So I've got a user that's telling me that there's no message read
pointers being synced correctly with what he's read via the web interface
as opposed to when he's scrolling through them now via text.
I'm not sure, it's been awhile, but it sounds familiar, and that may've been an issue close to, if not the primary, that I had stopped offering web based service at one point, too.
Is this known/handled at all?
So I've got a user that's telling me that there's no message read pointers being synced correctly with what he's read via the web
interface as opposed to when he's scrolling through them now via text.
Is this known/handled at all?
I'm not sure, it's been awhile, but it sounds familiar, and that may've been an issue close to, if not the primary, that I had stopped offering web based service at one point, too.
Yes, it's handled, but the user needs to logoff the terminal server session before the pointers will be updated in the web GUI. They can't
be accessed at the same time and expect them to be in sync.
I don't remember for sure, but I don't think that I ever made ecweb update pointers. I do remember pondering it and seeing various pros and cons.
I wouldn't personally want it to update my new-scan pointers, because the "flow" of how I read things on the web is different (less linear) than I prefer on the console side. I think there's a "last read" pointer which might be useful as a sort of bookmark as you come and go or switch between interfaces.
If you can let me know how you would like it to work, we can see if it's doable. Bear in mind that when browsing a sub in ecweb, the first thread in the list contains the newest message in the sub. Viewing that message & updating the new-scan pointer for that thread would essentially mark everything older as having been read. Unless there's some per-message read/unread flag that I don't know about but would be happy to hear of.
Yeah... Let's see there's got to be something better here. If nothing else, maybe an optional toggle, with a setting saved in JSON, though
that seems like a horrible kludge. :P
Okay, so you're trying to make sure that this is a module tied on to the core, without having any massive change to the BBS internals. I totally understand and agree with that.
mentioned that I was alluding to above was making a checkbox, as fugly
as that may be, with a boolean toggle, letting the user decide whether
or not they want the most recent message that they've read in whichever sub to be setting their last_read pointer. I'll get back to you soon as
better way than a JSON addition to save that option? I only thought
better way than a JSON addition to save that option? I only thought
about the JSON kludge, actually, because I just recently realized that
I'm going to talk with my friend a bit; like I said, he's been doing dev a lot longer than I have. Once he gets done scoffing and pointing
fingers and actually thinks about the extent of the issue I'm pretty
Re: message read pointers betwixt ecweb & synchronet terminal iface
By: Digital Man to Khelair on Tue Nov 18 2014 20:27:43
Yes, it's handled, but the user needs to logoff the terminal server session before the pointers will be updated in the web GUI. They can't be accessed at the same time and expect them to be in sync.
He accessed the message bases via web like a month or more before he logged on via the terminal server. :P The problem was opposite of what
was described here, and they weren't on at the same time at all.
If each user is going to have a preference, it would need to be stored somewhere, and that's normal / not what I would call a kludge. If you mean using the JSON DB service for this purpose, then that would be an unnecessary complication.
I don't mean to imply that it's a big complicated issue or would require a complicated fix. It'd be pretty simple to do, really. I just had several reasons for not doing it in the first place.
Sure, something like that is possible, and I would prefer it being an option than the standard/only behaviour.
It would be simplest to just make a per-user config file to hold ecweb preferences (only created when they change away from the defaults.) That file could contain JSON or be an INI, or whatever else. I believe that the 'data/user/' directory is the typical place to put these things.
There's no need to get hung up on JSON. Really all we would need at this point is to capture a user's preference on this matter, then store it somewhere, somehow, in some format.
Oh, well if that's his attitude, I don't really feel like helping. He can fuck himself.
system so far, so I haven't really looked into things. I thought JSON
would be the primary option available, and as such I was kind of
dreading it. It's good to hear that this isn't the only option to work
For sure. Now that I've thought about it from the aspects of looking into the user & message base object structures, I totally get it.
Yeah I'll look into what's hanging out in that directory in just a bit
here.
