Closed Bug 1243498 Opened 9 years ago Closed 8 years ago

Kill Eudora import for good


(MailNews Core :: Import, defect)

Not set


(Not tracked)

Thunderbird 48.0


(Reporter: jorgk-bmo, Assigned: jorgk-bmo)




(2 files)

As per
I think most of us agree that the Eudora days are over († 2006) and Penelope is dead also († 2010, was based on TB 3!!).

We should remove the Eudora import code from the code base.
I hope I haven't overlooked anything. This compiles and runs, so there shouldn't be much more to it.
Attachment #8712835 - Flags: feedback?(rkent)
Attachment #8712835 - Flags: feedback?(mkmelin+mozilla)
Attachment #8712835 - Flags: feedback?(Pidgeot18)
Comment on attachment 8712835 [details] [diff] [review]
Remove all traces of Eudora (v1) [checked in with some more hunks, comment #19]

Yeah, doesn't sound like it's worth keeping
Attachment #8712835 - Flags: feedback?(mkmelin+mozilla) → feedback+
Attachment #8712835 - Flags: review?(mkmelin+mozilla)
Comment on attachment 8712835 [details] [diff] [review]
Remove all traces of Eudora (v1) [checked in with some more hunks, comment #19]

Review of attachment 8712835 [details] [diff] [review]:

I've not really looked at the patch in substantive detail, so this is more of a "do I agree with the philosophy of the bug."

The last evidence of anyone trying to import Eudora was about 2 years ago--there's a bug filed 2014-01-11 (bug 958878), and the last comment on a bug by a putative importer was bug 958569. The last substantive fix was bug 228550, last touched 2011-08-06. That's pretty strong evidence that there is little interest in this code one way or another. Judging from the putative import test code of bug 1061086, the mail directory is largely in mbox format (which is a long-standing import feature request anyways), so a somewhat more complicated way of handling Eudora import could probably be hooked together for anyone who might be interested.

In other words, I have no objection to deleting the Eudora importer. Kill untested hacky code! ☺
Attachment #8712835 - Flags: feedback?(Pidgeot18) → feedback+
I'm not here precisely to argue against this bug, because there is clearly no appetite to fix this importer.  But I think we should be guided by facts - and Eudora is much more popular than you might think. One might even say wildly popular, considering how long since the product went unsupported.  "QUALCOMM discontinued sales for Eudora on May 1st, 2007.  As of October 31st, 2007 they no longer offer support for Eudora."

I've followed Eudora activity over the years and for the record - users still love it and are amazingly loyal.  But they do drop off as the years go on and look for something else. For example I have a person logged in my addressbook who switched from OSE 5/1/2013.  And last I looked there are still crash reports submitted to mozilla from Eudora OSE, previously Penelope - the Thunderbird-based port of Eudora.

In SUMO (I didn't check mozillazine which has a fair precentage of Thunderird support) we have 10 posts aout Eudora this year. And last year between 100 and 150 (hard to judge because of their stupid search). is an example, where Matt has posted.

Postbox has also killed the importer. is what they have to say. And we also need to have a support topic about importing. 

THere is also this - I found a support link ... it turns out all of is redirecting to
Wayne, we need to kill what is no longer working and not part of TB's future. Joshua argues that we should kill "untested hacky code". I would argue that we should kill code that wouldn't work even if we fixed the crash. Even if it didn't crash we would not maintain it any more.

Importing e-mail is a once-in-a-lifetime thing, so you can go through some hassle to achieve this. I dragged a very good friend of mine away from Eudora, so I know what I'm talking about. My recommendation stands: For Eudora import use Thunderbird 1.5 (one dot five) and pre-process the data with "Eudora Rescue", see here for details:
As I said, I'm not arguing against the decision to kill it. I agree. But all the cards should be on the table. And our support system should provide a path for users who want to move from Eudora to THunderbird.
See Also: → 1175055
See also: Bug 1175055 comment #37.
Comment on attachment 8712835 [details] [diff] [review]
Remove all traces of Eudora (v1) [checked in with some more hunks, comment #19]

Review of attachment 8712835 [details] [diff] [review]:

Killing Eudora import is fine with me.
Attachment #8712835 - Flags: feedback?(rkent) → feedback+
There seems to be more eudora references e.g. at:

Please look for more at:^[^\0]*%24&hitlimit=&tree=comm-central

Also, the patch does not remove anything from Seamonkey (/suite/*), like the strings and help texts. I CC the SM guys to say what they need here.

