The list of available banks could not be retrieved... forever.

I believe this still doesn't fully solve the problem: after ten minutes, the API should perform a new directory request, regardless of the cached data. Because the patch does not only check the cached data's expiration date, but also checks the actual data, it does not invalidate the cache item if the data is valid, but expired.

1) Yes the patch does solve the problem at hand, because if the data is valid (but expired) then the problem at hand does not apply.

2) This is what I pointed out in my last paragraph: you can delete the checks on the actual data if you want. (In case you want to make sure that lists of valid banks are only cached for a minimum time period. That's your decision, as far as I'm concerned. By the way, in that case the expiry time is not 10 minutes but 1 day. But that's detail.)

I thought my bugreport was clear enough in implying that, wasn't it?

---

, but I'm very much inclined to see if we can't fix this issue in core instead.

Fix what, and how? You don't give a hint to what you mean and even say in the next line that there is neither a bug nor a hiatus in the documentation of the current behavior.

---

As a general remark, maybe it's just some kind of language/attitude barrier... but I just don't get how you expect bug reports to be followed up. I would try to cooperate usefully if I got it, but...

In the first part, you rephrase something I already said, but then portray it as a problem that I had not thought of. In the second part you say 'should be fixed elsewhere' but not why or how. Both have the effect of making me completely unsure of what to do now - and halting the path to a resolution. Saying the fix is wrong and/or the direction should be different is cool, but there's always a "BUT" following which takes away any concrete direction.

Same as a previous report.

If you bring suggested fixes to a halt because you don't want them... and prefer I just fork the module privately... that's fine. Just say so. It's less work than hashing out fixes in issue queues.

OK, sorry for being snappy.

What remains, though, is that I have no idea what to do now to resolve this (or if we're even aiming to resolve this issue), until I know what this means:

fix this issue in core instead

(I'm hoping you are not waiting (for someone) to ask for Drupal Core to change the code and the specs of the current caching system, before allowing this fix in iDEAL API.)

Being busy happens to 100% of module maintainers. (Glad I'm not one, I just trawl random issue queues ;) )

Until that time, this issue is fine as a temporary fix for whoever encounters it... as long as it doesn't say "does not work" when in fact it does :p

Payment sounds nice! and I'd be interested in chatting about related stuff in person. I wish I could be there without giving up a full day for traveling back & forth, unfortunately I'm at too much of a distance to NL at the moment.

@Jawi: they are only solved after applying the patch, which is not in the -dev version yet.

@Xano: I am going to re-try asking questions to make this concrete.

1)

Do you still want to wait and (quoting #1) "see if we can't fix this issue in core instead"?

You have 2 options for an answer:

a) No (because I realise that the behavior of the D7 caching system -despite being confusing- actually works as specified, so I realize that nothing will get fixed.)

b) No (because even though I don't realize point a, I accept that such a behavior change will never be introduced in D7.)

:)

Seriously, I could hunt down issue numbers and other discussions for you, but I hope I don't have to. The caching system in D7 is braindead (in some aspects), but not broken. You'll need to work with the system as specified/documented.

2)

You mentioned that my patch does not "invalidate the cache item if the data is valid, but expired.". This is by design; it follows Drupal's standard caching behavior. IMHO Drupal's standard behavior is only a problem for the cached 'invalid' data.

But do you want this anyway?

a) yes, I do want to invalidate valid cached data immediately after 'expiry time' (which means treating my valid list of issuers differently from any other data in the {cache} table)

b) no, I realize that the caching system specifies a "minimum cache lifetime", and I don't mind treating my valid list of issuers just like any other data in the {cache} table

OK. So in the linked issue, I see you implemented an equivalent of this original issue summary/patch, which is apparently considered valid for 2.x in hindsight.

So since noone is interested in 2,x anymore, I guess this issue can now die.....................