Scroll without having to hold middle mouse click à la Firefox

[Herkules97]

On Firefox when I click or hold mmb it will start a scroller that moves up or down based on where I keep the cursor.
I am thinking of using this behaviour in EBV to scroll a result list for de-duplication because it's 70K files and manually scrolling that much will be a pain.
Granted I could manually scroll maybe 5K every day and eventually I'd have done it all.

I found settings that mention the wheel, but I changed them to anything else that is about scrolling and none of it changed anything. Idk if there is a setting I have to enable for EBV to modify the mmb behaviour within EBV. For example setting any of the ones with wheel in them to "Scroll page" does nothing, the hold-to-move-around remains the only action.
Holding is tedious for any button, particularly the mmb. Plus it doesn't stay at a consistent speed, it's dependant on how I move it. I can't move smoothly.

[void]

I will consider an auto scroll on mbutton like Firefox.

Thank you for the suggestion.



For now, there should be third party tools out there for scrolling on mbutton.
To disable the builtin Everything mbutton scrolling:

• In Everything 1.5, from the Tools menu, click Options.
• Click the Advanced tab on the left.
• To the right of Show settings containing, search for:
  mbutton
• Select: mbutton_action
• Set the value to: None
• Click OK.

Another option is to use the down key in the result list.
But you can't really control the speed without changing your keyboard repeat delay.



Another option is to export results to html and scroll the results in firefox.

[Herkules97]

Another option is to use the down key in the result list.
But you can't really control the speed without changing your keyboard repeat delay.

Arrow key down? Ctrl+down does prevent selection touching.
Never done it before or don't remember, but I tried ctrl+page down and that also does the same thing.
Page down wouldn't work, skips too much but arrow key might work..Maybe 40% slower than manual scrolling but doesn't require doing it.
Firefox method is better, as it doesn't require further input. But I did not think about ctrl+arrow key.

Is mice being able to be blown with air to reach high speeds actually scrolling? Or is it just a damage/modification to any mice where it no longer scrolls and is somewhat detached from the mouse..

Anyway I select 70K files in a 222k file list, then I scroll staring near the middle of the list for any abnormalities.
This list has mostly 3 total duplicates across 3 devices. So when the list reads the same while I scroll manually, it indicates the files scrolled all have 3 duplicates.

The thing with Firefox's is that it scrolls faster or slower depending on how far away it is from the starting scroller.
It could be that I'd use it to scroll 20% faster than I do manually..Without the tedium of doing it manually and it should be smoother too.
Or I find out I can actually do it 50% faster or maybe more. I have to stare and register it all so can't be too fast..
The percentages are not known to me in Firefox, just guesses.

Another option is to export results to html and scroll the results in firefox.

I haven't tried an export where I can select one of 3 copies, but I tried one where I only exported what I Intend to delete and I'm a bit blinded in the middle of my vision now from just having looked at the highlighted paths for 2-3 seconds(Lasts a few minutes I think)..
It might both be the pink in highlighted text and the white background. Maybe both can be modified in about:config..but can't select and de-select via Firefox anyway so little point trying this method.
EBV is optimal, at least for how I do it.

There are de-duplication programs but I don't trust them, at least not Digital Volcano's after it told me very different files are duplicates and I only knew that because I still manually scrolled some of those lists to be sure. Maybe something was corrupt somewhere, Idk. EBV is almost entirely trustworthy and if I'm manually scrolling anyway why not use EBV.

There are very few files that remain in the dbs after deletions..So if say I've deleted a file on 2 devices and they happen to both remain in the dbs, if I have 3 devices with the copies EBV will claim I have 3 duplicates when I only have one real copy left.
Rebuilds and ctrl+f5 don't fix at least broken names entries. Maybe a rebuild would fix entries of files that no longer exists and it failed to remove them from the db at the time of deletion. Wish there was a way to tell EBV to check if a file is there or not, ctrl+f5 probably should but then something else is wrong as it has done nothing on such a file..Maybe copying any file into the same path with the same filename then deleting it would fix it..hmm.

This is too vague of a memory now for me to be sure of anything written: I've also rarely discovered it doesn't have the live timestamps of files, Idk how widespread it is. It might only be for date modified..I think I've seen multiple instances have the same wrong timestamp of a file so maybe it's not a corruption. I think I used 7-Zip to confirm the live timestamp..Bulk Rename Utility was involved too. I did write about it when it happened so maybe I could look back at some point on my notes..But if times are wrong on some files but not other copies then it wouldn't show up as a duplicate anyway.

[void]

Arrow key down?

Yes.