Alright let me backtrack just a bit here. *grin* I was typing that in such a manner for a couple of reasons... He's burnt out on software development and he's much more a QA guy, now, and he's excellent at troubleshooting design issues. I shouldn't have said scoffing and pointing fingers. Any of that that _did_ exist was immediately
it was a reason that I originally stopped using the interface in the
first place... It really bothered me when I'd have to switch back and forth. It'd be nice to have at least a little something to hold place a bit... So it's not just him, or him & I, I'm guessing. Other people
have to have run into this a bit. Best wishes.
JavaScript Object Notation (JSON) is fairly easy to work with, and nothing to be dreaded. In fact it's probably what I would end up using for this. Don't confuse JSON itself with the JSON DB, which would be overkill here. One can easily just save a JSON string to a file, then read it back later.
Well, it's not really about that. It's easy enough to update a user's message pointers. I don't have any technical concerns about this. It's just a user-experience consideration.
Look at it this way:
this sub. As far as my pointers are concerned, I've seen all of the new messages in this sub (but I haven't really, just yet - there are seven others that I haven't seen, which won't be caught by my new scan the next time I SSH into the BBS.)
Of course this might be what people want or expect. I just don't think it's ideal. (Though neither is the way it currently ignores pointers altogether.) I can think of a bunch of ways to work with and around this, but they all involve compromises of one kind or another. This is why I shrugged and left things as they are for now.
What I could and probably should do, without complicating things unnecessarily, would be to add a visual indicator of which threads have unread messages in them. It would only be good until you browsed away from the current page, but would at least tell you where to look while you're there.
A previous version of ecWeb had a "new message scan" that was sort of okay. Maybe that could come back. I can't say that I really want to put any work into this right now. I've got other fish to fry.
Whatever you find there may or may not be informative in any way. It's really just a convenient place that's set aside for storing arbitrary per-user data files, as far as I know. (With the exception of the ptrs subdirectory there, but ignore it, because we'd only be operating on those files through abstraction.)
Well, I do care about design and user-experience, which is why I didn't foist something on users that might lead to a confusing diddling of their pointers.
Scoffing and finger-pointing don't bother me too much. That's just a part of the "fun" of this shit. I'm just more tolerant of it when it's about some glaring inconvenience-causing or security-compromising error than when it's about a missing feature. If the suggestion has merit I'll likely implement it eventually, no matter how I heard about it.
That said, "eventually" could be a long time.
ecWeb is not without its multitude of flaws, and I can't say I'm particularly wild about it in general. It does however have a lot of features. If it's missing one and you suggest it, then sure, I'll look into it. If you scoff and point fingers because I didn't think of every possibility right off the bat, well ... fuck yourself, and I'll get to it when I get to it. That, sir, is just the manner in which I locomote.
It's entirely possible that you or other people have raised this issue with me and that I've forgotten about it. I don't however remember any big ecWeb-and-pointers discussions having taken place. If you do stop using something because it's lacking in some way, it never hurts to make a feature request. If you did and I ignored it, well ... heh, sorry. :|
Anyway, I'll see what I can do about this the next time I'm poking around in ecWeb's intestines and lower bowels. If you or your friend have a particular vision for how things would look or operate re: this, please let me know.
JavaScript Object Notation (JSON) is fairly easy to work with, and
I'm actually starting to learn quite a bit of it, in the coding that
I've been doing. Haven't worked with dumping it to and from a string
yet, but I assume there's a pair of methods that make it pretty simple.
The JSON DB is, indeed, where the confusion came from. That was the beast that I was dreading here. Thank you for pointing out the
different terminologies to me.
I totally understand. Actually, I've been hampered in my career search lately because I haven't been working on any web apps, only console and system level stuff... Would you mind if I tinkered a bit myself?
That is a sentiment I can understand and respect. :) Here's my take on
ecweb: best alternative that I've seen so far. I like it, and when I've
got the ability to have web service available, it's the one that I put
up. Right now I just can't have any web servers open due to the
didn't mean that I stopped offering the option; I've left it up, and
multiple people used it while I still had it up. It was just a problem
that I'd noted, so I stuck to text.
