Bugzilla – Full Text Bug Listing
| Summary: | Wrong subject / message links in MHonArc mailing list archive | ||
|---|---|---|---|
| Product: | Community | Reporter: | David Williams <david_williams> |
| Component: | MailingLists | Assignee: | Eclipse Webmaster <webmaster> |
| Status: | RESOLVED FIXED | QA Contact: | |
| Severity: | normal | ||
| Priority: | P3 | CC: | daniel_megert, david_williams, markus.kell.r |
| Version: | unspecified | ||
| Target Milestone: | --- | ||
| Hardware: | All | ||
| OS: | All | ||
| Whiteboard: | |||
| Bug Depends on: | 401312 | ||
| Bug Blocks: | 484429 | ||
|
Description
David Williams
The threading and subjects are fine for me in IBM Notes 9.0.1 and in Thunderbird. Nabble also shows the correct subject, but it is unable to group the messages into threads (probably because they didn't grok that the algorithm should also work if you cannot guarantee that a Message-ID is available for everything referenced in a "References:" header): http://eclipse.1072660.n5.nabble.com/4-6-0-N-Build-N20160104-2000-1-failure-from-ep46N-unit-win32-td178610.html FWIW, here's a document that describes how threading is supposed to work, and that includes links to specs that show that the headers we generate are fine: https://www.jwz.org/doc/threading.html Note that we use the same approach as Bugzilla to create the threading headers. => The only thing I see that's wrong is the (white-on-blue) subject in the MHonArc archive: https://dev.eclipse.org/mhonarc/lists/platform-releng-dev/msg23820.html Moving to webmaster. The bug is either in MHonArc or in the way mails are fed into MHonArc. I don't see a connection to the "References:" header. Wheew, this is even worse! *All* recent messages from 2016 on https://dev.eclipse.org/mhonarc/lists/platform-releng-dev/maillist.html now link to that special https://dev.eclipse.org/mhonarc/lists/platform-releng-dev/msg23820.html Just one message was archived correctly: https://dev.eclipse.org/mhonarc/lists/platform-releng-dev/msg23821.html . David's reply is again back to msg23820 All other new messages are not reachable from the archive page. That's not a huge deal in this specific case, but I worry it could be a bigger infrastructure issue. For the records, here are headers from two of the mails as I got them: --- https://dev.eclipse.org/mhonarc/lists/platform-releng-dev/msg23820.html : Received: by build.eclipse.org (Postfix, from userid 7336) id 7D49A270025; Tue, 5 Jan 2016 03:10:26 -0500 (EST) To: platform-releng-dev@eclipse.org From: e4Builder@eclipse.org In-Reply-To: <N20160104-2000@build.eclipse.org/build/eclipse/> Reference: <N20160104-2000@build.eclipse.org/build/eclipse/> MIME-Version: 1.0 Message-Id: <20160105081026.7D49A270025@build.eclipse.org> Date: Tue, 5 Jan 2016 03:10:26 -0500 (EST) Subject: [platform-releng-dev] 4.6.0 N-Build: N20160104-2000: 1 failure from ep46N-unit-win32 --- The mail whose subject is used in msg23820.html : Received: by build.eclipse.org (Postfix, from userid 7336) id 6C08A270025; Fri, 1 Jan 2016 03:00:27 -0500 (EST) To: platform-releng-dev@eclipse.org From: e4Builder@eclipse.org In-Reply-To: <N20151231-2000@build.eclipse.org/build/eclipse/> Reference: <N20151231-2000@build.eclipse.org/build/eclipse/> MIME-Version: 1.0 Message-Id: <20160101080028.6C08A270025@build.eclipse.org> Date: Fri, 1 Jan 2016 03:00:27 -0500 (EST) Subject: [platform-releng-dev] 4.6.0 N-Build: N20151231-2000: 0 failures from ep46N-unit-win32 I've rebuilt the archive, and the recent messages seem to correctly map to individual messages. -M. After clearing my browser cache, the archive looks good again. The wrong subjects have been corrected. Marking as FIXED. Webmaster, please reopen if you plan further investigations. Problem is back: https://dev.eclipse.org/mhonarc/lists/platform-releng-dev/2016/Feb/index.php lists many "[platform-releng-dev] 4.6.0 I-Build: I20160301-0800- Posted 10:35 AM(ET) e4Builder" that all link to https://dev.eclipse.org/mhonarc/lists/platform-releng-dev/msg24153.html and misses the actual messages. I've regenerated the archives again. I'm starting to think there may be something in the filesystem/nfs that is causing this. -M. |