If you just want to select 70k files, what about:
Click the first item.
This will set the selection mark.
Use the scroll bar to scroll roughly 70k files.
While holding down Shift, click on a result.
Everything should select the entire range from the first click.

To make large adjustments to get closer to 70k selected files:
Use the scroll bar to scroll to the rough position.
Hold down Shift and click on a result.

Once you are close to 70k, to make small adjustments.
Hold down Shift and press Up/Down/Page-Up/Page-Down.

[Herkules97]

If you just want to select 70k files, what about:
Click the first item.
This will set the selection mark.
Use the scroll bar to scroll roughly 70k files.
While holding down Shift, click on a result.
Everything should select the entire range from the first click.

I think you're misunderstanding(or am I misunderstanding your reply?), I already select the files from the start. Search with the dupe: commands, sort by path to select everything on the device I want to delete from, then sort whichever I use to get dupe lines. Recently it's been date-created-descending but any of the ones I have dupe: for. Datetimes are more accurate than size and name probably. At least created, with yt-dlp and mtime on the modifieds are from server so are less useful.
This highlights what I intend to delete, I then scroll to find out if there are any abnormalities. Files that only exist on 2 devices instead of the 3 for example. In the case of the 70K/220K list they all should have duplicates. But all search results aren't as..uniform.
For many of the lists for these 3 folders it's all been 100% duplicates I can delete. Maybe it's safe to presume it will be the same for this one, but I'd like to be more sure than that so I scroll.

An image could help, here are opus files I am waiting to index with the media property instance first before deleting..I was too lazy to get the 220k export loaded..Idk what it's meant to help with, maybe it can be easier seeing it visually just.
All I2 files selected out of the 3 devices, then sorted by date-created-descending to see in motion where there are abnormalities. During manual scrolling it's uniform if there are only 3 total duplicates. When hitting something like the middle here with one device having 2 copies, it's obvious.

Album version https://imgur.com/a/AnF5u1z

[void]

gotcha, results are already selected and your visibly want to inspect each item.

What about:
Enable Tools -> Options -> Results -> Automatically scroll view.
Make your window larger or hide some columns so there's an empty spot in the results where there's no columns.
Press and hold down Ctrl
While Ctrl is down, left click on an empty spot in the result list.
Move the cursor below the result list.
Keep the cursor out of any columns (Otherwise your selection might change)
The view will auto scroll.
Move the the cursor further below the result list to increase the auto scroll rate.
You can let go of Ctrl once you have left clicked.

[Herkules97]

gotcha, results are already selected and your visibly want to inspect each item.

What about:
Enable Tools -> Options -> Results -> Automatically scroll view.
Make your window larger or hide some columns so there's an empty spot in the results where there's no columns.
Press and hold down Ctrl
While Ctrl is down, left click on an empty spot in the result list.
Move the cursor below the result list.
Keep the cursor out of any columns (Otherwise your selection might change)
The view will auto scroll.
Move the the cursor further below the result list to increase the auto scroll rate.
You can let go of Ctrl once you have left clicked.

What does the "Automatically scroll view" thing even do?
I drag toward edges and with it off and on it still scrolls? The description is a bit confusing with "periodically" as well.

This is after writing the italic stuff further down, kept all that in for posterity.
I realise that the manual scrolling always scrolls the same distance. It's probably why it's often not smooth with other scrolling methods, because they're dynamic and the amount depends on me. Maybe sometimes I am scrolling 1 row per whatever and other times 5 rows per whatever. Idk what measurement to use. Firefox-type scroller could be set to a smooth amount, but would still face the issue of needing to re-adjust any time I have to make a stop because I spotted something.
I also tested dupe-min:4 to find out if scrolling 3x rows is universal or if it's because I mostly have 3 duplicates per file...scrolling 9 rows per scroll had a similar effect as 7 had for dupe-min:3. Seems it has to match the amount of duplicate rows per group.
I've never realised this before, maybe because I've had it on 3 rows per scroll and it's slow enough that it works well enough? Or because most of the time I have 2-3 duplicates in any given list?

A Firefox-type scroller could have 2 modes, one is dynamic and the other is a set distance like the mouse setting for row per scroll is.
But the second mode could also be achieved with autohotkey by binding say page down to scroll wheel down? Then you can hold the page down button to simulate it?
Does this mean there is no point in implementing something like that? Windows' rows per scroll goes up to 100 and down to 1 so it's not like I need more. 15 is already a bit much.

