+++ This bug was initially created as a clone of Bug #1727008 +++

After bug 1727008, comm/mail/test/browser/cloudfile/browser_attachmentUrls.js is now failing. Here's the log.

I tried to push to trysever, but I got this error:

$ hg push -f try
pushing to ssh://hg.mozilla.org/try
no revisions specified to push; using . to avoid pushing multiple heads
searching for changes
warning: repository is unrelated
remote: waiting for lock on working directory of /repo/hg/mozilla/try held by process '30854' on host 'hgssh1.dmz.mdc1.mozilla.com/effffffc'
remote: got lock after 23 seconds
remote: adding changesets
remote: adding manifests
remote: adding file changes
remote:
remote: ****************************** ERROR *******************************
remote: a36e027fc622 adds or modifies the following symlinks:
remote:
remote:   mail/app/macbuild/Contents/CodeResources
remote:
remote: Symlinks aren't allowed in this repo. Convert these paths to regular
remote: files and try your push again.
remote: ********************************************************************
remote:
remote: transaction abort!
remote: rollback completed
remote: pretxnchangegroup.mozhooks hook failed
abort: push failed on remote

Hmm. (My environment is Windows 10)

You need to put one patch on try and one on try-comm-central. This page should sort you out:
https://developer.thunderbird.net/thunderbird-development/fixing-a-bug/try-server#testing-mozilla-central-patches

(In reply to Geoff Lankow (:darktrojan) from comment #2)

You need to put one patch on try and one on try-comm-central. This page should sort you out:
https://developer.thunderbird.net/thunderbird-development/fixing-a-bug/try-server#testing-mozilla-central-patches

Yeah, I did it, but I see the error. I'm still not sure what's wrong in my environment...

From that error message it looks like you're trying to push comm-central things to ssh://hg.mozilla.org/try instead of ssh://hg.mozilla.org/try-comm-central.

Set release status flags based on info from the regressing bug 1727008

Thanks. Oddly, the hgrc settings are ignored when I use hg push -f try. I now specify the URI directly, then, it works for me.

Perhaps, fix of bug 1727185 will hide the failure of the test. But I need to fix the root cause. Keep investigating which revision (part 13 - part 21) is wrong for the test or bug 1727008 is just not enough fixed the regression of part 13.

Hmm, it's odd even with the patches for bug 1727185, I couldn't get useful warning messages nor hitting MOZ_DIAGNOSTIC_ASSERT stack...

