AFTER supported y/n?

[markus_petrux]

Ah, there will probably be a good oportunity to scan all MODs for this kind of issues ...when Olympus...

[TerraFrost]

If someone tries to run an SQL query with a reserved word in it, the query will still be parsed, but an error will still be thrown, none-the-less.

Also, if someone has a query with TYPE=MyISAM, it'll be removed and a warning will be thrown.

Why not do the same thing with AFTERs? Parse the query but throw a warning? It seems like throwing a warning would be just as effective at getting people to write good SQL queries as halting the conversion would.

Alternatively, it seems like a more meaningful error message could be made when AFTER is used. Perhaps the justifications made in this thread could be given when an ALTER TABLE containing an AFTER is used?

[markus_petrux]

hmm... I may try to improve the parser so it generates more understandable warnings/errors. That's not easy though. Look for example the errors reported by MySQL itself "You have an error near ...".

However, these things should be detected by MOD authors in the first place, so they may have more knowledge to diagnose why their MODs don't work as expected, etc.

[-=ET=-]

However, these things should be detected by MOD authors in the first place, so they may have more knowledge to diagnose why their MODs don't work as expected, etc.

I didn't comment your decision before as I fully respect it, but you know I'm quite sure nearly 100% of MOD authors have no clue if an equivalent of AFTER exist for Postgres, Access or MSSQL, and I'm quite sure too nowhere the phpBB group ask not to use AFTER. So nothing will be detected, as they don't care about use AFTER or not.

But now we know, it's not that important (even if personally I would have preferred to have the possibility to use it, even if for an other database the fields are added at the end of the table), so that's fine

[TerraFrost]

I'm quite sure too nowhere the phpBB group ask not to use AFTER. So nothing will be detected, as they don't care about use AFTER or not.

Authors who've submitted their MODs for validation (atleast in the past few months) and whose SQL queries use AFTER usually get comments from us