Thank you for effort, but ...

[elgreko5]

Hello aaFn,

First, I am really grateful for your effort (I am sure many others are, too), but I also think you're going in the wrong direction.

This will be a bit long-winded, but please bear with me.

Bookmarks are absolutely essential part of a browser, and FF not having functionality like in Show & Go parent folder today not functioning add-ons on its own, is a sign of idiocy of FF developers. As is e.g., moving FF Home Page icon from right to left while not rectifying a bug which caused losing your session tabs (at times I had more than 1.200 of them; after futile attempts to find a solution online, I found it myself - replace "_closedW" with "w" in sessionstore.js), not rectifying a bug which caused losing years of browsing history and shrinking it to a week, etc.

Bookmarks, tabs, browsing history, are very important data (I have about 80.000 of bookmarks and some are 15 years or so old), and should be treated as such. Imagine just the time needed to bookmark 80.000 links. Now imagine the time to open 80.000 pages.
Imagine the time to find, read, select, etc. and finally bookmark 80.000 pages... And that's just about invested time.

So, bookmarks database must be safe. If I understood correctly you made the effort to safeguard favicons, but not bookmarks. Also, I see quite a bit of options for favicons. And favicons are nowhere near as important as bookmarks, if at all?! Favicons are a little more than eye-candy, and candy is for kids. (Also, when there are no loaded favicons, or they are disabled, BSP displays a blotch - it should be something neutral or nothing at all.)
So, safety first.
Regarding safety of data - earlier today, your add-on's statistics showed more than 80.000 bookmarks - now it shows 79.672. I assure you I didn't delete 500 bookmarks today. (?!)

Efficacy second.
3-4 seconds wait to "Load saved state" or "" (yes, sometimes there is a message and sometimes there is not - wait is always there) on an i5@3.2-3.7 8GB SSD computer is ridiculous.
Sorry, but it really is. Long time ago, when 486 or 386 was a computer, I wrote a program for POS that replaced "program" that a company was using that needed 20 or more minutes just to display information about available quantity of goods for a customer. My program gave that info instantaneously and I didn't do anything exceptional - the solution was trivial. In fact, there was nothing to solve, it was just that my predecessor didn't have a clue.
3-4 s wait when clicking on "Bookmarks menu" to close it, too. I don't know if that is your fault or FF's - if it is FF's and can't be resolved, it is time for me to finally stop using it.
Outside of your add-on, said opening and closing is almost instantaneous.

Information design - user interface
Possibility of changing appearance of bookmarks menu in your add-on is good, but forcing some unknown set of font size and item spacing and whatnot (it seems the font is changed too - I lost my leading spaces indentation) a new default is not!

Also, when right-clicking on a bookmark, I do not get context menu anymore (I did yesterday or day before - I think I then was on a "beta update channel")?!

Regarding my 80.000 bookmarks - it seem there are unique 40.000 of them. At one time, when I was importing bookmarks from IE, idiotic FF duplicated them. Since then I have two Bookmarks Menus - second is in Other Bookmarks. Problem is that those bookmarks in that second Bookmarks Menu are updated automatically as I edit them in original Bookmarks menu. EDIT: Checked, there are 80.000 bookmarks, only folder was duplicated.
This doesn't have anything to do with your add-on, but what does is that when BSP is active, instead of second "instance" of Bookmarks Menu in Other Bookmarks, now there is only a settings icon without any text beside it?!

Also, today after I edited one bookmark and searched for it through BSP, BSP showed name like it was before editing?! Disabling and enabling BSP didn't help - situation resolved only after restarting FF.

One more suggestion - it would be pretty useful if you could add counter of tabs to your add-on.

So, thank you for effort, but there is much room for improvements, some of those - essential.