I can't really review this as it only compiles on Windows. I only checked the patch does not break a Linux build. Maybe a successful run on try on all platforms would be enough.
Assignee: nobody → mozilla
Blocks: 1061086
Keywords: user-doc-needed
I added the user-doc-needed so that after this is landed I ask for adding a mention of the removal to and that there are clear steps how to import Eudora mail into a recent Thunderbird (e.g. Install TB 11, import from Eudora, then run Eudora Rescue, THEN you can upgrade to current TB).
Aceman, thanks for looking at it. My problem is that the patch hasn't been reviewed in six weeks and if I touched "high traffic" files, like nsMsgSend.cpp, it would rot all the time.

There references are all to comments:

We are early in the TB 48 cycle, so how about landing this and then fixing up any left over traces later?
I'm pretty sure I hit all the relevant code.

You're right, there is stuff in the suite, so I can prepare another patch to remove that, too.
Magnus, do you think what I said in comment #11 is a possible way forward here?
Looking for "eudora"^[^\0]*%24&hitlimit=&tree=comm-central
is really confusing since it finds stuff in /mozilla like psEUDORAndom ;-)
IMHO it would be good to remove the bulk of the C++ and build config and then do a cleanup patch for comments later and also a patch for SM.
Flags: needinfo?(mkmelin+mozilla)
I'm fine with that plan, if the cleanup follows quickly after the main patch.
(In reply to :aceman from comment #13)
> I'm fine with that plan, if the cleanup follows quickly after the main patch.
Thanks. You know that I will do it within two days or less. I let you review the clean-up, OK?
Right now though, with the current bustage we have, nothing will be done quickly :-(
I have a (semi-)green try run for it:
Builds on all platforms.
Comment on attachment 8712835 [details] [diff] [review]
Remove all traces of Eudora (v1) [checked in with some more hunks, comment #19]

Review of attachment 8712835 [details] [diff] [review]:

Yes this should all go.
Attachment #8712835 - Flags: review?(mkmelin+mozilla) → review+
Flags: needinfo?(mkmelin+mozilla)
OS: Windows → All
Hardware: x86 → All
Thanks. I'll land this when the C-C tree opens again. I'll submit more patches to clean up comments and also a patch for SM.
> patches to clean up comments and also a patch for SM.
Thank you.
Part 1 landed:
Note: This has more hunks than the attached patch (attachment 8712835 [details] [diff] [review]) since I cleaned up some more comments as well.

SM patch to follow in the next few hours.
Attachment #8712835 - Attachment description: Remove all traces of Eudora (v1) → Remove all traces of Eudora (v1) [checked in with some more hunks, comment #19]
Part 2: SM. This is straight forward, just a few string corrections and removing as we've done for TB.
Attachment #8735149 - Flags: review?(philip.chee)
Since Aceman complained about more "Eudora" strings, I'd like to note that I left three occurrences in place:

// Send the vCard out with a filename which distinguishes this user. e.g. jsmith.vcf
// The main reason to do this is for interop with Eudora, which saves off
// the attachments separately from the message body
So unless we also want to change the processing, this comment should stay in place.

// See also Bug 503690, where the same issue in Eudora import is reported.

Interesting fact: it's by Peter W. Resnick from Qualcomm (Eudora).
I've updated (at the top) since I have an account there. Aceman, shall we clear the user-doc-needed?
Cool, thanks. But please fix TB 17.09 to 17.0.9 :)
Keywords: user-doc-needed
Done. No idea whether this is the best version to use though. When using "Eudora Rescue", I used TB 1.5. Those were the days, my friend, ...
Comment on attachment 8735149 [details] [diff] [review]
Remove all traces of Eudora (SM part) (v1)

r=me Thanks!
Attachment #8735149 - Flags: review?(philip.chee) → review+
Thanks for the quick review!

Part 2 for SM landed:
Closed: 8 years ago
Resolution: --- → FIXED
Target Milestone: --- → Thunderbird 48.0
You need to log in before you can comment on or make changes to this bug.