Closed Bug 1509326 Opened 7 years ago Closed 7 years ago

Reply All Does Not Include Bcc-ed Recipients

Categories

(Thunderbird :: Message Reader UI, defect)

Product:
Thunderbird
Component:
Message Reader UI
Type:
defect
Priority:
Not set
Severity:
normal

Tracking

(Not tracked)

Status:
RESOLVED INVALID

People

(Reporter: mortonspam, Unassigned)

Details

User Agent: Mozilla/5.0 (Windows NT 10.0; Win64; x64; rv:63.0) Gecko/20100101 Firefox/63.0 Steps to reproduce: Sent an email with Bcc recipients. Later clicked Reply All on that email to send a followup. See https://stackoverflow.com/q/53353056/1745001 and https://support.mozilla.org/en-US/questions/1028980 for more details. Actual results: The Bcc-ed recipients were not included. Expected results: The Bcc-ed recipients should have been included, Bcc-ed again.
I stumbled across a workaround for this - see my answer at https://stackoverflow.com/q/53353056/1745001
I'm not sure what you're doing, but "reply all" on a message from the Sent box (which has the BCC header) works as you want it.
If it did what I wanted I wouldn't have opened the bug report. See the screenshots at https://stackoverflow.com/q/53353056/1745001 to see the default behavior and then the behavior if I change that flag I mentioned. I do have 2 email identities if that makes any difference.
Well, you can try on a new profile, create one running thunderbird -p. If you have two identities (or accounts?) you could see whether setting pref mailnews.reply_to_self_check_all_ident makes a difference. I have that set, it helps detect a "reply to self" reply in more cases. I think that only "reply to self" includes the BCCs.
I am using a new profile, it's a new TB installation on a new laptop. That flag IS what I set (again - see https://stackoverflow.com/q/53353056/1745001) to work around the problem, I just can't imagine why that's necessary (and if it is necessary why it isn't documented somewhere). If I have a Sent email that includes a Bcc list then including that Bcc list when I click Reply to All would be the obvious, intuitive behavior since it is Reply to All, not Reply to Some.
Looks like I didn't scroll down far enough in the Stackoverflow post. If you want us to take this further, you need to give more information. Do you have two identities, as you said, or are they two accounts? Where is the sent mail stored for those identities/accounts? In a common Sent box? As Jonathan Kamens, a TB contributor and add-on author confirmed, this is generally working. But in your setup it looks like TB only checks one identity, concludes that the message wasn't sent from the identity and the "reply to self" behaviour is not triggered, so BCCs are not included in the reply.
It's 2 accounts with a common local Sent folder.
I can confirm, that this happens in some cases. I see the logic as follows: If the sent email resides in a folder within the account that belongs to the address in the From header then the BCC headers will also be considered. If the email is in a folder of another account or in a local folder or can not be assigned to any account, the BCC headers will be ignored. Turning on mailnews.reply_to_self_check_all_ident fixes the problem here except in case the mail address can not be mapped.
(In reply to Ed Morton from comment #7) > It's 2 accounts with a common local Sent folder. Then you have no chance to find the sending identity unless you check all of them, that is, set the pref. It will work find if the two accounts use separate Sent folders.
Status: UNCONFIRMED → RESOLVED
Closed: 7 years ago
Resolution: --- → INVALID
Why are you checking identities at all though - I clicked "Reply To All" so why not simply Reply To All rather than excluding the Bccs in some cases based on some guesswork about what a user might want to do differently given some IMHO unrelated and certainly non-obvious aspects of the users configuration?
Or to put it another way - if I have 1 account and Reply To All includes the Bccs (as I think anyone would find reasonable and intuitive behavior), why is it then reasonable to exclude the Bccs from Reply To All just because I added a second account? I see the status on this has been changed to "Resolved Invalid" - I don't agree with that at all, this strongly seems to be a bug to me but if you think it's a "feature" instead then please document that feature and provide a reference to the documentation as I couldn't find any hint of this anywhere.
It's a feature. Take a single account and a message sent from EM to R(ecipient). When you reply to that message, the sender becomes the recipient, so the message would go from EM to EM. Oh, you don't want that? OK, that's why there is the concept of "reply to self". If we detect that, we don't sent the message to the original from, but instead to the original recipient, or recipients. But in order to do so, we need to detect that case, so we need to check your identities or accounts. If you have multiples and use shared mailboxes, it only works with the pref set. The only argument one could have is that the pref should be set by default. There is no bug, it works as designed. If you don't like it, submit a patch and we'll consider it. You can read the code here: https://searchfox.org/comm-central/rev/aa357565819d38429d69190fbcc80f149c6488fe/mailnews/compose/src/nsMsgCompose.cpp#2760 Sorry, I don't have more time for this discussion as there are some *real* problems waiting.
FWIW in that situation I'd intuitively expect Reply to go to the original sender and ReplyToAll to go to all of the recipients but that's not what we're discussing in this bug report so I'm not sure why you're bringing it up here. In this bug report we're talking about whether or not the Bcc list should be included when you Reply-To-All. Despite how it may be coded, that functionality is not intuitively related to who a single Reply should go to and the current implementation, whether by design or not, just doesn't make any sense so I'd have hoped you'd be willing to fix it or at least document it. Since you apparently don't think this is a real problem I expect if I coded a fix you'd dismiss that too so I'm not going to bother, I have a workaround and I've posted what that is on Stackoverflow so at least this bug is documented somewhere with a workaround.
Even for "reply to all", the system must detect a "reply to self" to eliminate the original sender from the list of recipients. Do we agree on that? As the code I quoted shows, we only add BCCs if "reply to self" is detected. Other received messages should NOT have BCCs, if they do, something has gone wrong and we don't want to promote that error further. Magnus, you're the expert on what should be sent to whom (see also bug 968270), so you might want to comment here further. While we're here, any opinion on changing mailnews.reply_to_self_check_all_ident to true by default?
Flags: needinfo?(mkmelin+mozilla)
Personally, I would find Reply To All simply replying to all, including the original sender, fine since I did say Reply to *ALL* but I don't have a problem with the email address of the current sender being removed from the list of recipients. idk why that would require looking at email accounts and folders or doing anything other than just looking at the address in the current From field and removing it from the To, Cc or Bcc field though. I guess if I really thought about it I do have a slight preference for TB removing me from the distribution list but I really don't care much - as long as *All* the original recipients in the To, Cc, and Bcc lists get it when I use Reply To All, then I'm good. Now that you've told me you're trying to avoid propagating incorrectly received Bccs, even though I disagree with the approach, I at least can understand why the functionality might have been coded to not simply always include Bccs on a Reply To All. I do find the behavior I get from TB when reply_to_self_check_all_ident is set to true far more intuitively obvious and much more likely to be what a user with my configuration (I'm not clear if it's having multiple accounts with a common local Sent folder that's the issue or if just a single account with a separate local Sent folder would have the same issues) would want though and again, it's completely non-obvious and undocumented as far as I can tell why TB behaves as it does when that flag is false or that setting that flag to true would produce the desirable behavior. So FWIW I'd like reply_to_self_check_all_ident to be set to true by default since that solves the 2 PITA problems I've been struggling with for weeks, see https://stackoverflow.com/q/53353056/1745001, and produces the behavior that everyone has been telling me I should simply get by default every time I ask anyone about it!
(In reply to Jorg K (GMT+1) from comment #14) > any opinion on changing mailnews.reply_to_self_check_all_ident to true by default? I guess we could yes.
Flags: needinfo?(mkmelin+mozilla)
Are you going to change the default for that flag?
Great, thanks!
You need to log in before you can comment on or make changes to this bug.