Kind regards

[aaFn]

Hello @elgreko5 ,

thank you for this extensive feedback
Quickly:

1. FF flawed approach
   Agree .. Purpose of this add-on is to disappear when FF incorporates the function. Not going to happen tomorrow I am afraid, given the duration this demand has been put on FF (basically since the beginning .. https://bugzilla.mozilla.org/show_bug.cgi?id=196509)

2. Bookmarks database
   Bookmarks are stored in Firefox (FF), not in the add-on. The add-on only has a cache of them (because of FF API limitations :-( ), but really the add-on main function is to display bookmarks and allow showing where they are

3. Safety of bookmarks
   If you lost 500 bookmarks in stats, this is due to FF. BSP2 as I said does not store them, only displays them. Only when you do an action like "delete" on BSP2 can you really remove something.

4. Favicons
   I didn't make efforts on favicon by choice, but because of https://bugzilla.mozilla.org/show_bug.cgi?id=1315616
   Basically, FF API doesn't give any add-on access to the favicon DB (contrary to Chrome) ... so I had to develop something to safely get them (no security exposure, no data privacy leak ..) by directly fetching them in background from the sites themselves. Meanwhile, an icon with a little blue T at bottom right to mean this is a temporary icon waiting for the fetch to occur, is displayed.
   You can turn this off in the options, and a standard "nofavicon" icon is displayed immediately.

5. Delay to load / efficiency
   A looong story .. again related to FF API limitation, which got even worse recently.
   See FF bug -> https://bugzilla.mozilla.org/show_bug.cgi?id=1396364
   and other details in Optimising loading #2
   The origin is awful coding and data mixing inside FF (I could cast an eye at it ...), with tags handled as bookmarks ..
   At the costs of many efforts, with building a local cache, I could accelerate all subsequent loads, after the initial one, by an exponential factor depending on the number of bookmarks .. but not the initial one
   The value I am at today is as far as i could get today with using FF API on local storage (which is far from efficient, it is storing json text, not binary, and it is slooow and hungry in memory..) and on the HTML display speed on FF, and I believe I got all the juice I could get from it.
   All the rest is marginal or so complex to code I do not want to enter into it, given the purpose of the add-on is to disappear one day when FF realizes the function.

6. Native bookmarks speed
   This is apparently using native OS primitives, or it has access to things add-ons cannot access, so display and storage access are much more speedy ... cannot do much about it, FF API is very limited.

7. Bookmarks menu speed to close
   Sorry, this is native FF and its speed of HTML display / manipulation, not much I can do about .. I already tried to optimize the HTML used to display to the simplest possible form fitting with needed features.

8. Right click context menu
   Should work. You might have memory issues though, FF has a memory hungry monster inside it, and it appears to be triggered when fetching favicons.
   Try disabling favicon fetching in the options.

9. Leading spaces
   You appear to be the second one saying that .. could you be more specific (a screenshot ?), I tested, this is displayed and working on my Windows or Linux machines.

10. Duplicate Bookmarks Menu inside Other Bookmarks
    Looks like this is something special ... what do you get when you hover on the "Setting" favicon ?
    Can you look at its properties ?
    Normally, this favicon corresponds to the "Most visited" bookmark item, or the "Recent Tags" one.
    I presume something is corrupted in your bookmarks structure.

11. Edit of bookmark not retrieved immediately
    Did you do that in BSP2, or in the native bookmarks toolbar ?
    It the first, it is possible that the update to the native DB got delayed in its notification back to BSP2 display front end, because of FF memory state or cpu load.
    Again, try disabling favicon fetching in case this is due to the memory monster ...

12. Count of tabs
    Do you mean count of open tabs on screen ?
    I have been toying with the idea of incorporating tabs search and handling in the add-on since the beginning, but for now decided to stay focused on bookmarks.
    Maybe one day I will add tabs indeed ..; I didn't mature enough on that decision yet.

Let me know if you have any additional feedback,
kind regards, aaFn.

[elgreko5]

Hello aaFn,

Thank you for your prompt response!

1. I am sure your add-on will be needed for years to come. The fact its been 16+ years and that they are writing garbage like - they lack the resources (16 y no resources for show and go folder?!) or that 95+% of their users have <2.000 bookmarks (like 2.000 is 20) assures me of that. If they, God forbid, wrote Total Commander, we would be able to find and open files, but not to know where they are and wouldn't be able to copy them...