[task 2021-08-25T02:05:19.102Z] 02:05:19     INFO - TEST-START | comm/mail/test/browser/cloudfile/browser_attachmentUrls.js
[task 2021-08-25T02:05:19.239Z] 02:05:19     INFO - GECKO(9455) | ### XPCOM_MEM_BLOAT_LOG defined -- logging bloat/leaks to /tmp/tmp_2rwifwi.mozrunner/runtests_leaks_tab_pid9667.log
[task 2021-08-25T02:05:19.239Z] 02:05:19     INFO - GECKO(9455) | [9667, Main Thread] WARNING: XPCOM_MEM_BLOAT_LOG is set, disabling native allocations.: file /builds/worker/checkouts/gecko/tools/profiler/core/platform.cpp:248
[task 2021-08-25T02:05:19.260Z] 02:05:19     INFO - GECKO(9455) | [Child 9667, Main Thread] WARNING: could not set real-time limit in CubebUtils::InitLibrary: file /builds/worker/checkouts/gecko/dom/media/CubebUtils.cpp:642
[task 2021-08-25T02:05:19.328Z] 02:05:19     INFO - GECKO(9455) | Warning: asking to enable_gpu_markers but no supporting extension was found
[task 2021-08-25T02:05:19.529Z] 02:05:19     INFO - GECKO(9455) | [Parent 9455, Main Thread] WARNING: NS_ENSURE_TRUE(presShell) failed: file /builds/worker/checkouts/gecko/dom/base/nsGlobalWindowOuter.cpp:4231
[task 2021-08-25T02:05:19.731Z] 02:05:19     INFO - GECKO(9455) | [Parent 9455, Main Thread] WARNING: NS_ENSURE_TRUE(rootFrame) failed: file /builds/worker/checkouts/gecko/dom/base/nsGlobalWindowOuter.cpp:4233
[task 2021-08-25T02:05:19.873Z] 02:05:19     INFO - GECKO(9455) | [Parent 9455, Main Thread] WARNING: NS_ENSURE_TRUE(uri) failed: file /builds/worker/checkouts/gecko/netwerk/base/nsIOService.cpp:1679
[task 2021-08-25T02:05:19.874Z] 02:05:19     INFO - GECKO(9455) | [Parent 9455, Main Thread] WARNING: NS_ENSURE_SUCCESS(rv, rv) failed with result 0x80070057 (NS_ERROR_ILLEGAL_VALUE): file /builds/worker/checkouts/gecko/netwerk/base/nsIOService.cpp:1701
[task 2021-08-25T02:05:19.875Z] 02:05:19     INFO - GECKO(9455) | [Parent 9455, Main Thread] WARNING: NS_ENSURE_SUCCESS(rv, nullptr) failed with result 0x804B000A (NS_ERROR_MALFORMED_URI): file /builds/worker/checkouts/gecko/extensions/permissions/Permission.cpp:45
[task 2021-08-25T02:05:19.875Z] 02:05:19     INFO - GECKO(9455) | [Parent 9455, Main Thread] WARNING: NS_ENSURE_TRUE(principal) failed: file /builds/worker/checkouts/gecko/extensions/permissions/Permission.cpp:57
[task 2021-08-25T02:05:20.074Z] 02:05:20     INFO - GECKO(9455) | Warning: asking to enable_gpu_markers but no supporting extension was found
[task 2021-08-25T02:05:20.079Z] 02:05:20     INFO - GECKO(9455) | Warning: asking to enable_gpu_markers but no supporting extension was found
[task 2021-08-25T02:05:20.082Z] 02:05:20     INFO - GECKO(9455) | Warning: asking to enable_gpu_markers but no supporting extension was found
[task 2021-08-25T02:05:20.087Z] 02:05:20     INFO - GECKO(9455) | Warning: asking to enable_gpu_markers but no supporting extension was found
[task 2021-08-25T02:05:20.112Z] 02:05:20     INFO - GECKO(9455) | Warning: asking to enable_gpu_markers but no supporting extension was found
[task 2021-08-25T02:05:20.117Z] 02:05:20     INFO - GECKO(9455) | Warning: asking to enable_gpu_markers but no supporting extension was found
[task 2021-08-25T02:05:20.122Z] 02:05:20     INFO - GECKO(9455) | Warning: asking to enable_gpu_markers but no supporting extension was found
[task 2021-08-25T02:05:20.169Z] 02:05:20     INFO - GECKO(9455) | [Parent 9455, Main Thread] WARNING: NS_ENSURE_TRUE(presShell) failed: file /builds/worker/checkouts/gecko/dom/base/nsGlobalWindowOuter.cpp:4231
[task 2021-08-25T02:05:20.169Z] 02:05:20     INFO - GECKO(9455) | [Parent 9455, Main Thread] WARNING: NS_ENSURE_TRUE(rootFrame) failed: file /builds/worker/checkouts/gecko/dom/base/nsGlobalWindowOuter.cpp:4233
[task 2021-08-25T02:05:20.170Z] 02:05:20     INFO - GECKO(9455) | [Parent 9455, Main Thread] WARNING: NS_ENSURE_TRUE(presShell) failed: file /builds/worker/checkouts/gecko/dom/base/nsGlobalWindowOuter.cpp:4231
[task 2021-08-25T02:05:20.206Z] 02:05:20     INFO - GECKO(9455) | [Parent 9455, Main Thread] WARNING: Failed to retarget HTML data delivery to the parser thread.: file /builds/worker/checkouts/gecko/parser/html/nsHtml5StreamParser.cpp:1365
[task 2021-08-25T02:05:20.403Z] 02:05:20     INFO - GECKO(9455) | [Parent 9455, Main Thread] WARNING: Failed to retarget HTML data delivery to the parser thread.: file /builds/worker/checkouts/gecko/parser/html/nsHtml5StreamParser.cpp:1365
[task 2021-08-25T02:05:20.417Z] 02:05:20     INFO - GECKO(9455) | [Parent 9455, Main Thread] WARNING: '!aChild->IsContent()', file /builds/worker/workspace/obj-build/dist/include/mozilla/EditorDOMPoint.h:465
[task 2021-08-25T02:05:20.427Z] 02:05:20     INFO - GECKO(9455) | [Parent 9455, Main Thread] WARNING: HTMLEditor::GetAbsolutelyPositionedSelectionContainer() reached <html> element: file /builds/worker/checkouts/gecko/editor/libeditor/HTMLAbsPositionEditor.cpp:95

Perhaps, I've being investigated wrong area.

The exact regression cause is https://phabricator.services.mozilla.com/D123069

Unfortunately, this is a victim of the intentional change of Gecko.

Previously, we allowed editor to insert content after invisible <br> element (last content of a block or immediately before a block). However, inserting content after an invisible <br> element means that it becomes visible, and the other browsers don't do this odd behavior. Therefore, I made HTMLEditor stop doing it. So, from the point of view of a Gecko developer, this new behavior shouldn't be reverted.

According to the comm-central code, Tb tries to insert a <div> after an invisible <br> element here:

• https://searchfox.org/comm-central/rev/ddd293b13d7052f64e51dca70678bfdf88fe2088/mail/components/compose/content/cloudAttachmentLinkManager.js#340
• https://searchfox.org/comm-central/rev/ddd293b13d7052f64e51dca70678bfdf88fe2088/mail/components/compose/content/cloudAttachmentLinkManager.js#354

But, it's not allowed right now. If they don't need to be undoaable, I recommend to use raw DOM API such as Node.appendChild. Otherwise, please put another <br> element after the invisible one. Then, insertElementAtSelection will work.

Could you update the composer behavior in the Thunderbird side?

I think I have this under control. Thanks for figuring out what the heck is going on. I'd assumed it was from bug 1727008 because the failure turned up on the Try run I did for that, I guess the change for bug 1726064 must've also been present there.

Adds a <br> ahead of the FileLink <div> if the insertion is at the start of the document or following a <span>, which are the most likely cases where it is needed.

Pushed by mkmelin@iki.fi:
https://hg.mozilla.org/comm-central/rev/5024c2c13b47
When inserting FileLink URLs, restore missing line breaks. r=mkmelin