Closed Prinzhorn closed 2 years ago
Why do you create a PR for each of these? I mean it's okay to have distinctive PRs but this is all about the gridfs documentation. I think this could all be handled in one PR having several commits for each changes. Makes reviewing also much easier.
cc @dr-dimitru
Because I'm following along the docs and open a PR when I run into an issue (using the GitHub UI). That is much easier on my end than to first keep a list of them and then go back to create a single PR and edit my fork and then push the changes. I though it would be a lot easier for you to hit merge five times than for me to keep updating a PR. Blame GitHub for having a limited "edit" button :smile:
Also I wasn't expecting to make multiple changes when I started and thought it would would remain a single PR.
I though it would be a lot easier for you to hit merge five times than for me to keep updating a PR
That's not how a review works :-) So I have to make 5 reviews now and in the end one more final review to see if all the changed code examples work, when applying to a real app, for example like this one: https://github.com/veliovgroup/files-gridfs-autoform-example
@Prinzhorn let me see if I can put them into one PR, git is fortunately not limited to what GitHub offers
Thanks!
Two of them are just text, nothing to test. One is literally syntax error (semicolon in an object literal). The two forgotten export
make copy paste harder. And createAfterUpdate
is a typo because the code that uses it has import { createOnAfterUpload } from 'path/to/createOnAfterUpload'
and onAfterUpload: createOnAfterUpload(imageBucket),
. They all currently prevent the code form working at all, so I'd be surprised if the example uses it.
I'm sorry for just throwing these PRs at you but I wanted to keep my flow and just spend the five seconds to open a PR via the UI.
I did get GridFS fully working now with these changes. (plus https://github.com/veliovgroup/Meteor-Files/issues/818)
I will add code for gird-fs straming with GridFSBucket, since it is slightly different. However I first need to make this running in our project at work.
Correct me if I'm wrong, but this is not accurate. The files are streamed to the client using
bucket.openDownloadStream
, see https://github.com/veliovgroup/Meteor-Files/blob/master/docs/gridfs-bucket-integration.md#4-create-download-handlerThat means the entire file does not need to be held in server memory. Correct?