Nils Goroll

generally in favor, have not looked closer yet, still :(

@mbgrydeland can you please explain why you think the ordering of the STV_Ban* calls would change? (FWIW, I would guess you actually care about the `storage_baninfo_f` ordering, do you?) I...

@mbgrydeland I understand that the mutex in current code enforces strict serialization. I would think, however, that the order of bans, at least as persistence is concerned, is very well...

Summary of bugwash: Because there is not much documentation for it, the stevedore code defines its API. by this interpretation, this PR constitutes an API change. @mbgrydeland has no particular...

FTR, Weak validators can not be used for [`If-Match`](https://www.rfc-editor.org/rfc/rfc9110.html#field.if-match) requests

I would not see how, The VCL_PRIV can not be reached through the VRT_CTX, it is a static in the vcl object. One obvious possibility was to create an association...

Yes, if that function exists. But the filter interface is designed that a simple ``` set resp.filters += " myfilter" ``` is sufficient. And there is no vmod function called...

@Dridi if we go with parameters, shouldn't those constitute aliases in the `storage.*` VCL object namespace? And shouldn't they be changeable at runtime? e.g. ```varnishadm param.set transient_storage storage.mystore``` -> VCL...

@Dridi so we have now three parameters on our list * `h2_rxbuf_storage` * `shortlived_storage` * `transit_storage` I am not sure where I possibly misunderstood you, I can see your point...

@Dridi it absolutely would work for me and I had the same idea yesterday, but thought it would be reaching too far. But now that there's two of us...