Closed Bug 17811 Opened 22 years ago Closed 22 years ago
[DOGFOOD][PP]Subject is blank in the thread pane
build: 1999110212 OS: RH6.0 On today's latest build, when opening the smoketest folder, not all the mails in the folder are loaded. There are 5 mails in the folder, only one got loaded. Inbox has some mails missing too.
Reassign to putterman
what is this attachment?
does this work for you on windows? What makes this an intl smoketest folder versus a non intl smoketest folder?
attached is a zip file which I used for intl smoketest. It contains 5 messages containing 8 bit and Japanese characters. It works fine on Windows.
Drop the Smoketest file into your POP mail directory and try displaying 5 msgs. On Linux, you will only see one msg displayed in the thread pane. This only fails on Linux and starting today. This is International Smoketest file for Display and the display passed on all 5 msgs yesterday.
Do you get something that says: Break: at file nsXulDocument.cpp, line 1761 Assertion: "error creating content" I can't even load my Sent folder on Linux.
No. I don't see those errors when trying to load the smoketest folder.
earlier today Kipp fixed a linux font problem. Perhaps this is what is happening here?
Linux Redhat 6.0 (1999-11-02-12 M11) 233 MHZ 128 MB Ram Using today's Linux build, I am unable to load message into all folders, including special or non-special folders both in POP or IMAP accounts. I click on Get Msg, it pops up the password dialog. I enter the password, but still nothing happens.
I can't download or unzip/untar the attachment here. Is there an alternate location where I can get it internally? Thanks.
I have smoketest.zip file in my home directory: /u/ji
Summary: [DOGFOOD]Does't load all the mails in the smoketest folder → [DOGFOOD][PP]Unable to display list of msgs in thread pane
dding waterson to cc list. Although ji isn't seeing this, I can't get this far because of assertions I'm seeing. About 9 of them in a row. akkana saw the same thing in a message she sent to mailnewsstaff earlier. Chris, does this mean anything to you: Break: at file nsXulDocument.cpp, line 1761 Assertion: "error creating content"
Putting on PDT+ radar.
cc'ing alec. Alec, given that the builder is apparently failing, I'm wondering if there's something that might not be parsing right on Linux and thereby returning a failure code somewhere.(it's the only think I can think that we are doing that is different than Windows).
see alec's comments in bug http://bugzilla.mozilla.org/show_bug.cgi?id=17214
oh my god, I've been debugging this all afternoon, I thought it was my tree! here's the deal: the builder is failing because we can't get the Mime2Encoded subject out of the database.
Summary: [DOGFOOD][PP]Unable to display list of msgs in thread pane → [DOGFOOD][PP]Subject is blank in the thread pane.
Ok, I checked in a fix so that the messages would display, but the subject will be blank. This means the bug is simpler, so I'm updating the summary.
I backed out ftang's checkin to nsCharsetConverterManager.cpp in my tree and it started working just fine. It looks like there's some problem with folder charsets. I'm calling ftang now. If I can't contact him, I'm backing him out.
I am working on a fix....
I think the right thing to do is to back it out. I will do that when the tree is green.
I back out my change. If that fix the problem, please mark it as FIXED.
with Alec's change, I can't load my folder at all - I don't think his change is good. RDF asserts when we return success with a null result. I'm getting asked for the thread property, which nsMsgMessageDataSource doesn't handle. Help!
I have a fix for the assertions, but since the tree is currently red, I don't plan on staying up long enough to check it in and watch the tree so it'll have to get checked in tomorrow unless someone else is going to be around. Basically, at the end of nsMsgMessageDataSource::createMessageNode we need to return rv instead of NS_OK. Otherwise we are assuming that none of the other functions return a success code besides NS_OK, which isn't true because some of them return NS_RDF_NO_VALUE. This gets rid of the asserts in my build.
And as David B pointed out to me, we either need to return rv or have another case where we return NS_RDF_NO_VALUE in case we don't handle the property which is what it sounds like was causing his assertions.
Status: NEW → RESOLVED
Closed: 22 years ago
Resolution: --- → FIXED
scott checked in his extra fix to my fix, and ftang has backed out his stuff. Thanks everyone
Lisa, you can assing this to ji if you prefer her to check with her original examples.
ok. Ji, can you verify that this is fixed for your original scenario? Thanks.
Verified with linux 1999110508 build. It's fixed.
Verified with Linux 1999111809 build. It is fixed.
You need to log in before you can comment on or make changes to this bug.