2. I knew that, of course.

3. Of course, but:
   
   I know I saw BSP2 displaying more than 80.000 so it must be it's fault.

EDIT - FALSE ---> EDIT: Just found out what's going on - you don't count opened and pinned tabs in bookmarks, although FF does.

4. Knew that also. Maybe I wasn't clear - I think that favicons are the least important thing in your add-on. Again, they are next to useless. What's more, in a case like mine, I would be watching fetching them for years, and we don't need another memory hog provoking thing in FF...

5. If I understand you correctly - now even alice0775's Show & Go parent folder would have to have that delay?

6. No memory issues, I would have spotted them:
   
   Already disabled favicons yesterday because of uselessness.

Font has changed. Also, icons are not what you described but what I called blotches.

EDIT: Tested on another computer - the same.

10. Hovering gets me nothing. Right click doesn't get me properties but "New bookmark, New folder, New separator, Paste info" menu - the same as when I right click on Other Bookmarks.

EDIT: This is obviously bug in FF and another bug in BSP2.
Vivaldi correctly imported bookmarks without that clone folder.

11. Don't remember that, but I am pretty sure it wasn't memory or CPU - I would have noticed that.

12. Yes, I meant that.

Right click context menu still doesn't work.
Tried on another computer which never had BSP2 installed - doesn't work.
As I said, it seems it works on beta versions.

EDIT: Confirmed, another bug - on beta it works - tested on another computer with 150 bookmarks. After reinstalling regular FF doesn't work.
Yesterday I thought it worked in sidebar, but it obviously didn't. On beta it works in Library and Bookmarks drop-down menu, but on regular it only works in menu (where it is useless, don't understand why you even put it there).
I read you wrote that it can't work in sidebar because of FF limitations.
I also read what their devs wrote that if bookmarks are so needed someone should write add-on on its own. What they didn't though, is that it can't be done because they basically don't allow it.
Well, after wasting several hours in only these two days because of their bugs and limitations, I can only wish them happy landing:

Kind regards

[elgreko5]

Another bug in BSP2:
Sort By Name puts "1" before "_"

Also, it is so slow you can actually see as bookmarks are printed in new order.
There is also another anomaly (seems to be caused by slowness) which I am not sure I can explain exactly but it looks like if you do Sort By Name and immediately scroll down and click on a bookmark, sidebar bookmark list jumps.

EDIT: Another problem - sometimes "progress circle" freezes for about 15 s:

EDIT 2: Another bug:
It is possible to close FF without first closing Properties' box "moz-extension...", and it shouldn't be. (After next opening of FF, that box appears as a new FF window without close and other buttons.)
Also, in above mentioned box, "Path: ..." is unnecessary.

EDIT 3: Another bug:
Clicking on a light blue space beyond icon and text doesn't open folder although it should.

EDIT 4: Another bug:
Dragging tab to sidebar bookmarks and without dropping, dragging it back, outside the sidebar, causes sidebar contents to scroll to the left a little. Scrolling repeats with repeated crossing of the sidebar line.

[aaFn]

Hello @elgreko5,

3. Indeed, BSP2 only counts bookmark items (= bookmarks, folders and separators), as it is using the bookmarks interface https://developer.mozilla.org/en-US/docs/Mozilla/Add-ons/WebExtensions/API/bookmarks

4. Yes agree, this is a cosmetic thing, although many people like them because this helps visually find bookmarks, as an acceleration.
   Still cosmetic though, this is why I made an option to disable them.

5. In its time, alice0775's extensions (not add-ons, note the difference) were directly integrated within the FF interface, thanks to XUL technology. This is not the case anymore with the WebExtension API https://developer.mozilla.org/en-US/docs/Mozilla/Add-ons/WebExtensions, add-ons are totally external and can only interact with FF and the user through these API's and through HTML.
   So alice0775's extensions could be fast, as relying on FF interface itself and its OS level primitives.
   Not possible anymore nowadays ..

