Build marked as failed due to VARCHAR(100) limit on license field(?)

[justsmth]

Crate name

aws-lc-sys

Build failure link

https://docs.rs/crate/aws-lc-sys/0.40.0/builds/3143766

Additional details

The aws-lc-sys crate has been failing to build on docs.rs since version =0.39.0. The build logs show the documentation building successfully, but the build is marked as failed with a database error.

The build log for aws-lc-sys 0.40.0 (build #3143766) shows:

# error kind
Other

# pre-build errors
error returned from database: value too long for type character varying(100)

I believe the cause is our license field in Cargo.toml, which is a 135-character SPDX expression:

license = "ISC AND (Apache-2.0 OR ISC) AND Apache-2.0 AND MIT AND BSD-3-Clause AND (Apache-2.0 OR ISC OR MIT) AND (Apache-2.0 OR ISC OR MIT-0)"

This exceeded 100 characters starting in version 0.39.0 when we updated our licensing metadata. The last successful build was 0.38.0, which had a 43-character license string.

I'm not aware of SPDX specification imposing a 100-character limit on license expressions, so this seems like an unintentional constraint in the database schema.

[syphar]

created the fix in #3302

when that's reviewed & deployed, we can re-run the builds.

[syphar]

I wonder if / how crate authors would want to be actively informed about these build failures.

Once we wanted to do this via the docs.rs interface (#2471), I could also imagine emails to the owners, if that's possible, not sure.

[justsmth]

I wonder if / how crate authors would want to be actively informed about these build failures.

Once we wanted to do this via the docs.rs interface (#2471), I could also imagine emails to the owners, if that's possible, not sure.

I wonder whether one could automate the creation of an issue in the associated repository when such a failure occurs? This should be straightforward with GitHub-hosted crates.

[syphar]

Not sure, this might be too noisy. Example: we create the issue for the error, you try again for the same problem, still fails. (we would create two issues)

[syphar]

@justsmth fyi, this not deployed yet, I have to see if I can get #3296 reviewed / merged soon so that would be included.

Generally:
Since the queue is so long right now, can re-run some builds with higher prio when this is deployed.

You can give me a list of crates+release, then I'll do that after deploying

[justsmth]

Not sure, this might be too noisy. Example: we create the issue for the error, you try again for the same problem, still fails. (we would create two issues)

Right, to avoid duplicates you would need to track for a given crate release whether or not an issue had been generated. I assume this could be added into an existing database schema?

[justsmth]

@justsmth fyi, this not deployed yet, I have to see if I can get #3296 reviewed / merged soon so that would be included.

Generally:
Since the queue is so long right now, can re-run some builds with higher prio when this is deployed.

You can give me a list of crates+release, then I'll do that after deploying

Thanks! I think the docs just need to be regenerated for the three most recent releases of aws-lc-sys: 0.39.0, 0.39.1, and 0.40.0.

[syphar]

all requeued