currently active file without a file switch. As long as there's no file switch or metadata reload, I think the diagnostics render fine.

There's also the consistency issue of a step's file being changed, so eg. changing main.h doesn't refresh the reports in main.cpp, and main.cpp's steps still render in an invalid way. That might be better suited for a separate PR though.

[Discookie]

As discussed in a meeting, because CodeLens doesn't move properly with the code on Linux, hiding the decorations in the current file is fine.

Once we find a solution for #35, the decorations won't need to be hidden in the currently edited file anymore.

Otherwise LGTM!

[csordasmarton]

If there are unpersisted changes in the current document we will hide the provided codelens items and bug step decorations.

[Discookie]

• Stepping into a dirty file is very unintuitive this way. The user clicks next step, switches from eg. a clean main.cpp to a dirty main.h, and no repr. steps are shown anywhere except the sidebar.
  There should be a popup or some other indicator that the steps are not shown because of the dirty flag. (Some extensions do this via a warning on line 0 column 0.)

• When the user presses accidentally types a character and then undo's it, CodeLens reappears, but the highlights do not.

• I'm not sold on the idea of hiding the highlights in the currently active file without a file switch. As long as there's no file switch or metadata reload, I think the diagnostics render fine.

There's also the consistency issue of a step's file being changed, so eg. changing main.h doesn't refresh the reports in main.cpp, and main.cpp's steps still render in an invalid way. That might be better suited for a separate PR though.

[Discookie]

As discussed in a meeting, because CodeLens doesn't move properly with the code on Linux, hiding the decorations in the current file is fine.

Once we find a solution for #35, the decorations won't need to be hidden in the currently edited file anymore.

Otherwise LGTM!

[csordasmarton]

If there are unpersisted changes in the current document we will hide the provided codelens items and bug step decorations.

[Discookie]

• Stepping into a dirty file is very unintuitive this way. The user clicks next step, switches from eg. a clean main.cpp to a dirty main.h, and no repr. steps are shown anywhere except the sidebar.
  There should be a popup or some other indicator that the steps are not shown because of the dirty flag. (Some extensions do this via a warning on line 0 column 0.)

• When the user presses accidentally types a character and then undo's it, CodeLens reappears, but the highlights do not.

• I'm not sold on the idea of hiding the highlights in the currently active file without a file switch. As long as there's no file switch or metadata reload, I think the diagnostics render fine.

There's also the consistency issue of a step's file being changed, so eg. changing main.h doesn't refresh the reports in main.cpp, and main.cpp's steps still render in an invalid way. That might be better suited for a separate PR though.

[Discookie]

As discussed in a meeting, because CodeLens doesn't move properly with the code on Linux, hiding the decorations in the current file is fine.

Once we find a solution for #35, the decorations won't need to be hidden in the currently edited file anymore.

Otherwise LGTM!

[csordasmarton]

If there are unpersisted changes in the current document we will hide the provided codelens items and bug step decorations.

[Discookie]

• Stepping into a dirty file is very unintuitive this way. The user clicks next step, switches from eg. a clean main.cpp to a dirty main.h, and no repr. steps are shown anywhere except the sidebar.
  There should be a popup or some other indicator that the steps are not shown because of the dirty flag. (Some extensions do this via a warning on line 0 column 0.)

• When the user presses accidentally types a character and then undo's it, CodeLens reappears, but the highlights do not.

• I'm not sold on the idea of hiding the highlights in the currently active file without a file switch. As long as there's no file switch or metadata reload, I think the diagnostics render fine.

There's also the consistency issue of a step's file being changed, so eg. changing main.h doesn't refresh the reports in main.cpp, and main.cpp's steps still render in an invalid way. That might be better suited for a separate PR though.

[Discookie]

As discussed in a meeting, because CodeLens doesn't move properly with the code on Linux, hiding the decorations in the current file is fine.

Once we find a solution for #35, the decorations won't need to be hidden in the currently edited file anymore.

Otherwise LGTM!

[csordasmarton]

If there are unpersisted changes in the current document we will hide the provided codelens items and bug step decorations.

[Discookie]

• Stepping into a dirty file is very unintuitive this way. The user clicks next step, switches from eg. a clean main.cpp to a dirty main.h, and no repr. steps are shown anywhere except the sidebar.
  There should be a popup or some other indicator that the steps are not shown because of the dirty flag. (Some extensions do this via a warning on line 0 column 0.)

• When the user presses accidentally types a character and then undo's it, CodeLens reappears, but the highlights do not.

• I'm not sold on the idea of hiding the highlights in the currently active file without a file switch. As long as there's no file switch or metadata reload, I think the diagnostics render fine.

There's also the consistency issue of a step's file being changed, so eg. changing main.h doesn't refresh the reports in main.cpp, and main.cpp's steps still render in an invalid way. That might be better suited for a separate PR though.

[Discookie]

As discussed in a meeting, because CodeLens doesn't move properly with the code on Linux, hiding the decorations in the current file is fine.

Once we find a solution for #35, the decorations won't need to be hidden in the currently edited file anymore.

Otherwise LGTM!