8. Re-reading you, I may be misinterpreting what you call "Right click context menu".
   Are you talking of the context menu inside BSP2 sidebar itself, or of the BSP2 submenu addition which is giving the path to bookmark as a string ?
   If the second, this is normal .. FF currently doesn't support that menu extension on the native Bookmarks sidebar, nor within the Bookmarks Library, as per -> https://developer.mozilla.org/en-US/docs/Mozilla/Add-ons/WebExtensions/API/menus/ContextType
   I am glad to hear that this is working on beta, that means this is coming :-) (FF people are progressing after all ...)

9. Ok, I spotted the problem .. it appears HTML is reducing all the leading spaces to only one space character.
   I will see if I can oblige it to respect all characters, and not condense spaces to only 1.

10. BSP2 only gets what this API call gives to it https://developer.mozilla.org/en-US/docs/Mozilla/Add-ons/WebExtensions/API/bookmarks/getTree, without interpreting it.
    I will check nothing is misinterpreted if a bookmark id is seen twice in the tree (they are supposed to be unique), but I do not think so, and in that case there's nothing I will be able to do.
    I'll keep you posted.

11. Ok, not sure what is the cause of delay. Did you make the edit from inside BSP2 using the "Properties" dialog, or from the native bookmarks sidebar (or Library) ?

[elgreko5]

EDIT: Had to put dots in front of numbers because they kept changing?!

EDIT: 3. Seemed strange I have so many tabs so I rechecked after installing Tab Counter - there is still discrepancy:
I have 498 tabs, Restore option shows 81894 items, while your statistics show: Bookmarks - 79744, Folders - 2169, Separators - 7. However I add and/or subtract, it just doesn't add up. Either FF errs, or BSP2, or maybe both, who should know :)

.4. Yes, they can be useful (they are really useful in tabs), it's just that I thought there are still wrinkles that should have been ironed out before spending time on favicons - e.g. changed fonts.

EDIT: .5. Well, that is dubious, FF calls them add-ons as well as extensions.

.8. Seems you skipped last paragraph - excerpt: "Yesterday I thought it worked in sidebar, but it obviously didn't. On beta it works in Library and Bookmarks drop-down menu, but on regular it only works in menu (where it is useless, don't understand why you even put it there).
I read you wrote that it can't work in sidebar because of FF limitations."

.9. No, as I already wrote - font has changed (as well as interlinear spacing) - it is obvious from the screenshots. Number of spaces is the same - three - it's because of different proportionality of different fonts. Also, look at the favicons - blotches. EDIT: Number of spaces stays the same, it is just that when printing, 3 spaces are printed as 1.
Correcting this, if you ask me, should be a priority. Also, it seems it would be trivial.

.10. Ok, that's why I reported it to you. I will eliminate that clone folder eventually, it's just that I'm not really looking forward to that endeavor...

.11. Can't be sure, I'm afraid. Hasn't happened since.

[aaFn]

Ok, thank you for the quick return.

On the total number of items, on my side, this matches perfectly = (number of bookmarks + number of oflders + number of separators) as counted by BSP2 exactly matches the counts in the import menu in the Library.
So I am not sure I can tell what is wrong in your case, but the double presence of the same tree / objects under different places could impact such counts maybe.

Changed fonts = I do not know which font is used by FF natively, they vary by OS and by version, so I have to try to find them manually (if available to the add-on):

• On all my Windows 10 machines (physical and virtual), BSP2 visually has the same font than the native Bookmark sidebar so I believe I found the right one there.
• On my Linux with Mate and KDE desktops, same (but by increasing the BSP2 font size to 14), so should be ok.
• Can't tell about IOS, I have no such machine.
• On Windows 7, the font is different .. It appears Windows 7 is using Tahoma, and that font is not availble to the add-on (or was not last time I tried). Plus, windows is Windows, hard to make any difference from within the add-on between Win 7 and Win 10 .. so I favored Win 10, which is more numerous now.
  So I confirm, things look different on Windows 7 (which appears to be your case from screenshots).

I do confirm also the collpasing of multiple spaces into one by HTML. I found the way to prevent that by using white-space: pre; in the CSS. This improvement will be in 2.0.54.

