Thanks for the detailed response, Jack!
On Thu, Jan 22, 2009 at 19:07, Jack Repenning <jrepenning_at_collab.net> wrote:
> On Jan 22, 2009, at 2:24 AM, Greg Stein wrote:
>> In random order while trying to fix Bert's posting issues (mentioned via
>> * when admin'ing dev@ list, to add a subscriber, hit "Browse Users"
>> ... the filter doesn't work
> Can you describe more what you mean by "doesn't work"? I know of one
Eek. Pilot error. I typed in "Bert" and got *60* pages of stuff.
Seemed like it did nothing. BUT! Now that I look at it... sure enough,
there are lots of "Robert" and "Albert" and stuff like that. My
mistake was assuming it was somehow restricted to the Subversion
project. It appears it is "everybody" known to the site.
When I filter using his last name, I get a respectable two lines.
Sorry 'bout that.
>> * tried to subscribe one of bert's secondary email addresses, but it
>> says "already subscribed", yet I can't see/filter on that in the
>> subscriber list
> By "secondary email addresses," do you mean "the extra addresses in his
Yes, we figured that one out.
> That is (as you might not yet have noticed), a registered user can
> now provide a list of email addresses that are all "me," and all of them get
> the same rights (ability to post, for example). But a "subscription" (that
> is, by a registered user) will always send the mail to the "primary"
> But conversely, as mentioned just above, there's this distinction between
> "addresses known to belong to registered users" (aka "subscriptions"), and
> "addresses not so-known" (aka, "anonymous guest email addresses"). Maybe, by
> "secondary email addresses," you mean "an address Bert hasn't added to his
> profile," in which case it sounds like you're running, again, into that
> "filtration ignores the latter" problem.
In this case, his emails were being held up, so I attempted to add the
address he was using to send. It said "already there", yet I couldn't
set a filter to *show* that secondary address to me. So it *looks*
like his other address is not Trusted because it doesn't show up. But
it is actually in there somewhere.
(in this case, you figure out that tigris knows about it when you try
to add it, and it says it is already there)
>> * for bert's record, I tried to alter the Trusted flag -- the changes
>> don't take
>> * to actually *get* the Trusted flag set, I had to unsub and resub Bert
> Several people have reported this. In all cases to date, it turns out that's
> because they clicked the "submit" button, just below the "trusted" check
> boxes. As it happens, that has nothing to do with the trust settings, only
> to do with the adjacent "number of lines per page" control. To make your
Yeah, I figured that out after I punched it the first time and nothing happened.
> trust changes stick, you need to click the "Save changes" button, a line
> lower and way over at the other side of the window. Very bad UI. But I think
Woah. Didn't see that.
> you'll find it works. Fixing that UI confusion is also in the queue.
I'd suggest moving the "per page" controls over next to the pagination
controls (imagine that!), and the Save Changes where the per page
stuff is. That should go a long ways to help.
Even better: use Ajax for that Trusted flag :-) (not quite "just a
template change" tho :-))
>> * Bert posts from one of N addresses, but they all get converted to
>> From: his one profile address (not desired; it isn't an "active"
> Can Bert change his profile primary address the the one he wants to be known
> as? That is: Bert can certainly change his primary profile address, as well
> as provide others if that's useful; my question is, is Bert's life simple
> enough that he can afford to be known by a single address?
Nope. He uses other email addresses in his daily life, but prefers an
@sharpsvn.net for (say) commit emails. But the @sharpsvn address is
just some light branding -- he doesn't use it for anything.
> If this boils down to "Bert should set is active address as primary," then
> there's the solution.
> If this boils down to "Bert needs different addresses for different
> purposes," that would be an enhancement, which is not (so far as I can tell)
> presently in the queue. Poke me, and I'll put it there. In the mean time,
> there might be a work-around, such as *not* putting some of the addresses
> into the profile, but rather "anonymous guest email address" subscribing
> them instead. Bert, if you're within the sound of my voice and want to talk
> about that, drop me a line.
That sounds like a reasonable workaround to prevent the From: address
rewriting that is occuring.
All that said, I would say that rewriting should never be done, regardless!!
>> Then we come back to the "convert to attachment" thing. While it may
>> appear that the mailing list is fixed in that attachments finally go
> Whoah! I never ever EVER said it was "fixed"! It's still massively
> busticated! What I said was only "data is no longer being lost." We have
> several engineers feverishly working on the rest of the breakage.
hehe... I think we're saying the same thing. "fixed [... go through]"
as in "doesn't get vaporized".
> And, as we discovered yesterday, we seem to have re-broken the "data getting
> lost" part recently; as of the last report I received, we were restoring
> that fix and trying to understand how it got broken again.
>> I found that the web-based list browser still shows them as
>> attachments (tho hard to tell if this example is auto-converted or
>> what). And it forces you to download the thing, so some messages are
>> simply blank unless/until you download an attachment to your box in
>> order to simply read a posting. For example:
> Yup. Yup. Yup. And for what it's worth: even message-by-message subscribers
Okay. For some reason, I had thought I remembered somebody saying
"well, the web archive browser makes this better". But... well.
> (who do get the attachments, or will once again when we re-fix that) get
> highly variable display, depending on which mail client they prefer to use.
> In some cases, it's quite a pain to get to the attachments; in other cases,
> they may not even know anything's broke. I mention this only to emphasize
> again that I do NOT think it's "fixed," only "improving," and "not actually
> LOSING data."
Yup. I meant the same thing, but with (obviously) ambiguous langauge.
> Obviously, that's a decision for the community to consider. I am actually a
> very minor member of this community, but probably not objective on these
> matters since I'm also a representative of the service provider; FWIW, and
> assuming we can re-fix the "lost data" part, it seems to me we're at the
> "flamingly annoying, but I can just barely still get my work done" stage of
That's about it. But all these little bits are adding up :-(
Received on 2009-01-22 20:28:02 CET