Truncated Posts

[Spider]

11 hours ago, jetsetdanny said:

It got truncated - as might be expected, with the benefit of hindsight - at the very spot where I used a diacritical Czech character. More of them certainly followed.

Yes. The missing part (and just the missing part only) was:

ální jeskyně" (to be precise, with no diactritical marks and capitalised, so "CENTRALNI JESKYNE"), which is Czech. The Slovak name would be (if Google is to be believed) "Centrálna jaskyňa" (which actually shows you nicely how similar and at the same time different the two languages are).

11 hours ago, jetsetdanny said:

I think these characters will have to go, but if the post could be restored to its full content, it would be appreciated 🙂 .

Its been restored 🙂 Check here

No, no need for them to go. Because now we are on utf8mb4 which allows 4bytes per "character" vs utf8 which was 3 bytes, and in a previous incarnation a few years ago I think it was iso-8859-1 (I think, I can't remember that far back!)

 

[JianYang]

1 hour ago, Spider said:

Because now we are on utf8mb4 which allows 4bytes per "character" vs utf8 which was 3 bytes

I don't know if this is relevant, but all three characters have 2 byte utf8 codes. You might have a different problem than that.

 

[Spider]

5 hours ago, JianYang said:

I don't know if this is relevant, but all three characters have 2 byte utf8 codes. You might have a different problem than that.

 

Its is (well , it was) slightly more involved. The forum started originally on ISO-8859-1 (I think) and after a couple of years I ran a script to convert it to 'regular' UTF8, which allowed more characters to be accepted properly in the database (mysql being upset as usual) , this was in effect v3 of the software, although I did upgrade it as/when required.

At this point in time, the 'data' was all there, or so it seemed. It turned out that some but not actually all posts (data) were effected. Have to bear in mind I'd also upgraded php/mysql a couple of times too.

Later on after the sites small unintentional 'vacation' , I upgraded to a newer software version, this would of by default demanded the data be in at least UTF8 format, if not UTF8MB4 format. I chose the latter even though it took more db space it did guarantee at least what I thought was a problem free way. It turns out upon inspection (I did post a section of the code way back) this contained some odd bugs, causing it in effect ignore a 'special character' and although that was not ideal, it then ignored the rest of that data, resulting in the truncation we sometimes see.

^ It is a bit more involved than that, but I wanted to skip the detail and just provide a summary.

To be fair once I spent a fair bit of time fixing the majority of the issues on this, the remaining ones are a case of "as and when , please tell me" as Danny posted above. 🙂

 

[JianYang]

I apologise that I made you type out the whole involved story.

 

[jetsetdanny]

Thanks for restoring my truncated post, Andy! 👍

[Spider]

12 hours ago, JianYang said:

I apologise that I made you type out the whole involved story.

 

It's not a problem at all. 🙂 Please do not worry about that. The 'issues' would of likely of been worse had we at some point switched between different software platforms rather than 'stay with one' and merely upgrade.

8 hours ago, jetsetdanny said:

Thanks for restoring my truncated post, Andy! 👍

Most welcome. Don't worry about posting 'special characters' , there should not be any issues anyway since the previous (from @IRF's original enquiry) a few months ago now with anything 'new' from that point onwards, its only the small handful of ones remaining that may need attention but as I think I mentioned at the time, from a cursory glance I could not see anything else amiss and it would be near impossible to read everything! 🙂

 

As an aside if anyone does note anything amiss (this is likely only for any posts previous to the middle of last year or older, nothing newer) please feel free to either post in this topic about it or if you prefer you can send me a PM instead.