When a review already exists, fedora-create-review offers to create a new one anyway. As the review is named after the package and package names should be unique, I think this is the wrong thing to do in virtually all cases. Therefore I think it is a mistake to offer this option to the user interactively.
fedora-create-review
If there are special cases (whether in production or just for testing purposes) where creating another review is fine, it should IMO be done using a --force option that would suppress the check.
--force
In general, one should be required to first close old reviews and only then create a new one. But I'm not sure whether the current code distinguishes open and closed reviews.
In addition, when fedora-create-review fails due to such a check, the user should IMO be informed about the possible next steps.
An alternative way would be to default to --amend when the bug already exists and offer --no-amend for the special case of creating a duplicate review bug.
--amend
--no-amend
On the other hand forcing the user to confirm amending a bug (when using --amend) looks superfluous to me and especially the wording suggest that the user is doing something unexpected.
{{{ $ fedora-create-review --test --debug --amend --no-scratch-build *.spec netresolve-0.0.1-0.7.20151110git.fc24.src.rpm DEBUG:fedora-create-review:Summary: Generic name resolution library DEBUG:fedora-create-review:Description: Netresolve is a package for non-blocking network name resolution via backends intended as a replacement for name service switch based name resolution in glibc as well as a testbed for future glibc improvements. DEBUG:fedora-create-review:Name: netresolve Reviews for a package of the same name have been found: #1090499 NEW - Tomas Hozza - Review Request: netresolve - Generic name resolution library -
#1099435 CLOSED - Nobody's working on this, feel free to take it - Review Request: netresolve - Generic name resolution library - DUPLICATE
Do you want to proceed anyway? [Y/N] }}}
Well, f-c-r could certainly need an overhaul with respect to this usecase. Re-classifying as enhancement - patches welcome!
We need also a possibility to update (own) review reqests with new sources, after there's some feedback given from the reviewer.
@raphgro: Indeed, but from the command line? Why, then?
Metadata Update from @leamas: - Issue close_status updated to: None
@leamas Not sure what you mean with "but from the command line". I wish to have an option to automatically upload new sources and add a new comment to an existing review bug. f-c-r could check if there's already one for the given package name from command line.
Well, in any case: patches welcome! Personally, i have the ENOTIME error code...
Login to comment on this ticket.