Steps to Reproduce: 1. Use the 'Standard' layout from 'View - Layout' in the menu. Show the message view and adjust its size until there are six messages visible in the message list. Hide the message view. 2. Go to a folder that has enough messages to fill at least half or more of the message list's visible area. 3. Select a message near the bottom of the message list's visible area. 4. Show the message view. Actual Result: The selected message is not visible in the message list. Expected Result: The selected message would be visible in the message list as well as the message below it. Note that the expected result is how 3.18.0 (gtk2) currently works, in case that is a helpful reference. The reason this feature is helpful is because being able to see the selected message and the one directly above and/or beneath it informs your decision about whether you will proceed directly with some action on the message or whether you will first hide the message view and then proceed.
This is very similar to bug #4521, however while I can reproduce #4521, I cannot reproduce this one. It seems that if #4521 it would fix this also, so am closing this one. *** This bug has been marked as a duplicate of bug 4521 ***
I guess you could be right; however, from a user perspective, they seem different. It's interesting that you can't reproduce it; I wonder if it has something to do with how my system (Linux Mint 20 XFCE) interacts with GTK3. I specifically mention GTK3 because 3.18.0 (GTK2) works as expected on the same system. I would also note that this bug occurs for me even in instances where Bug 4520 does not occur; this also makes me think they are separate issues. However, I don't know how the underlying code is connected, so perhaps they would both be solved by fixing the one. Just to confirm, if you go through the STR for both 3.18.0 and 4.0.0, do you get exactly the same result in regards to whether or not the specific messages visible in the message list changes or not when message view is shown?
(In reply to Paul from comment #1) Maybe I can shed more light on this now. I was working with my email and realized that I wasn't able to reproduce this bug either anymore. What I soon figured out was that the bug doesn't manifest in folders of an IMAP account, but it does manifest in folders of a POP account. When you get a chance, could you see if you can reproduce this in the context of a POP account?
I was using an MH mailbox in my tests, but there is no reason that either a POP account or an IMAP account's mailbox should behave any differently.
Okay, another thing I noticed that might be the key. Based on some limited testing, my hypothesis is that this bug only shows up when an HTML viewer plugin is loaded, the message that is selected has an HTML part, and when one's config is set so that if an HTML part is present then that is the default to be shown when the message view is shown. This seems to be consistent for both MH and IMAP mailboxes. Would you mind running the STR again with the following additional details added? * Have one of the HTML Viewer plugins loaded. * In 'Config - Prefs - Message View - Text Options - HTML Messages, make sure the "Select HTML Part..." option is enabled. (If it's relevant, I also have the other two options in that section enabled.) * For the selected message, use one that has an HTML part. On my system the bug shows up with LiteHTML, Fancy, and Dillo plugins. The only difference is that, with the Dillo plugin, after the STR is complete, if you hide the message view, Claws will crash and terminate. Will be interested in finding out your results.
I can confirm that fixing Bug 4521 did not fix this bug. However, there is some inconsistency in the reproduction of this bug. For instance, if I run about ten cycles of showing / hiding the message view for the same message (without doing anything else in-between), the bug will only show up on some of those cycles. I will continue to look into this as I have time, to see if I can find out any other useful information. If there are specific things I can do to help with troubleshooting / testing in relation to getting this bug fixed, let me know.