Re: message read pointers betwixt ecweb & synchronet terminal iface
By: Digital Man to Khelair on Tue Nov 18 2014 20:27:43
Yes, it's handled, but the user needs to logoff the terminal server session before the pointers will be updated in the web GUI. They can't be accessedat the same time and expect them to be in sync.
He accessed the message bases via web like a month or more before he
logged on via the terminal server. :P The problem was opposite of what
was described here, and they weren't on at the sametime at all.
system so far, so I haven't really looked into things. I thought JSON
would be the primary option available, and as such I was kind of
dreading it. It's good to hearthat this isn't the only option to work
For sure. Now that I've thought about it from the aspects of looking into the user & message base object structures, I totally get it.
Yeah I'll look into what's hanging out in that directory in just a bit
here.
Alright let me backtrack just a bit here. *grin* I was typing that in such a manner for a couple of reasons... He's burnt out on software development and he's much more a QA guy, now, and he's excellent at troubleshooting design issues. I shouldn't have said scoffing and pointing fingers. Any of that that _did_ exist was immediately
it was a reason that I originally stoppedusing the interface in the
first place... It really bothered me when I'd have to switch back and forth. It'd be nice to have at least a little something to hold place a bit... So it's not just him, or him & I, I'm guessing. Other people
have to have run into this a bit. Best wishes.
JavaScript Object Notation (JSON) is fairly easy to work with, and nothing to be dreaded. In fact it's probably what I would end up using for this. Don't confuse JSON itself with the JSON DB, which would be overkill here. One can easily just save a JSON string to a file, then read it back later.
Well, it's not really about that. It's easy enough to update a user's message pointers. I don't have any technicalconcerns about this. It's just a user-experience consideration.
Look at it this way:
this sub. As far as my pointers are concerned, I've seen all of the new messages in this sub (but I haven't really, just yet - there are seven
others that I haven't seen, which won't be caught by my new scan the next
time I SSH into the BBS.)
Of course this might be what people want or expect. I just don't think it's ideal. (Though neither is the way it currently ignores pointers altogether.) I can think of a bunch of ways to work with and around this, but they all involve compromises of one kind or another. This is why I shrugged and left things as they are for now.
What I could and probably should do, without complicating things unnecessarily, would be to add a visual indicator of which threads have unread messages in them. It would only be good until you browsed away from the current page, but would at least tell you where to look while you're there.
A previous version of ecWeb had a "new message scan" that was sort of okay. Maybe that could come back. I can't say that I really want to put any work into this right now. I've got other fish to fry.
Whatever you find there may or may not be informative in any way. It's really just a convenient place that's set aside for storing arbitrary per-user data files, as far as I know. (With the exception of the ptrs subdirectory there, but ignoreit, because we'd only be operating on those files through abstraction.)
Well, I do care about design and user-experience, which is why I didn't foist something on users that might lead to a confusing diddling of their pointers.
Scoffing and finger-pointing don't bother me too much. That's just a part of the "fun" of this shit. I'm just more tolerant of it when it's about some glaring inconvenience-causing or security-compromising error than when it's about a missing feature. If the suggestion has merit I'll likely implement it eventually, no matter how I heard about it.
That said, "eventually" could be a long time.
ecWeb is not without its multitude of flaws, and I can't say I'm particularly wild about it in general. It does however have a lot of features. If it's missing one and you suggest it, then sure, I'll look into it. If you scoff and point fingers because I didn't think of every possibility right off the bat, well ... fuck yourself, andI'll get to it when I get to it. That, sir, is just the manner in which I locomote.
It's entirely possible that you or other people have raised this issue with me and that I've forgotten about it. I don't however remember any big ecWeb-and-pointers discussions having taken place. If you do stop using something because it's lacking insome way, it never hurts to make a feature request. If you did and I ignored it, well ... heh, sorry. :|
Anyway, I'll see what I can do about this the next time I'm poking around in ecWeb's intestines and lower bowels. If you or your friend have a particular vision for how things would look or operate re:
Sysop: | MCMLXXIX |
---|---|
Location: | Prospect, CT |
Users: | 325 |
Nodes: | 10 (0 / 10) |
Uptime: | 05:24:46 |
Calls: | 510 |
Messages: | 220570 |