Now, on the other bugs you mention;

1. Sort by name
   This is using ASCII code, and using the Javascript native '<' operator.
   See http://www.asciitable.com/ : 1 character code is 49 (decimal), and _ character code is 95 (decimal)
   So it is normal that 1 is put before _.
   I do not know what character ordering the native bookmark is using, but that seems non natural. If there is a simple logic I can reproduce, ok to try to change, else I will remain with ASCII and Javascript sorting order.

2. Slowness = each bookmark move is sent to FF by this API call https://developer.mozilla.org/en-US/docs/Mozilla/Add-ons/WebExtensions/API/bookmarks/move.
   There is no way to send all or several moves in a bunch.
   Then, slowness comes for a part from this "one-by-one" send and FF API reactivity to the calls.
   In all truth, this is not slow with 2000 bookmarks .. but for a second part, things also slow down with high numbers of bookmarks, because the whole list of bookmarks in BSP2 is an HTML table and not an OS level list. With high number of rows, HTML gets slower in individual delete / insert operations (there is no row move operation on HTML tables).

3. Jumps = not really surprised by that, since bookmarks continue to move .. this is impacting where is the cursor I guess, or making things jerky on the screen movements because there are HTML refreshes in the middle (related with slowness / cpu intensive operations on the HTML table as said above).

4. Properties window
   This is an FF / Chrome security choice = Modal OS dialogs are not possible anymore.
   They can only be browser windows ... and so such windows can remain open when other windows are closed, like any other browser window.
   And of course, FF restores them when reopening, like any other browsing Window ...
   Can't do anything about it, sorry :-( .. talk with Chrome and FF about their choices for our good ... (some of those choices look like the Gnome 3 diktats, with these developers know better than us what is good and intelligent for us .. as a developer myself, I call that self-justified developer lazyness instead of intelligence)

5. Auto scrolls when dragging
   A complex handling. There is no way to disable or to control auto scrolling on drag. I tried and spent many hours on that, I can tell you .. and found FF bugs in the process btw !
   At the end, I resorted to a trick of removing scrollbars in some specific cases, when I am able to detect that a drag is coming from outside to inside, but once we come to deeper inside the sidebar, I have to restore the scrollbars, so as to allow the user to use natural auto-scrolling for placing the bookmark at the right place ...
   So what you observe is normal, and wishable since it allows somebody to auto-scroll where he/she wants to go (imagine a very deep tree, with bookmarks out of sight on right of the sidebar window .. need to auto scroll there on drag to be able to drop at the right place .. something not possible with the native bookmark sidebar by the way, since it doesn't have any horizontal scrollbar).

Hope that answers all the points you raised.
I am still looking details on item 10) to verify if there's something I can do or correct.

[aaFn]

Ah yes, forgot abouth what you call the "blotches".
This was the favicon used by the native Bookmark sidebar in FF 57 for bookmarks without fetchable favicons.
I just captured the screen and made it my own "nofavicon" icon.

It has changed now, probably around FF 62 (and I guess you could still call it a "blotch" btw .. myself, I don't like the new any better than the previous one ... both look ugly to me), so I will recapture it, and refresh in 2.0.54.

[aaFn]

So looking at 10) .. I believe the duplicate "Bookmarks menu" is represented as a "place:" URL bookmark.

These are special bookmarks for which I have to look at the "sort=xx" part in the URL to identify what they are (if xx = 8, then this is "Most Visited", if xx =14 then is this is Recent Tags and if xx =12 then this is Recent Bookmarks).
And then I have to emulate them, because the FF API does not return to me their content, and add-ons do not have access to "place:" now ("place:" appear to be live queries of sets of bookmarks and other objects, executed by the native Bookmark sidebar directly on sqlite DB .. kin of like a "View").

If I am correct, this case is too specific and rare, I won't enter into coding things to emulate it.
What I'll do is just show the URL as its title, instead of having no title, so that it is more obvious that it is a special thing, and also to verify my hypothesis ..
To be in 2.0.54.

