EdgeViewer - Modern HTML viewer plugin

[AntonyD]

For me, the main thing when working with this type of documents is the instantaneous jump
from viewing to editing. Since sometimes I can copy something only by switching to such a mode.
And wasting time on opening/scrolling can be neglected.... For me...

[larry99]

I only now tried this plugin, installed the most recent version from Github releases and the Evergreen Standalone Installer (was necessary though the Edge Browser comes with the OS).

I tried it with a HTML page downloaded with the Firefox "Save Page WE"-Plugin but I always get an error.
With the default ini settings (no switches):
"Die Server-IP-Adresse local.example wurde nicht gefunden" (Server IP local.example not found)
With the settings recommended somewhere in this thread:

Code: Select all

Switches= --disable-gpu-shader-disk-cache --disk-cache-dir=nul --disk-cache-size=1 --disable-background-networking --disable-component-update --disable-breakpad --disable-crash-reporter --disable-databases --disable-default-apps --disable-file-system --disable-local-storage --disable-remote-fonts --inprivate --no-pings --proxy-server="0.0.0.0:0" --enable-features=OverlayScrollbar
UserDir=%temp%
CleanupOnExit=1
OfflineMode=1

I get a HTTP 403 error "Der Zugriff auf local.example wurde verweigert" (access to local.example denied), perhaps because 0.0.0.0 is blocked after a vulnerabillity was found. When I remove the proxy switch I get another 403 error message:
"Sie besitzen keine Benutzerrechte zum Anzeigen dieser Seite" (no user permissions to show this page)
When I change OfflineMode back to the default "0" I am back to the initial error message (not found).

I did two control checks:
1. Open the "Save Page WE" page with the Edge Browser that comes with Windows: Opens just fine, looks really good
2. Save a Page from this forum with the standard Firefox "Save page as..." Dialog: Works but looks horrible (no CSS? no JS?)

But I depend heavyly on the pages from WE, since there everything is in one file, including images, css...
I use these pages as part of my internal documentation and they work just fine in any modern browser, just not within TC.

OS is a fairly fresh Windows 10. Any ideas?

[larry99]

Sorry, a few seconds after I posted the former message I noticed that the pages that gave error messages were all located on an UNC path.
After I copied one of them to a local drive (c:\temp) it worked just fine!

Still, since most of the time I do not set drive letters for network shares but access them through UNC paths: Is this possible to fix?

[rg_software]

Sorry, a few seconds after I posted the former message I noticed that the pages that gave error messages were all located on an UNC path.
After I copied one of them to a local drive (c:\temp) it worked just fine!

Still, since most of the time I do not set drive letters for network shares but access them through UNC paths: Is this possible to fix?

Will put into my todo list. I don't really know why it doesn't work with UNC out of the box. Just for the record, do I understand correctly that the issue is with paths that look like

Code: Select all

\\server\path\file.htm

?

[larry99]

Yes, exactly like that.

Thanks for looking into it!

[rg_software]

Sorry, it seems that UNC is a real pain. It must be mapped to a local drive letter. Even if I do it automatically, the share might (and commonly is) password-protected, so in this case it must be necessary to request username & password, quite a lot of work. On the other hand, if TC already accessed it, user credentials are already there, so maybe it is still doable.

[larry99]

I don't think mapping to a drive letter is a good idea. On the other hand authentication is never a problem since I must have already (at least) read access to see the file in TC. So I don't think the 403 I get for UNC paths is a Windows file system "permission denied". I can open it with the standard lister as plain text without any permission problems. I guess it is a security feature of the fake web server(?) used in your plugin. I had hoped that it is just a configuration setting to allow this sort of network access.

Perhaps you can temporarely copy it to %TEMP% if the file is on an UNC path and open from there? But then you should also clean up when the viewer is closed.

[rg_software]

Perhaps you can temporarely copy it to %TEMP% if the file is on an UNC path and open from there? But then you should also clean up when the viewer is closed.

The core issue with all "copy" approaches already suggested here before is relative links. If you have HTML with pictures, I need to copy basically all of them preserving the original directory structure. Currently I map the file location as a "virtual host", but this approach does not work with UNC at all, as I learned.

[larry99]

You are right. This kind of problems is exactly why I use the "Save Page WE" Firefox plugin. With this everything is in one file. But I understand that not everyone uses this or a similar tool.

But still: Isn't it better than the current situation? Now all you get is an error message. With the copy of just the HTML file you have either a perfect solution if you have all-in-one HTML files or at least the pure, perhaps ugly content of the HTML page itself -- which is still more than nothing.
And as I said, I would only do the copy if the path is an UNC path, so no change for files that work anyway.

[rg_software]

But still: Isn't it better than the current situation? Now all you get is an error message. With the copy of just the HTML file you have either a perfect solution if you have all-in-one HTML files or at least the pure, perhaps ugly content of the HTML page itself -- which is still more than nothing.
And as I said, I would only do the copy if the path is an UNC path, so no change for files that work anyway.

Yeah, agree. I am sorry it takes so much time (kind of extended busy period), I still have hope to invent something better, but if everything fails, it will do. After all, we already have a similar situation, e.g., if you press F3 on HTML file inside an archive.

[larry99]

Thanks for working on it! I know very well that progress needs time, especially if you are very busy. But I am certainly looking forward to a solution...

[rg_software]

Hi all,
Please check out the latest update of the EdgeViewer plugin.

Here is the summary of changes:
- Any string can be searched inside open documents (previously, e.g., apostrophes did not work).
- URL shortcuts pointing to local files work now.
- Zoom can be controlled both by Ctrl + mouse wheel and Ctrl+/Ctrl- keys. Furthermore, zoom level can be optionally preserved between subsequent document openings via KeepZoom ini setting (but NOT between TC reloads).
- Directory view works even for "special directories" and symlinks (crashed previously).
- UNC paths work via file copy to a temp location, meaning it won't work for folders and won't display linked resources.

[AntonyD]

2rg_software
What about https://github.com/rg-software/wlx-edge-viewer/issues/20?
Do you plan to move on for it?

[rg_software]

Yes, I will take a look. Technically this part is inside a third-party JavaScript HTML component, which I am less comfortable with than with my own code.

[rg_software]

2AnthonyD:
Turned out to be far more challenging, but please try downloading this new mhtml2html.min.js version. Replace the old one in assets\mhtml and restart TC.