Alicornutopia

I agree, especially about layout

I just decided to try using the Constellation's unread-tracking features instead of keeping lots of open tabs and manually syncing with the daily report.

Is there a way to see a list of all of my “opened” threads, including those with nothing unread?

Got another Lexile blip just now while selfthreading in they broke the mold

Man so many things to reply to, here goes:

If my Ruby processes allocate themselves more than 512MB of memory, Heroku starts throwing R14 errors, no one wants errors.

I can probably do a streaming reply, it's a thing all my separate pieces each conceptually support, but I've never done it before personally and brief Googling is not readily yielding Rails 3 + Heroku + Puma server + streaming responses, so it'll take a while.

Filter/sort by # unread is already part of issue #103, as is by author, but I added by # total replies.

I have no idea about the certificate errors and continue not to have touched anything. I've been hoping it's a brief incident with one of my providers; if someone reports it again, I'll open tickets with my host (Heroku) and my DNS provider (PointDNS) and my domain host (NameCheap), can probably skip the SSL issuer (DNSimple) unless the others have no leads, aaaand it's possible this site has a lot of moving pieces, jeez.

I'm using a prebuilt tag UI called Select2 so I did not in fact build that thing but I'll add it to the list.

Stuck the StreamingResponse and precomputed view=flat and mobile Load Only (But All) Unread buttons into a single "wtf do I do about view=flat" issue. Once I have solutions covering internetless browsing, searching the thread for specific text, and downloading the thread as a backup, I anticipate limiting max per_page to something much lower than the existing version, likely 50 or 100. But no touching without solving those first.

Made tickets for account UI and opened posts.

In other news:

This morning we achieved our first 24 hour period with no memory errors! (This may in part be due to me occasionally needing to restart for deploys; this is the first time we went 24 hours and hit the auto Heroku restart. We've gotten really close and generally hang out above 95%, so this is not like Case Closed Forever, but it's probably stable enough that I could theoretically look at other things.

Also, I'm going through a really miserable cold. I've cycled through fever and chills, headache and sore throat, miserable cough, and am now onto the "horribly congested to the point where it makes me tear up constantly and my eyes hurt and I can't look at screens". I'm presently drugged enough to have managed responding, since I didn't want to get too behind on it, but I could do a response at a time then take a break; actually doing things that require uninterrupted screen staring (like coding) is so extremely on hold till I'm functional. Apologies to everyone and in particular Throne, who is wonderful and very helpfully sent me many pull requests I'm in no shape to review.

Haven't posted yet but Kappa just posted so preemptive edit: I'm submitting tickets about the Lexile thing now, this is ridiculous.

*hugs a marri*

For those who have seen the Lexile/Bluefin thing, I would greatly appreciate if you remember/could tell me if you were using www.glowfic.com or straight glowfic.com.

Straight glowfic.com for me, the latest time, but when I checked the rest of my tabs it turns out I usually use www, so that was probably where I was the previous times.

I usually use straight glowfic.com and expect that's where at least one of my errors happened.

I don’t see how NameCheap could have anything to do with it; registrars are basically pure middlemen between you and the WhoIs database. NameCheap does provide a DNS service but you don’t use that. I’d focus on Heroku and PointDNS for now.

*Hugs*

I submitted tickets to Heroku and PointDNS already; past!me agreed with you about the unlikelihood of NameCheap being remotely relevant.

I expect Heroku to blame the fact that my glowfic.com record "appears to be an A record" (quoting their DNS check link), despite the fact that if you follow the link they provide in the error message it explicitly says to use ALIAS with PointDNS. I look forward to the argument, particularly given that I set this all up almost two months ago and (unless you guys have been weirdly hiding things from me) we have seen no problems until four days ago. This is why I asked about glowfic.com vs www.glowfic.com.

I also do expect this to be caused by my Heroku SSL somehow; PointDNS was mostly for completeness.

Straight glowfic.com for me as well.