[elgreko5]

I finally took to resolve the clone folder situation.
With situation being that when a bookmark is deleted from a clone folder, it is also deleted from original folder, I presumed deletion of a clone folder would delete original folder too. So I delved into bookmarks.json and manually excavated this:
BOF
,"children":[{"guid":"ho1czFgP8sJX","title":"","index":0,"dateAdded":1320785492187000,"lastModified":1550759278609000,"id":31850,"typeCode":1,"annos":[{"name":"places/excludeFromBackup","value":"1","expires":4,"flags":0}],"type":"text/x-moz-place","uri":"place:parent=menu________"}]"
EOF
The part immediately before removed text is this:
BOF
{"guid":"unfiled_____","title":"Other Bookmarks","index":3,"dateAdded":1313873732468000,"lastModified":1550971885685000,"id":5,"typeCode":2,"type":"text/x-moz-place-container","root":"unfiledBookmarksFolder"
EOF
After removal I was left with a .json file of 39,548,576 bytes.
Then I deleted the offending folder just to see if original folder will be deleted too, and, lo and behold - it wasn't! Did a backup of bookmarks and guess what - resulting file had a length of 39,548,576 bytes?!
Being used to so much more pain from Mozilla idiots, I couldn't believe it - I wasted only 20 minutes?!
Library's Restore option now shows 81949 items, while BSP2 shows 81947 (79771+2169+7+0), so it seems OK now. (On another W7 PC (with 165 bookmarks) there is also a discrepancy of 2, as well as on another W10 PC (17 bookmarks).)

Changed fonts

Yes, using W7 on this machine. I overrode interlinear spacing to 6 and it is much better now, although the font is uglier than original and, more importantly, less legible - it looks like condensed print on a 9-pin dot matrix printer.
Maybe you could ask Mozilla which is the original font or try Windows 7 system font - Segoe UI (it sure looks like it) and then make it an option? Just looked it up - W10 - 53%, W7 - 35%...

1. Sort by name
   See http://www.asciitable.com/ : 1 character code is 49 (decimal), and _ character
   code is 95 (decimal)

Yes, I know that (for more years than I care to admit :) )

   I do not know what character ordering the native bookmark is using, but that seems
  non natural. If there is a simple logic I can reproduce, ok to try to change, else I
   will remain with ASCII and Javascript sorting order.

In fact, natural sorting exists and it is not ASCII sorting. Also, sorting in BSP2 is not ASCII sorting, e.g. it puts "_" before "A" but after "1".
But, that's not the point here - the point is that BSP2 sorting is different from original and leaving it like this is inconsistency.

2. Slowness
   In all truth, this is not slow with 2000 bookmarks

Please don't get me wrong when I say - and if it was slow with 2000 bookmarks, it could be said - but it isn't with 20... I understand that there is nothing more you can do.

4. Properties window
   They can only be browser windows ... and so such windows can remain open when other
   windows are closed, like any other browser window.

Wow! Mozilla Geniuses at work!

Can't do anything about it, sorry :-( .. talk with Chrome and FF about their choices
for our good ... (some of those choices look like the Gnome 3 diktats,
with these developers know better than us what is good and intelligent
for us .. as a developer myself, I call that self-justified developer lazyness instead of intelligence)

Yes, but that's not just laziness, it is already a pathological state, so, no, I won't be talking to them :)

5. Auto scrolls when dragging
    but once we come to deeper inside the sidebar,
   So what you observe is normal, and wishable

It seems you didn't understand - it happens even when you have JUST crossed the vertical scrollbar (into the sidebar) and moved back across it. If you do the same SHORT movement again, it scrolls more.

Hope that answers all the points you raised.

It seems you skipped this: "Clicking on a light blue space beyond icon and text doesn't open folder although it should.". There is a screenshot above.

"nofavicon" icon.

If it's not too much to ask, could you just replace it with something neutral - less distracting.

Kind regards

EDITed because of buggy quoting

[elgreko5]

Another inconsistency:
Open in a New Tab in Bookmarks opens a new tab as last of opened tabs.
The command with a same name in BSP2 works differently - opens a new tab after current tab.
This is even better but name should be changed to reflect difference. Also, I think the best solution would be to have both commands in BSP2.

