I'd propose building each line and then table.concat(tbl, "\n") to avoid extra concatenation.

@hius07

I've tried using the concat "\n" but the result is inconsistent..?
I'm not sure why that would be so, since I'm matching the same structure as the one with the concat "\n"

Can you help me out? I'm not very familiar with lua

local function format(booknotes)
    local tbl = {}

    for ___, entry in ipairs(booknotes) do
        for ____, clipping in ipairs(entry) do
            if booknotes.title and clipping.text then
                table.insert(tbl, booknotes.title)

                local header = T(_("- Your highlight on page %1 | Added on %2"), clipping.page,
                    os.date("%A, %B %d, %Y %I:%M:%S %p", clipping.time))
                table.insert(tbl, header)

                table.insert(tbl, "")
                table.insert(tbl, clipping.text)
                table.insert(tbl, "==========")

                if clipping.note then
                    table.insert(tbl, booknotes.title)

                    header = T(_("- Your note on page %1 | Added on %2"), clipping.page,
                        os.date("%A, %B %d, %Y %I:%M:%S %p", clipping.time))
                    table.insert(tbl, header)

                    table.insert(tbl, "")
                    table.insert(tbl, clipping.note)
                    table.insert(tbl, "==========")
                end
            end
        end
    end

    local content = table.concat(tbl, "\n")
    return content
end

The outputs; notice how some titles are in the same line as the divider ==========

My experience with Anki
- Your highlight on page 1 | Added on Thursday, June 27, 2024 08:38:52 PM

I prepared some questions that are primarily for reflection, feel free
==========
My experience with Anki
- Your highlight on page 1 | Added on Thursday, June 27, 2024 08:38:54 PM

Do you have a special philosophy/methodology behind
==========
My experience with Anki
- Your highlight on page 6 | Added on Thursday, June 27, 2024 08:38:58 PM

ease hell](https://readbroca.com/anki/ease-hell/)
- Progress
==========KOReader Quickstart Guide
- Your highlight on page 1 | Added on Thursday, June 27, 2024 08:35:36 PM

You can access the complete user manual from
==========LinuxCommand.Find and replace
- Your highlight on page 1 | Added on Thursday, June 27, 2024 08:39:19 PM

-rli "1.0.5" . | xargs -i@ sed -i "s/1.0
==========
LinuxCommand.Find and replace
- Your highlight on page 1 | Added on Thursday, June 27, 2024 08:39:57 PM

the matching string with [[LinuxTool.grep]].
`r` recursively, `l` print name of files with matching strings, `i`
==========
LinuxCommand.Find and replace
- Your note on page 1 | Added on Thursday, June 27, 2024 08:39:57 PM

This is my note
==========

At the beginning insert an extra "\n" (so it'll turn into two). Or just an empty string if you only want one.

That is, if you do concat with \n you get line1\nline2\nline3. Attach that to another set and you're missing a newline.

Oh, I've got it now! It actually needed to be added to the end part :)
To avoid having a newline at the start of the file and output the same content as before the fix.

I lost focus of the fact that the format function is called separately for each book, instead of the entry i'm iterating through being a continuous stream.

tysm, fixed ! :)

@Mochitto and @pazos In order to include the features in #12106 (now with conflicting changes) you need to change the following in line 19 (OOM fix version). I have added the author as "Unknown" if the author is not presenting preserving Kindle syntax as mentioned in @hius07 in #12106

- table.insert(tbl, booknotes.title)

+ if booknotes.author then
+     table.insert(tbl, booknotes.title .. " (" .. booknotes.author .. ")" )
+ else
+     table.insert(tbl, booknotes.title .. " (Unknown)" )
+ end

And the following in line 30 (OOM fix version)

- table.insert(tbl, booknotes.title)

+ if booknotes.author then
+     table.insert(tbl, booknotes.title .. " (" .. booknotes.author .. ")" )
+ else
+     table.insert(tbl, booknotes.title .. " (Unknown)" )
+ end

Also @Mochitto and @pazos, I have noticed something, that is worth fixing. The time in the added export are in local time. But I assume Kindle uses GMT/UTC timestamps so when importing to myclippings, it auto-converts them to local time based on browser timezone/system time. I have illustrated an example here -

As you can see, 05:30 was added to the timestamp given in the file ( I am from GMT + 5:30 time zone )

I am not sure what is the format of the time in clipping.time , but it's best if you can somehow convert that to the UTC time zone matching Kindle syntax for seamless importing by clippings.io. I think Koreader already has the time zone information, so you need to use that for the conversion. Maybe @Frenzie or @pazos can help with this issue better.

