On 22 January 2017 at 20:46, Steve Armstrong <[email protected]> wrote:
> Posted separate review so Camlistore, when getting a request for an app, > will strip off the app's prefix from the url before prepending the app's > host url: https://camlistore-review.googlesource.com/9207 > > Finally, what is the policy for updating someone else's review? I've got > mpl's scancab review (https://camlistore-review.googlesource.com/#/c/5416/) > rebased onto ToT and working (here on my Github fork https://github.com/ > stevearm/camlistore/tree/scancab). Should I post a separate review with > mpl's changes and my two small commits squashed into a single commit, or > should I be updating the existing review? > The main reason my CL isn't up to date is because the main interested party (Brad) does not have time to review it, and because it's not in our first priorities. Also because obviously I wouldn't merge it without review. That said, if it gets reviewed and tested enough by others, it is more acceptable for me to land it without waiting for Brad. So yes, if you want you can add a patchset on top of the existing CL, and we can move forward from there if there's enough interest. > Thanks > > On Sunday, January 22, 2017 at 11:12:23 AM UTC-8, Steve Armstrong wrote: >> >> I've posted the following to setup defaults for the keys I thought should >> be optional: https://camlistore-review.googlesource.com/#/c/9206/ >> >> On Sunday, January 22, 2017 at 9:32:50 AM UTC-8, Steve Armstrong wrote: >>> >>> I'm trying to get the scanningcabinet working ( >>> https://camlistore-review.googlesource.com/#/c/5416/). I've rebased it >>> up on ToT and it seems to be functional, but I'm having a lot of confusion >>> with the app interface with camlistore. >>> >>> I added the following to my config: >>> >>> "/scancab/": { >>> "handler": "app", >>> "enabled": true, >>> "handlerArgs": { >>> "prefix": "/scancab/", >>> "program": "scanningcabinet", >>> "listen": "127.0.0.1:2055", >>> "backendURL": "http://127.0.0.1:2055/"; >>> } >>> }, >>> >>> I'd like to drop as much redundant info as possible. Why is "prefix" >>> required twice? Can we plumb through the "/scancab/" key down into the >>> handler's constructor so it has it as a default? This key seems to be how >>> scannincabinet would call back to Camlistore. Shouldn't that not be >>> configurable and just be a randomly generated URL that's passed into >>> scanningcabinet when it's run (like the auth is)? >>> >>> Also, "listen" or "serverListed" are required. Can't we just default to >>> a random port added to Camlistore's ip if neither one is specified? >>> >>> "backendURL" or "serverBaseURL" is also required. Shouldn't it just >>> default to whatever "listen" defaults to? >>> >>> Finally, when I go to http://localhost:3179/scancab/test it routes to >>> http://localhost:2055/scancab/test. I'd expect it to trim off the >>> "/scancab/" prefix and route to http://localhost:2055/test. >>> >>> I'm probably missing the goal of how this contract is supposed to work, >>> and which knobs are intended to be configurable. Could someone help me out, >>> as I don't want to submit a pull request that removes/changes a bunch of >>> these keys and behaviours if they were intended. >>> >>> Thanks >>> >> -- > You received this message because you are subscribed to the Google Groups > "Camlistore" group. > To unsubscribe from this group and stop receiving emails from it, send an > email to [email protected]. > For more options, visit https://groups.google.com/d/optout. > -- You received this message because you are subscribed to the Google Groups "Camlistore" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. For more options, visit https://groups.google.com/d/optout.