EDIT: typo

[aaFn]

Interesting ideas. And by the way, thanks for all the testing :-)

---

You are correct, from what I know, the redundant folder you had was not really a folder, but a view at the real folder.
Deleting it had no effect on the real one, it just deleted the view at it.
However, deleting an object inside was in fact deleting the true object, because this was a view.

---

On fonts, yes I do set Segoe on Windows, whatever version it is:

.fontwin {
  font: 12px "Segoe UI";
}

What you see in Windows 7 is the result of that setting.
Btw, if you activate traces in the options, you will see that under Windows 7. It is a checking that I make afterwards to see how FF transformed what I have set:

...
PlatformOs: win
Setting Windows variations
...
fontFamily = '"Segoe UI"'
fontSize   = '12px'
...

So it does not look like the native Bookmark sidebar is "Segoe UI" since we see it is different.
I do believe it is "Tahoma", but last time I checked I cannot access that font from the add-on.
What that means also is that Segoe UI does not look the same under Windows 7 and Windows 10 ...

If you want to have a closer look, go to C;\Windows\Fonts, and look at both "Segoe UI" and "Tahoma" .. I believe you should conclude like me that this is Tahoma they are using, but once again I can't ...

---

Sort by name ..
Something I didn't say = like the native Bookmark sidebar, I am sorting with ignoring caps.
So all Caps are first converted to Small letters, and then sorting is done.
This is why _ is before A .. =because in fact it is before a

I still don't know what the native Bookmarks sidebar is using as sort order, but I believe I am not too far from it.

You are correct though that as long as it is not exactly the same this is inconsistent. If we can find what it is, then I'll be ok to switch to it ...

---

Horizontal scrolling when crossing the vertical separation.

Well, I do not have that effect on any of my machines (Win10, Win7 or Linux).
Here is what is designed.
It happens exactly like that for me, and so should happen for you also (plenty of hours of tuning to get that ... because of FF bugs and insufficient API to manage these things ..):

• Take the (i) icon in the address bar or anything other to drag, from outside the sidebar.
• Drag by the right of the sidebar = when crossing the sidebar limit, an interdiction symbol appears as long as you are over the vertical scrollbar
• Continue towards left, as soon as you leave the vertical scrollbar, both scrollbars disappear (horizontal and vertical), and (normally) no scrolling happen.
  I am saying "normally", because in fact a scroll does happen, I cannot avoid it :-(, but I am catching it and forcing the frame back to its last remembered position before the scroll .. not visible "normally" because very quick ..
  And I cannot avoid it because I cannot control the auto-scroll area size and position ..
• At that point if you go back to right, nothing happens until you get out of the sidebar, at which time the scrollbars are restored after a timeout (another trick .. a bug in FF is removing one key event supposed to notify me that the drag exited ... so I have to rely on interruption of the supposed_regular [another bug ..] mouseover event to detect the mouse is no more on the sidebar and then I can restore the scrollbars).
• If you go deeper left instead, after some distance the scrollbars are restored to allow the wishable auto-scrolling function.

Apparently, this is not what is happening on your side.

Are you able to describe exactly in the same terms what is happening to you ?
And would you be so kind as to open the Browser console before doing the drag test (Ctrl+Shit+J), empty it with the trash icon at top left, and let me know if you see any message / error happening there while you do the experiment ?

---

Click on space before (or after) the "favicon + text" ... sorry I missed this point.

Correct, that does nothing by design, see -> https://github.com/aaFn/Bookmark-search-plus-2/wiki/Mouse-clicks-and-modifiers-summary-table-(as-of-2.0.46-)
I find this is an inconsistency in the native Bookmark sidebar = when you click on the space at left of a bookmark, nothing happens. So why should it when it is a folder ?

Doing so allows to select a folder like any other item, without any action, and then with the keyboard up/down arrows you can go up/down.
On the native Bookmark, to do the same, you have to click twice on the folder to be able to navigate with keys after that ... very cumbersome :-(

So in short, if you want an action, click on favicon or text. If you click outside of that, no action, apart from setting the cursor at that point.