I do think for larger lists like 220K entries, it's worthwhile to have a more automatic method, my right hand is not a fan of all the scrolling at least. If the AHK method works, then that could help a lot with that.
I do wish I could do 9-15 but I don't have enough pixels for that I think. Maybe if I had a 4K screen.


I'll keep this long thought process below but will write a "shorter" version detailing the relevant stuff above
I was able to try the empty space method and it doesn't work well. Way too choppy.
MMB is more useable and that is worse than only dragging the wheel. There are spots where it does smoothly scroll but it's maybe 20% of the time. 80% of the time I can barely see.
It could depend on whether or not files are selected, what the current sort is and maybe the properties of files too. Idk if there is a way to increase its processing ability, if say it's GPU-reliant.
The Firefox-type scroller could have the same issue. I might be stuck doing it manually if so. Maybe manual scrolling hits a sweet spot of speed and smoothness and any speed increase will quickly go to too laggy territory. I also don't know if increasing the smoothness would fix the issue, I'd have to try it out first.
Reducing the window size and shrinking more columns did result in better performance..But there are still large areas where it just stops being smooth and such a small window makes it hard to spot differences anyway, a lot less is visible at any given time.

But if there was a way to create empty columns spots, that don't interact with selection so full row select still works, and somehow helping against smoothness could be an alternative to a Firefox-type scroller. IF such a scroller doesn't rely as much on the same smoothness. Using https://store.steampowered.com/steamawards/2024 because Steam pages lag hard on Firefox..It does show it's not smooth scrolling. So I think it's smooth in Firefox because the page is smooth. Just making sure that's the case rather than assuming.
So I think the issue with this empty space method is that it's just too laggy and the scroller would have the same issue.

From what I can tell, manual scrolling acts like a nerfed page up/down. It scrolls a lot less, but it scrolls the same distance.
I guess this means there is nothing I can do to make it faster besides maybe increasing the gap each scroll does which I think is inside the mouse properties.
..And it seems like 6(translated it says rows, so 6 rows per scroll), I was on 3, is better. 9 is too much and can risk missing stuff. It also seems like it has to be in thirds.
When I put it on 7 I couldn't see, similar to the empty space method most of the time.
The empty space method, if you can hit it right could scroll in thirds then..But any time something stands out I'd have to stop to check it out and then again have to figure out the sweet spot. Manual scrolling has the same distance every time and doesn't need precision.
12 sort of works so 9 should too and I guess it was just in a laggier spot, 15 causes too much lag and stuff doesn't render in time.

Firefox-type scroller would not require input until reaching an oddity, but then you'd have to figure out what the sweet spot is again.
If such a thing was introduced, it could have a second mode for set distance to ensure it works the same as the scroll wheel.
Alternatively..autohotkey scroll wheel down action to a different button? Temporarily hijack the page down for scroll wheel then you can just hold it down and never need to figure out the sweet spot after every stop.

[void]

What does the "Automatically scroll view" thing even do?

auto scroll view off: only scroll when you move the mouse.
auto scroll view on: periodically scroll.



Everything doesn't use pixel scrolling, it scrolls by lines.
Scrolling will not be smooth.

Enabling Tools -> Options -> update_display_after_scroll might help "smoothness".



To adjust how many lines Everything scrolls on mouse wheel:

• In Everything 1.5, from the Tools menu, click Options.
• Click the Advanced tab on the left.
• To the right of Show settings containing, search for:
  scroll
• Select: wheel_scroll_lines
• Set the value to: 3
  (where 3 is the number of lines to scroll, use -1 for a full page)
• Click OK.

If you want to do something with autohotkey I recommend just sending
UI_ID_RESULT_LIST_SCROLL_UP or UI_ID_RESULT_LIST_SCROLL_DOWN for the most control.

I was able to try the empty space method and it doesn't work well. Way too choppy.

If I added mbutton scroll like firefox, it would behave the same way as the empty space method.
It will not be smooth because Everything scrolls a line at a time.

But if there was a way to create empty columns spots, that don't interact with selection so full row select still works, and somehow helping against

There's a small edge on the left of the icon that doesn't select files, but that only works if Name is the first column.

Tools -> Options -> Advanced -> listview_margin_left can add a margin to the left side where you can drag-select.

[Herkules97]

If you want to do something with autohotkey I would just send
UI_ID_RESULT_LIST_SCROLL_UP or UI_ID_RESULT_LIST_SCROLL_DOWN for the most control.

If I added mbutton scroll like firefox, it would behave the same way as the empty space method.
It will not be smooth because Everything scrolls a line at a time.

