Hi, Thanks… I had a terrible time getting the mail client to behave and failed never the less.
Attached are the split patches as requested. Hopefully it makes it through the email. Thanks, Chris.
0001-MINOR-backend-srv_queue-helper.patch
Description: Binary data
0002-MINOR-backend-srv_is_up-converter.patch
Description: Binary data
> On 26 Sep 2025, at 05:03, Willy Tarreau <[email protected]> wrote: > > Hi Christopher, > > On Thu, Sep 25, 2025 at 04:04:59PM +0100, Christopher Staite wrote: >> There is currently an srv_queue converter which is capable of taking the >> output of a dynamic name and determining the queue length for a given >> server. In addition there is a sample fetcher for whether a server is >> currently up. This simply combines the two such that srv_is_up can be >> used as a converter too. > > Good idea! > >> To avoid code duplication, extract the server lookup into a function >> from the existing converter and re-use in both implementations. >> >> Future work might extend this to other sample fetchers for servers, but >> this is probably the most useful for acl routing. > > OK. I'd ask you 3 small things: > > - first, when modifying an existing feature to prepare a new one, it's > better to split that into two patches: one that splits the existing > function, and a new one that adds the new feature. This way, possible > regressions on the current functions are easier to detect and fix. > Typical regressions that happen when splitting functions are mostly > build-time issues that could produce warnings on older or newer > compilers for example (e.g. unused parameter in the new function). > And this way the new feature is easier to review, and you'll be happy > to see that your patch looks small, self-contained and cleaner. > > - mention in the srv_queue doc (then on srv_is_up) a small warning > like this one: "Before using this, please keep in mind that using > this converter on uncontrolled data might allow an external observer > to query the state of any server in the whole configuration, which > might possibly not be acceptable in some environments". > > - your patches were space-mangled by your mailer below, I suggest that > you simply attach them so that they remain intact. > > Otherwise at first glance this looks good to me. > > Thank you! > Willy
