Closed mayaCostantini closed 2 years ago
/assign @harshad16
[APPROVALNOTIFIER] This PR is NOT APPROVED
This pull-request has been approved by: VannTen
Once this PR has been reviewed and has the lgtm label, please ask for approval from harshad16 by writing /assign @harshad16
in a comment. For more information see:The Kubernetes Code Review Process.
The full list of commands accepted by this bot can be found here.
From the look of the license-solver the license bit should be dict https://github.com/thoth-station/license-solver/blob/28c29dbf560c7165fd67419a3abfe3b53a1c6f1a/thoth/license_solver/package.py#L57 there was issue in default value sent in case of uncatched license. There is already a pr for fix. https://github.com/thoth-station/license-solver/pull/67
I think that would fix the issue, and we would have to catch this in storages.
On Thu, Sep 15, 2022 at 06:00:33AM -0700, Harshad Reddy Nalla wrote:
From the look of the license-solver the license bit should be dict https://github.com/thoth-station/license-solver/blob/28c29dbf560c7165fd67419a3abfe3b53a1c6f1a/thoth/license_solver/package.py#L57 there was issue in default value sent in case of uncatched license. There is already a pr for fix. https://github.com/thoth-station/license-solver/pull/67
Hum. Reading that code, maybe license
should be an Optional[Dict[str, str]]
and replace UNDETECTED with None ?
Unless something else can be there in place of UNDETECTED ?
Hum. Reading that code, maybe
license
should be anOptional[Dict[str, str]]
and replace UNDETECTED with None ? Unless something else can be there in place of UNDETECTED ?
maybe it , seems like the code of license solver needs more of a closer look. As of now, we can follow the pattern of existing one i feel and explore more going forward.
in any case, i feel we might not have to do the type checking ? or do you feel it just better to check as there is no much effect on runtime?
On Thu, Sep 15, 2022 at 08:11:24AM -0700, Harshad Reddy Nalla wrote:
in any case, i feel we might not have to do the type checking ? or do you feel it just better to check as there is no much effect on runtime?
I'm not the person to ask ^^. I'm totally sold on strong static type-checking, so I'll always feels more type-correctness is good to have. That's a time allocation problem, though (for example the whole graph module in storages is a bit much) ; so if we solve the problem at hand, that could probably wait.
no much effect on runtime
Type hints are discarded at runtime (at parsing time I guess) so makes that virtually no effect.
Don't know why github keeps dropping my email replies :/
in any case, i feel we might not have to do the type checking ? or do you feel it just better to check as there is no much effect on runtime?
I'm not the person to ask ^^. I'm totally sold on strong static type-checking, so I'll always feels more type-correctness is good to have. That's a time allocation problem, though (for example the whole graph module in storages is a bit much) ; so if we solve the problem at hand, that could probably wait.
no much effect on runtime
Type hints are discarded at runtime (at parsing time I guess) so makes that virtually no effect.
This might need more static type-checking. for now, the already include changes are keeping this in place. we should work on license solver to get the default value and then include the type-checking. Closing this for now. lets address this in future, if we face same issue. thanks for all the discussions and work.
Related Issues and Dependencies
Fixes https://github.com/thoth-station/storages/issues/2703
This introduces a breaking change
This should yield a new module release
This Pull Request implements
Handle the case when the license metadata is undetected during document sync job.