Hmm..But it would be automatic, so no need to do a bunch of scrolling yourself. Seems the second best is this empty column method, it gives similar control as the Firefox scroller. Add some more complexity to it and you could probably automate holding ctrl+left click to not have to do it but ahk method will probably be less tedious to work with so that will be used for my first real de-duplication attempt with one of these changes.
I've already done some de-duplication recently so it won't be right now that I try anything.
Idk if one has to make it more complex by using the command stuff, but I suppose it can give more control.
Wheel down re-bind is probably going to achieve a similar effect, not that I've tried it before.

At least this got me to change the scroll from 3 to 6 in Windows so now I have to scroll less in general.

Just a curiosity, if scrolling too much lags..Where does the issue lie?
Weaker single or multi-core CPU performance?
Weaker GPU?
Weaker RAM?
As in if I upgraded one or more parts, which would help against it..
I'm not upgrading anytime soon, but for this system I bought low/mid-end stuff so there is improvement to be had in the hardware whenever I decide to splurge there. I found out I don't even have 12 cores, I have 12 threads(Maybe I knew when I bought and forgot, more likely I didn't care about it at the time and "suffer" for it since). Buying based on the iGPU was a mistake, I also thought most CPUs don't have iGPUs to render at least the desktop but they do so..A near-pointless CPU purchase. There was a few days where I didn't have a dGPU(I did but I never tried my old dGPU), but I could've waited instead of basing my entire purchasing decision on something that benefitted me for a few days especially as I don't usually get new parts unless old stuff fails in some way and that can be a few- or several years.

Tools -> Options -> Advanced -> listview_margin_left can add a margin to the left side where you can drag-select.

Seems to work as expected, no need to mess with the columns to get it..
But why is it only left? For the right one you have to maximise the window or at least the horizontal and then reduce column sizes to get it.
If there was one for right you wouldn't have to mess with columns. Much worse if you have a lot of columns, though for my monolithic instance I don't.

Alternatively you could have column sets? In Process Explorer I can enable whatever columns then save them and load them back if I change stuff some more.
I don't remember if it works as expected but it's there at least.
Saving from the ini would work but requires a restart of the whole instance every time.
I do use Clipdiary so I could add the whole thing to it, label it something useful then paste it in after I'm done messing with columns and exited the instance..Not ideal but works.
Copying the entire .ini and then replacing the newer version later could change more than wanted so the Clipdiary path is better.

[therube]

(

There are de-duplication programs but I don't trust them, at least not Digital Volcano's after it told me very different files are duplicates

Will depend on your definition of "duplicate" & settings used in Duplicate Cleaner.

If you're doing a duplicate Content search, you'll get true duplicates.

Other modes; Image, Audio, Video, all use various algorithms that may find true duplicates & may also find "similar" duplicates.
And then may also have settings to check "tags" or "data" or both... So yes, it is not unlikely that very different files could end up as being "duplicates".

Same would be true for any duplicate file finder program.
)

yt-dlp and mtime on the modifieds are from server so are less useful

Not sure, but I would think yt-dlp would have a setting to set the download file data/time to current.
--no-mtime, looks to do it.

Files that only exist on 2 devices instead of the 3 for example.

dupe-min:<count>

Find duplicates where the number of duplicates is greater than or equal to the specified minimum count.

Examples:
dupe-min:3 dupe:size

(would actually find a minimum of 3 duplicates, not necessarily on 3 separate devices - so you'd have to verify that part
- oh, you're already aware of dupe-min:, i see)

[void]

Just a curiosity, if scrolling too much lags..Where does the issue lie?

It could be the gathering of unindexed information for display.
Please try showing only indexed information and see if the issue persists.

Gathering unindexed information is done in background thread, but this can still freeze the system.

If you are showing only indexed information and Everything freezes while scrolling, its likely Everything was paged to disk.

No hardware change is going to help here.

But why is it only left?

A right margin is too complicated.

Consider disabling full row select under Tools -> Options -> Results.



Column sets will be added in a future version.
For now, you can set columns with bookmarks.

[Herkules97]

It could be the gathering of unindexed information for display.
Please try showing only indexed information and see if the issue persists.

Gathering unindexed information is done in background thread, but this can still freeze the system.

If you are showing only indexed information and Everything freezes while scrolling, its likely Everything was paged to disk.

No hardware change is going to help here.

I was trying to replicate the issue, not that I remember it clearly(One issue was overlapping entries making it impossible to see what was needed in scrolling)..But after doing it for a few minutes I remembered I restarted foobar2000 after that and now it's too smooth to at least overlap like it did before.
Yet another time I forget foobar2000 has been on for 2-4 days and is the cause for the lag.