21

I tried to reply to a user by way of comment on my answer, which normally I do all the time to make sure they get the comment, but this time, when I tried, I got the generic "Comments cannot contain this content" error.

It appears that their name may be the issue (it's "Алёна Опарина"). The system allowed me to auto-complete the name when I typed @, but when I try to post it, it gave me this error. This is question that was posted.

I also tried several variations of the comment, and all of them were rejected, so I suspect that this is a problem with the username and not anything I've tried to say. I've also never had this problem with similar comments on any other comment I've written, hence why I'm asking.

Is there somewhere we can see the list of disallowed phrases? Should the system block a comment because of the username?

enter image description here

3
  • 1
    As a third party, I tried and can't post it, either. - And I used different text, so it's not the text of the comment, either.
    – Catija StaffMod
    Commented Jul 18, 2020 at 18:31
  • @Catija could it be site specific?
    – VLAZ
    Commented Jul 18, 2020 at 18:33
  • @VLAZ Yeah. If they've had issues in the past with Cyrillic, maybe we put in some special casing... It'll take a dev to look at.
    – Catija StaffMod
    Commented Jul 18, 2020 at 18:35

1 Answer 1

17

Thanks for bringing this to our attention!

Back in January 2014, Salesforce was hit by an influx of Russian Spam. The same day the meta post was created, we added Cyrillic characters to the blacklist on the Salesforce site to block that spam.

Screenshot from SEDE of the blacklist entry for Cyrillic text being blocked.

I'm guessing this was intended to be a temporary block as our spam responses are generally designed to make spamming difficult until the spammers move on to somewhere else. In the time since the blacklist was established, it blocked something just shy of 2000 attempts. That's about one per day and probably a handful of those have been people testing it today alone.

Based on that, I'm guessing that we can remove the blacklist and see if the spam picks up again. I'll dig into the data a bit more to see who's hitting the blacklist to double check and then will likely recommend removing it.

I'll see if we can get this fixed next week.


Some additional data I pulled:

Y'all did used to have a big issue and there was a smaller attempt at posting in Cyrillic in 2017 but, for the last two years or so, it's been about 15-25 attempts per month

Graph showing the attempts to post Cyrillic characters on the site over the last six years. There's a big spike over 1000 in 2014 and a smaller one up to 200 in 2017 but relatively low numbers per month otherwise

Of these events, 175 were attempts to ask a question and 50 were attempts to post an answer. The most frequent reason to hit the blacklist were attempts to edit - 2195 edit attempts, most of which happened on five posts which attracted 660, 593, 492, 203 and 105 edits each.

Additionally, only 58 attempts have ever been in a comment - all time - on 20 total posts. Six of which were related to this specific event and on another post, someone attempted to leave a comment 15 times.

Some things have changed since y'all had this blacklist in place - we have the community group, Charcoal, and their bot, Smoke Detector, scanning for and removing spam. So, considering the above, do you think the complete blacklist - or any - is necessary? I'm leaning towards get rid of it all and if there's a huge resurgence, we can re-establish part of it but in a way that doesn't overly-block this content.


Adam was kind enough to remove this block so Cyrillic characters are now allowed. Please let me know if y'all run into issues!

7
  • Thanks for the follow up! I just wasn't sure where to take this, next time I'll know for sure. 😊
    – sfdcfox Mod
    Commented Jul 19, 2020 at 0:24
  • I noticed you marked the original Meta Stack Exchange copy of this question as status-review. Mind doing the same here?
    – gparyani
    Commented Jul 19, 2020 at 1:38
  • 1
    @gparyani I do mind. Marking it here will make a second ticket about the same issue. The first ticket will still link here, so tagging is unnecessary.
    – Catija StaffMod
    Commented Jul 19, 2020 at 1:46
  • 1
    FYI industry standard terminology is shifting away from "blacklist" and towards "blocklist".
    – Adrian Larson Mod
    Commented Jul 21, 2020 at 13:35
  • @AdrianLarson I'm aware... but for the time being, it's the literal name of the feature and until we change it internally, I kinda have to keep using it. :)
    – Catija StaffMod
    Commented Jul 21, 2020 at 13:38
  • 2
    Fair enough. It was really surprising to hear about at first and culture shifts take time and many small interactions, so I just figured I'd mention it. Thanks for chiming in here. Always appreciate you guys.
    – Adrian Larson Mod
    Commented Jul 21, 2020 at 13:42
  • 3
    @AdrianLarson And we are having those discussions! I think we have actually made changes in some terminology internally - things like master/slave... I'm guessing this is one we just haven't discussed but I'll make sure we talk about it. I do appreciate you mentioning it! :D
    – Catija StaffMod
    Commented Jul 21, 2020 at 13:43

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .