As far as the issues mentioned by me are concerned, these are my findings in 1.0.190:
[x] The generated HTML is identical to v1.0.179’s.[ ] H1: the style for #rendered-md H1:first-child is still missing. As a consequence, there is 1.2em whitespace above the first header of the note.
[x] H5-2 and H6-2 are lightgreen, as in 1.0.179.[x] The word “red” is red, as in 1.0.179.[x] <mark>bla</mark> marks “bla”, as in 1.0.179.So one of my issues is still open.
I removed this CSS as the logic was making it remove the top margin from lists and probably other elements too.
If someone has a clue about CSS and know how to implement this, I'd accept a PR. The idea is to remove the top margin from H1 and other elements when they are on top of the page (because there's already a margin there).
I'm not 100% sure what you mean by "this", but I have a suspicion. The original CSS was:
#rendered-md h1:first-child,
#rendered-md h2:first-child,
#rendered-md h3:first-child,
#rendered-md h4:first-child,
#rendered-md ul:first-child,
#rendered-md ol:first-child,
#rendered-md table:first-child,
#rendered-md blockquote:first-child,
#rendered-md img:first-child,
#rendered-md p:first-child {
margin-top: 0;
padding-top: 0;
}
This covers all h1..p elements that are a first child of some element AND are also a descendant (in whatever degree) of #rendered-md. That's a bit too much. If it is replaced with:
#rendered-md > h1:first-child,
#rendered-md > h2:first-child,
#rendered-md > h3:first-child,
#rendered-md > h4:first-child,
#rendered-md > ul:first-child,
#rendered-md > ol:first-child,
#rendered-md > table:first-child,
#rendered-md > blockquote:first-child,
#rendered-md > img:first-child,
#rendered-md > p:first-child {
margin-top: 0;
padding-top: 0;
}
then it will cover all h1..p elements that are the first direct child of #rendered-md. That's exactly what we want, isn't it?
This is really weird! My experiments have been executed on v1.0.179, and the CSS works there as might be expected.
On v1.0.190 however, the greater-than sign in the selectors part make the entire rule invalid. Even one > operator is fatal. I have no explanation for this. The rules that follow (coloring the H5 and H6) still work, so the <style>...</style> block has not completely been blown up.
This concerns only the CSS specs in the note itself. When I copy them to userstyle.css, they work again as expected. So there must some extra intermediary interpretation of style blocks embedded in notes that is malfunctioning now. New issue!
Yes this is fixed, it's because the style was incorrectly escaped. The fix will be in the next release.
Yes thank you, that works, so I'll re-enable it in the next release.
New pre-release with more fixes:
https://github.com/laurent22/joplin/releases/tag/v1.0.192
I think we’re getting close to production ready release but if you still notice some issue, please post here.
Note that due to the encryption method change, non-updated clients won’t be able to decrypt the notes created by this new version (assuming you use E2EE). All clients are going to be updated soon but something to keep in mind.
If you try to sync these newly encrypted notes with a non-updated client, the notes will simply remain encrypted on that client, and will be decrypted when you upgrade.
This would have been a good point to increase the minor. So that all clients with the same minor are compatible.
With how many pre-release builds, I’d definitely expect Joplin 1.1.0. Unless you’re planning on going the Vim route and be up to 8.1.9253 or something crazy before incrementing the minor version. Ha
Hello everybody, I’m experiencing the graphical bug in the screen recording below since I updated Joplin app on MacOS 10.15.3 to v1.0.189. Right now I’m running v1.0.192, always on MacOS.
When I’m scrolling down in a note, even if it’s short and it completely fits in the page, I have a white blank margin coming up from below. As you can see, even when not scrolling, there’s also a persistent scrollbar on the right.
If it’s necessary I can create an issue on GitHub.
Let me know, thanks for your help @laurent!
@xplosionmind, there was no attached screen recording.
I'm sorry, I uploaded the video but probably Discourse doesn't support it. Now I fixed it inserting a link to an unlisted YouTube video.
I am using 179. Will there be non revertable changes to the database? i.e. can I just swap back to previous version if need be? If so I’ll give it a spin as it’s easy for me to swap versions the way I have it installed.
I can definitely answer this. I'm not sure what's been changed but all versions past at least the draft version 182 all had where i needed to delete and rebuild my config directory when moving back to 179
Ah, always good to be reminded to start with a blank ~/.config/joplin-desktop/ (after having it backed up some way) in case of a rollback. I have read various instructions for such cases on this forum, but I would very much appreciate a complete guide in one place for all platforms.
Currently, I do all my experiments on the Mac, because rolling back is so easy there: just rename/move the ~/.config/joplin-desktop/ folder, copy the old Joplin version to /Applications/ (overwriting the newer version), run it, define the wanted configuration settings again (actually the most annoying operation is this procedure), copy the old userstyle.css and userchrome.css to the new ~/.config/joplin-desktop/ folder, exit Joplin, rerun it, and find that everything works again as it did before. And oh yes, the backup of the configuration can now be deleted.
I wouldn’t know how to downgrade on Windows or Linux. And mobile is completely dark matter to me.
That's a good point, and I believe you cannot downgrade because the new version will modify the database structure.
There have been talks of having distinct channels for beta versions, and perhaps that's we should do eventually. It's just particularly tricky with Joplin since there are three different clients.
It seems most issues are fixed now? I’m thinking about releasing desktop and the mobile apps tomorrow, but if you notice anything let me know.
Hi,
this version works now. I’ve retrieved all the content and my booknotes.
Thanks to @Laurent and all testers.
The final version, v1.0.193, is now available. Thanks everyone to help test this release. It was a big one, with several in-depth changes, so that beta testing was really helpful.
Android and CLI versions have been released too, and iOS will come soon.
If you notice any other bug, please open an issue directly on GitHub as it’s easier to keep track of them there.