cfml

And beware that those who did this before the Dec and may updates found it to be removed. Cf updates always remove any such hotfix jars

ok, so basically it has to be the &lt; .23 driver

All this is indeed frustrating, most important that the may and Dec updates had NO bug fixes. I was livid when I saw that in May. Most people (even Brad) felt I was overreacting and that it was "good" that Adobe now offered sec-only updates. I've long argued they should offer both

Yes, on Mysql driver. That's where the problem started. It's a myslql change, that Adobe had to accommodate, and did.

It basically makes serializing MySQL queries impossible

So is the fix solving things? You just restart cf, of course

Not related to serializing at all. Just dump the query and you see any dates there, alteady, as an object. Will affect ANY subsequent use of that date col in cfml

no, what I am saying is that if you do `serializeJSON( query) `

But with 23 or up, didn't the hotfix solve the problem? It should

because it tries to recursively serialize the java localDateTime object

Luis, I said ANY processing of that date col will fail. It's not UNIQUE to serializing

Just so folks are aware, who might not have read the MySQL JDBC driver release notes in depth, there was a big change in the 8.0.23 driver compared to the 8.0.22 driver in the area of date/time fields. In theory, it's a bug fix but the reality is that a lot of code depended on the bug so this _patch-level_ release is a breaking change for a lot of people. We're stuck on 8.0.22 at work, completely unrelated to CFML, because we haven't had the time/energy to dig into our codebase and make the changes necessary for this!

See <https://dev.mysql.com/doc/relnotes/connector-j/8.0/en/news-8-0-23.html#connector-j-8-0-23-feature> for more details. They changed the _default_ behavior  for timezones and you need to add a new connection parameter but they also changed the return value for `getObject()` that requires code changes!

Always applicable to semantic version conversations: <https://xkcd.com/1172/>

Gotta love breaking changes in patch releases...

I have to say, it's one thing I really like about the Clojure community: they value backward compatibility to the point that instead of introducing breaking changes, they will add a new function or package with the modified behavior, allowing users to adopt the changes as and when they want while still guaranteeing existing code continues to run (and libraries tend to be very small and focused in their functionality so keeping that "legacy" code around has very little cost). There is a lot of Clojure code out there written against, say, version 1.3.0 from a decade ago that still runs unchanged on 1.11.1 (the latest stable version) -- and can usually have all its 3rd party (Clojure) libraries safely updated too...

Please use threads for conversations like this :slightly_smiling_face:

So don't refer to it as a serialization bug in your talk. It's not, at all. Again, the problem is in the query result

Anyway, Luis, it sounds like you're opting not to try the hotfix, is that right? Any compelling reason, since your doing something for a talk? Folks using Mysql and drivers from the past year SHOULD apply that hotfix. Never any reason NOT to.

Maybe I misunderstood you. Is there an actual hot fix?

I replaced the driver manually for now so I could continue to work. But I did not see that there was a hot fix provided

I said previously that that tracker link you offered had the hf jar (for cf2021). You asked where to put it and I said cf's lib/updates. You said you were looking and found it. Or were we miscommunicatinf then? This problem HAS that fix. Folks should NOT stick with the old driver, for its own sec issues and bugs, long since fixed.

I was talking about the jdbc driver itself

Would be great if I could say `cfpm update`

and I could patch it like we do with CommandBox

Sure, but there's no adobe patch repo, and many are not even offered in tracker, so will likely never happen.

You don't offer such "patches" (special hotfixes) for cf in commandbox, right?

Am about to start a troubleshooting session with a client, so won't reply for a while