From:  "ISHIKAWA,chiaki" <ishikawa@yk.rim.or.jp>
Date:  16 Mar 2019 01:00:59 Hong Kong Time
Newsgroup:  news.mozilla.org/mozilla.dev.apps.thunderbird
Subject:  

Re: Current state of |make mozmill| from a developer's point of view.

NNTP-Posting-Host:  63.245.210.105

Jörg,

Thank you for your feedback.


On 2019/03/15 20:21, Jörg Knobloch wrote:
> On 14/03/2019 17:58, ISHIKAWA,chiaki wrote:
>> I inserted my observation where appropriate. 
>
> Chiaki,
>
> thanks for looking and reporting. It's surprising TB works at all with 
> all those warnings and even assertions :-(
>

I have a similar sentiment when I created the summary after these months.

>
> IMHO, the editor and spellcheck errors are pretty much a hopeless case 
> since M-C people would have to look at them. I raised a few of these 
> or similar warnings with Masayuki-san a while ago, they fixed some and 
> others got postponed.
>
I see. I will leave editor and spellcheck errors as low-priority work 
(for TB developers/patch submitters).

> What stands out for me are
>
> 293: Main Thread] WARNING: NS_ENSURE_SUCCESS(rv, rv) failed with 
> result 0x80550006: file 
> /NREF-COMM-CENTRAL/mozilla/comm/mailnews/local/src/nsMailboxService.cpp, 
> line 666
> 293: Main Thread] WARNING: NS_ENSURE_SUCCESS(rv, rv) failed with 
> result 0x80550006: file 
> /NREF-COMM-CENTRAL/mozilla/comm/mailnews/base/util/nsMsgDBFolder.cpp, 
> line 5132
>
> and of course all the assertions, particularly those in our code base:
>
> 10 [13726, Main Thread] ###!!! ASSERTION: not-null 
> m_mime_delivery_state: 'm_mime_delivery_state != nullptr', file 
> /NREF-COMM-CENTRAL/mozilla/comm/mailnews/compose/src/nsMsgAttachmentHandler.cpp, 
> line 1052
>
> 2 [13726, Main Thread] ###!!! ASSERTION: RemoveCurrentDraftMessage 
> msgSend is null.: 'msgSend', file 
> /NREF-COMM-CENTRAL/mozilla/comm/mailnews/compose/src/nsMsgCompose.cpp, 
> line 3989
>
The above two assertions *might* be related to the shutdown cleanup issues.
>
> 1 [15766, Main Thread] ###!!! ASSERTION: fix this: '!(*pResultKey != 
> nsMsgKey_None && resultThreadId && *resultThreadId == nsMsgKey_None)', 
> file 
> /NREF-COMM-CENTRAL/mozilla/comm/mailnews/base/src/nsMsgDBView.cpp, 
> line 7844

I think I filed a bugzilla for this a few years ago. I did not realize 
this assertion is triggered now again.

>
> All these need a bug and careful looking into.
>
While I clean up local patches to work with the latest source, I will 
try to see if I can look into the above issues.
[The local patches to deal with low-level I/O errors worked beautifully 
up until about April/May 2018, but somehow suddenly started to fail in 
June 2018, and I still can't figure out exactly why. I am planning to 
split the local patches into smaller chunks to figure out what is to 
blame. And maybe smaller patches could be merged one by one more easily.
What delayed the clean up effort was  my tool for comparing/merging the 
patches under linux, a GUI-tool called MELD started to have an issue 
with Japanese locale or something that is very opaque to me. It must 
have something to do with obscure python issue. I reported the bug to 
MELD developers and there is a better error reporting facility, but I 
think I need to change my locale when I merge the patches. This is a bit 
awkward.]

> Oh, and the some/most of the JS errors are from recent de-XBL efforts 
> where the new custom elements now spew out errors :-( - like: 
> chrome://messenger/content/multimessageview.js. That should also be 
> filed as a bug with CC/NI :mkmelin and Arshad.
>
I will look into this also. I will try to file a bugzilla after waiting 
for a few more source updates. They may be fixed in a newer revision.

 > Jörg.

Thank you again!

Chiaki