1. pastebin crash
- Posted by katsmeow 3 months ago
- 426 views
I just tried to paste some code, and found more OE stuff is broken : https://openeuphoria.org/pastey/create.wc
Internal Error
Fatal run-time error: Couldn't insert new pastey: Incorrect string value: '\xC3\xA7ais:...' for column 'body' at row 1
2. Re: pastebin crash
- Posted by katsmeow 3 months ago
- 381 views
Can someone let us know when this is fixed?
Thanks, Kat
3. Re: pastebin crash
- Posted by ghaberek (admin) 3 months ago
- 373 views
Can someone let us know when this is fixed?
I'm sure once someone looks at they can give you a time estimate on getting it fixed. But as it stands, I don't know who that would be, so I don't know when that will be.
-Greg
4. Re: pastebin crash
- Posted by jimcbrown (admin) 3 months ago
- 367 views
Can someone let us know when this is fixed?
I'm sure once someone looks at they can give you a time estimate on getting it fixed. But as it stands, I don't know who that would be, so I don't know when that will be.
-Greg
At least I can give a hint as to what's wrong. The error comes from MySQL and it seems that the fix requires direct database access to update the character set to support full four byte UTF-8 characters.
(Alternatively, we could add a line of Eu code in to replace the problematic lines with question marks or something, but that seems less than ideal.)
5. Re: pastebin crash
- Posted by katsmeow 3 months ago
- 361 views
The text i was trying to upload was 100% compatible with 16/32 bit win95 and 8-bit text editors of the day. And being it contains an irc interface, it was and is acsii/ansi compatible as far as i know. I see no reference in the error code to clue me in on what to not do.
Kat
6. Re: pastebin crash
- Posted by jimcbrown (admin) 3 months ago
- 357 views
I see no reference in the error code to clue me in on what to not do.
Kat
Yeah, understandably frustrating. We're not very good at taking the MySQL error and making it into something more sensible. And furthermore it shouldn't have been an error at all, that sort of thing should have just worked.
The text i was trying to upload was 100% compatible with 16/32 bit win95 and 8-bit text editors of the day.
That's true. The problem, as described in the link I posted earlier, is that the default character encoding used by MySQL doesn't support certain 8bit values. Some 8bit values are not allowed under UTF-8 at all; others do work but require more bytes than MySQL allows for. Meanwhile 95 and the 8bit text editors of the time easily supported all possible 8bit values either under some kind of "extended ASCII" guise or else some other encoding like Latin-1. These are oddly more permissible (any combination of 8bits is valid) but also have fewer overall characters than UTF-8.
And being it contains an irc interface, it was and is acsii/ansi compatible as far as i know.
I'm guessing 8bit extended ascii? Something 7bit acsii compatible shouldn't have triggered this sort of error. But I vaguely recall that some extended 8bit values were sometimes used for screen control..or to render some slightly fancier graphics on what was otherwise a text console.
7. Re: pastebin crash
- Posted by petelomax 3 months ago
- 333 views
The offending character is c with cedilla:
Windows 1252: https://bytetool.web.app/en/ascii/code/0xe7/
Unicode: https://www.compart.com/en/unicode/U+00E7
The hint in the message is it is the character just before "ais".
Apparently in his day and age, only 1.6% (global average) of the internet copes with Windows 1252 characters: https://en.wikipedia.org/wiki/Windows-1252 (Of course, replacing it with '?' would be better than crashing)
I reproduced the error with "çais" even though I'm pretty sure I'm using the proper utf8... so I now doubt it has anything to do with Windows 1252.
8. Re: pastebin crash
- Posted by katsmeow 3 months ago
- 318 views
The offending character is c with cedilla:
Pete got it right, i opened the (potential) upload in Textpad and did a "search for ais", and found
set %Bable.tag.1 *En Français:*
There's also, for whatever future view warning this is worth:
set %Bable.tag.4 *Em Português:* set %Bable.tag.5 *En Español:*
I spose i could intentionally trash those lines, since the site and the intended environment to run the code no longer exists anyhow...
Kat
9. Re: pastebin crash
- Posted by katsmeow 3 months ago
- 311 views
I'm guessing 8bit extended ascii? Something 7bit acsii compatible shouldn't have triggered this sort of error. But I vaguely recall that some extended 8bit values were sometimes used for screen control..or to render some slightly fancier graphics on what was otherwise a text console.
The olde mirc versions used 8bit "control codes", i know this because i wrote a fairly complete irc client in Eu. It was "fairly complete" because i never got a useful Windows window, despite begging for assistance. I used a dos console to display, which was nice enough to have a multiple screen line entry partition with word wrap. It was quite sufficient for a bot, and could use mirc solely as a scriptable "plug-in" gui.
For a time, "leet" non-english characters were the rage with some people. This was bad enough i wrote a translator to convert that to english too.
Kat
10. Re: pastebin crash
- Posted by katsmeow 3 months ago
- 316 views
It uploaded, but the text formatting is way off in places, and one control code is not displaying.
Kat
11. Re: pastebin crash
- Posted by jimcbrown (admin) 3 months ago
- 291 views
It uploaded, but the text formatting is way off in places,
That kinda makes sense, considering how different a web forum environment is from the text consoles of heydays past.
and one control code is not displaying.
Kat
That's worse than having it replaced by a ? I suppose - a casual reader might not even realize something is missing.
Pete got it right,
+1
I spose i could intentionally trash those lines, since the site and the intended environment to run the code no longer exists anyhow...
Kat
Or you could preserve them in this thread. I mean they've lasted this long. Either way.
I'm a bit surprised that it works fine in the forum but not in pastebin actually.
The olde mirc versions used 8bit "control codes", I used a dos console to display, which was nice enough to have a multiple screen line entry partition with word wrap. It was quite sufficient for a bot, and could use mirc solely as a scriptable "plug-in" gui.
Ah yes, makes perfect sense now!
For a time, "leet" non-english characters were the rage with some people. This was bad enough i wrote a translator to convert that to english too.
Kat
Smart move. Reminds me of that judge who found rap slang so unreadable as to comment that it should be a different language.