---

Nofavicon icon = in 2.0.54, I am replacing by the new "nofavicon icon" of the latest FF version. Hope this is less distracting. At least this is aligning with their icon design change.

---

Open in new tab = have both forms of next and last

Indeed, that is an idea. However, mind that depending on an option of FF, open in new tab can be the next tab instead of the last tab .. so we cannot say that "Open in new tab" is always open as last.

What (short please) labels would you suggest for open next and open last ?

---

Thank you, aaFn

[elgreko5]

Interesting ideas. And by the way, thanks for all the testing :-)

Glad to help.

On fonts, yes I do set Segoe on Windows, whatever version it is:

You're right for Segoe UI and Tahoma. To be absolutely sure I am seeing the same, I checked with the help of layout.css.devPixelsPerPx=3.
Could you try with Arial - looks like a good substitute. I think it might be even better because of a tad bigger inter-character spacing than that of Tahoma.

Sort by name ..

Finding out the system would be easy, if not a little boring. Will do it when I catch some time.

Horizontal scrolling when crossing the vertical separation.

Tried on a W10 - the same thing happens.
On a W7 caught only this: TypeError: event.originalTarget.getAttribute is not a function[Learn More] tabbrowser.xml:1588:30

(plenty of hours of tuning to get that ... because of FF bugs and insufficient API to manage these things ..):

I know exactly what are you talking about - had my share of building UIs from scratch because standard UIs weren't neither ergonomic nor efficient. Would like to put that fella Marco from M. through the steps, bet he'd change his tune.

Are you able to describe exactly in the same terms what is happening to you ?
And would you be so kind as to open the Browser console before doing the drag test (Ctrl+Shit+J), empty it with the trash icon at top left, and let me know if you see any message / error happening there while you do the experiment ?

Tried, just one error (above). When I catch some time will do, or maybe better - make a video.

I find this is an inconsistency in the native Bookmark sidebar = when you click on the space at left of a bookmark, nothing happens. So why should it when it is a folder ?

I agree, it is.

Doing so allows to select a folder like any other item, without any action, and then with the keyboard up/down arrows you can go up/down.
On the native Bookmark, to do the same, you have to click twice on the folder to be able to navigate with keys after that ... very cumbersome :-(
So in short, if you want an action, click on favicon or text. If you click outside of that, no action, apart from setting the cursor at that point.

However, I think this would be even better: clicking on a left space selects item - folder or a bookmark, but clicking on the right space opens item. Notice that opening an item is far more used operation than selecting and right space opening is reducing mouse mileage. Also, it'd be consistent with FF Bookmarks thus eliminating frustrating failed attempts at clicking (which would result in selecting). Also, often there is no right space next to a bookmark, especially if it's in a subfolder, hence no chance of selecting.

While at it, I read what you wrote about multi-selection and that you postponed it's implementation, but - would it be easier to implement it first to work with keyboard?

What (short please) labels would you suggest for open next and open last ?

Open in a New Next Tab
Open in a New Last Tab

I see you shortened it in BSP2 to "Open in new tab", then above without an "a" :)

[aaFn]

Unfortunately, as said previously, I cannot distinguish at 100% and in a durable way between Windows 10 and Windows 7. This is Windows and that's all the add-on can safely determine.
So I am favoring Windows 10, by setting the correct one for Windows 10.

For horizontal auto-scrolling, the error you catch is not coming from BSP2 .. not sure what it means or if this is related to what is happening.
I'll wait that are you are able to describe things more in details then to see if I can figure out what could be happening.

Currently, click on right or left of (favicon+text) is the same, as seen from the code, with no easy way to distinguish. I'll see if there is a more obvious way. Meanwhile, the pointer anyway is changing when hovering on things, to indicate where an action will be triggered (= hand pointer), and where not (= simple pointer).

Multi-selection requires some deeper structure change, so really that work needs to happen first, before I can make it all happen. This is why I delayed it for some time where I will be in vacations .. :-)

Thank you for the suggestions, I will look at implementing that then, without the "a" probably indeed to make them shorter ..

Btw, 2.0.54 will be out tonight.