I am not sure what is the format of the time in clipping.time , but it's best if you can somehow convert that to the UTC time zone matching Kindle syntax for seamless importing by clippings.io.

It is converted exactly to the Kindle format:

I am not sure what is the format of the time in clipping.time , but it's best if you can somehow convert that to the UTC time zone matching Kindle syntax for seamless importing by clippings.io.

It is converted exactly to the Kindle format:

koreader/plugins/exporter.koplugin/target/my_clippings.lua

Lines 21 to 22 in 6b81b51

	local header = T(_("- Your highlight on page %1 | Added on %2"), clipping.page,
	os.date("%A, %B %d, %Y %I:%M:%S %p", clipping.time))

The problem is outlined above. What I meant is I do not know what object clipping.time variable is ( is it a formatted string, datetime object, Epoch time integer ... ) . The format there is fine, but the time is in local time zone. Which is not how Kindle exports it. So when uploaded to myclippings.io (as intended for this plugin by @Mochitto), the timestamp gets modified to wrong value ( It always assumes the given time is in UTC and stores accordingly, shows the user in local time by calculating from browser time ). See the images I attached in my comment. That might help you understand the issue.

What does myclippings.io import from the Kindle?
My Clippings.txt or something else?

What does myclippings.io import from the Kindle? My Clippings.txt or something else?

Yes, It imports the My Clippings.txt file (The exact format we are trying to replicate here)

And our Export plugin provides the same timestamp as the Kindle does.
Maybe myclippings.io wants to connect to the device to get its timezone and to recalculate the time?

And our Export plugin provides the same timestamp as the Kindle does. Maybe myclippings.io wants to connect to the device to get its timezone and to recalculate the time?

I don't have a kindle right now with me, but can you check if the exported file has the timestamp converted to UTC?

Okay, I got your point.
I am looking at the My Clippings.txt made just now.
It contains the timestamp exactly with the current time of my Kindle. It is UTC+2.

Okay, I got your point. I am looking at the My Clippings.txt made just now. It contains the timestamp exactly with the current time of my Kindle. It is UTC+2.

Ah, Thank you. So you mean it's the local time and not UTC. Do you mind uploading that to the service ( you just need a simple https://clippings.io account ) to see if the time appears right?

May I see a preview of that file please?

I haven't used apps that integrate with my clippings, only obsidian plugins which worked fine; I just made the plugin since kindle has a huge market-share and thought it would make for a very compatibility with other projects.

I'll wait for the updates but my 2 cents are: I don't think kindle end-users would know what timestamps other than local timestamps would be.
I'm rather perplexed about the way this service handles time, as we are just exploring a timezone-less datetime string...
So any parsing should be pretty forward and not apply mutations; if kindle devices export in local timezones (which I'd expect, else I could imagine all the calls to Amazon's support services), could it be that the app applies the timezones erroneously?

Maybe you can shoot an email to the app's support services and ask about it :)

Maybe a new issue can be opened about this as it's out of scope with the out of memory fix and the authors fix and, maybe, a bug on the app's end instead of koreader's, so if we have to wait for updates this PR wouldn't be the best to store that information for the future.

I haven't used apps that integrate with my clippings, only obsidian plugins which worked fine; I just made the plugin since kindle has a huge market-share and thought it would make for a very compatibility with other projects.

I'll wait for the updates but my 2 cents are: I don't think kindle end-users would know what timestamps other than local timestamps would be. I'm rather perplexed about the way this service handles time, as we are just exploring a timezone-less datetime string... So any parsing should be pretty forward and not apply mutations; if kindle devices export in local timezones (which I'd expect, else I could imagine all the calls to Amazon's support services), could it be that the app applies the timezones erroneously?

Maybe you can shoot an email to the app's support services and ask about it :)

Then I assume that is the issue with the web service (clippings.io) itself. Don't bother about it. For now, I will be happy with merging #12106 and #12107 (both looks approved, but have conflicting changes). I am not sure how @pazos or @Frenzie will resolve the conflict, but in case of any issue, I am down to help.