eric99

Is it possible to ignore anyone below a specific number of posts too? 😏

Probably because these users don't try hard enough with different notebooks, different search strings and with different result list (length). It doesn't occur in all circumstances... But then again, what's the point of the very slow rollout if no action is taken when these issues were reported in a very early stage? I would expect them to at least analyze/debug the issues to get a better understanding of the impact before moving forward with the rollout...

Sure, in the 1980s it was common practice to wait a few seconds for data retrieval from a database. If you had told developers back then that this would still be the case in 2023 on computers that are 10,000 times faster, they would have laughed at you in disbelief!

is this error message: Minified React error #185

I noticed that all these problems do also occur in the latest WEB client, they aren't windows client related: searching in a notebook is denied with the pop-up message: "Permissions lost, you can no longer access notebook XYZ" sometimes global search in all notes returns the full note list instead of the filtered one. deleting a note by pressing the delete button doesn't always respond the note menu is sometimes completely grayed-out for some notes when scrolling to the bottom of a result list with length > 128 ( 2^7 ), it always jumps back automatically to the top of the list. So the notes at the end of the list are unreachable sometimes a global search does return a result list, but then after just waiting 10-30 seconds, it spontaneously starts repeating the search with intervals of a few seconds. This happens only when the result list sort-order is set to 'relevant' In order to reproduce this, I must emphasize that you should try with several notebooks and/or several different search strings to reproduce the issues. For instance, the result list oscillation when the sort-order is set to 'relevant' doesn't always occur, it seems to be dependent on the length of the list (the longer the better) or the returned notes or the search string (my strings are always single word but not all of them can trigger it) Once it starts oscillating, it will automatically refresh the result list with a period of a few seconds, sometimes a burst of a few refreshes, sometimes continuously for a minute or so. Sometimes you can start it by just clicking or hoovering in the list a few times. I wonder whether this is a failing BendingSpoons AI prototype to find the most relevant search results? Is there anyone who can reproduce this oscillation on any platform? I need this info for another ticket again...

Divide and conquer is sometimes better than trying to force very different things into one 'unified' unmaintainable program. Do you know how much accidental complexity has been added by the electron framework and the code above it to accomplish this? It's already taking more than 2 years now to get V10 stable...

Or the opposite: these 'experienced' developers created a monolithic software, full of dependencies all over the place, which is unmaintainable by other developers: if you touch one thing it completely collapses. Maybe the right time to rewrite it again in a few simple native implementations ?

Yes, I also wonder what's the point of a phased rollout if they don't take action when problems are reported. Even if an issue happens only once, you can't ignore it because it will explode in your face after the rollout to 200 million users is complete.

I hope with all that regression that they don't forget to regress the prices too 😏

Yes, because of all the other issues, I didn't even mention the performance regression in this release: note creation: 3 to 10 seconds, note deletion: several seconds or sometimes no delete at all, moving to other notebook: seconds

I'm not so sure that the first line is completely new, unless they are all called Diezel...

This is not my experience: sometimes support seems to be completely ignorant of an issue that has already been submitted multiple times by other affected users. I can't find the issue status at https://help.evernote.com/hc/en-us/requests . The only status I get is that it's solved which isn't the case. For some tickets I'm still waiting for 2 years to be solved.

The major problem with tickets is that related issues/info from different users get scattered over several 'independent' tickets. And the worst problem is that it is extremely difficult to get the ticket to the developers, it is always shielded by various impenetrable levels of support and the status of the issue is kept secret for the end-user once it has been fed into the pipeline. I believe much more in the agile interactive way of bug reporting/solving on the forum, like Federico already did successfully for some urgent problems.

Here's a list of the regression issues I'm experiencing in 10.59.5. Have you already tested this yourself? https://discussion.evernote.com/forums/topic/146555-👉-input-needed-– what-issues-did-you-experience-in-the-last-couple-of-days/?do=findComment&comment=692061

To prevent this kind of chatGPT reactions from support, I always added a forum link to my tickets to provide more context. Apart from that, this "notebook access denied" problem has been introduced in the latest release, they just have to check the changes they made in the latest release to find the problem...

Yeah, since we are on @Federico Simionato 's thread, I hope to get some response too on the v10.59.5 regression list: https://discussion.evernote.com/forums/topic/146555-👉-input-needed-– what-issues-did-you-experience-in-the-last-couple-of-days/?do=findComment&comment=692061

"...Those inline images are CDATA encoded base64 strings. They don’t include a hash, but I figured out that decoding the base64 string and then running it through MD5 generates the hash that is used in the corresponding <en-media> tag..." https://simonwillison.net/2020/Oct/16/building-evernote-sqlite-exporter/ So, as I read this, it's a MD5 calculation on the raw image data.

I found out that the global search oscillation is caused by the sort order of the result list which was default set to 'relevant' . Apparently it's trying endlessly to find the most relevant order and gets never satisfied, probably because I write only highly relevant notes 😉. When I set the sort order to 'update time', the global search seems to work normally. The other problems I mentioned before are still there. edit: even with the 'update time' sort order set, there is another problem: when scrolling to the bottom of a result list with length > 128 ( 2^7 ), it always jumps back automatically to the top of the list. So the notes at the end of the list are unreachable.

That's the advantage of a unified code base 😉

There is an obsidian search plugin that handles OCR and PDF

EN data is just encrypted at rest, it can still be accessed by Evernote employees. Obsidian or joplin employees can not access your data due to full E2EE with your own private password.

Yes, very frustrating. I'll further check Obsidian and Joplin in more detail. Both have full E2EE AES 256 encryption, a big plus.

Yes, I have several search problems too:

Yes, race conditions bugs are hard to pinpoint because these are very difficult to reproduce. Actually, race conditions should have been handled in the early design and not as an afterthought. That's the only way to get rid of it.

Yeah, but this might be the result of taking over a monolitic code base: you turn one ***** to fix a specific problem, and suddenly an unexpected side effect occurs in a completely different piece of the code base. Therefore, before changing anything, much more (unit) tests should be written to verify that the behavior hasn't been changed after refactoring.