add no macros example, remove bogus print, fix crash on empty schedule #113
Conversation
All our examples use the macros, but we should also demonstrate the ugly truth behind those macros.
Add some comments, shorten imports a bit
This fixes a crash on the hypervisor, if no partition is scheduled. Instead, in each partition window a warning is emitted, pointing out the empty schedule for that partition window.
| a653rs = { workspace = true } | ||
| a653rs-postcard = { git = "https://github.com/DLR-FT/a653rs-postcard.git", branch = "main" } | ||
| a653rs-linux = { path = "../../partition" } | ||
| once_cell.workspace = true | ||
| serde = "1.0" | ||
| log = "0" | ||
| humantime = "2.1" |
There was a problem hiding this comment.
This should probably be workspace dependencies.
There was a problem hiding this comment.
Yeah ok for the example it is probably easier to copy-paste. But then again it should all not be workspace dependencies. Pick one?
There was a problem hiding this comment.
I agree. I just copied what was already there. I'm in favor of adjusting this, however, I think we should prepare an extra PR for that.
| //! - What is [`StartContext<Hypervisor>`]? | ||
| //! - Some parts of the ARINC 653 API are only available on a specific mode. | ||
| //! In particular, the process and port creationg API can only be called | ||
| //! during {`COLD`,`WARM`}`_START` mode (see [`OperatingMode`]). | ||
| //! `StartContext` is a struct that exposes this API only during | ||
| //! {`COLD`,`WARM`}`_START` mode, and can not be constructed by the user. | ||
| //! This makes misuse of the API (as in attempting to create a port/process | ||
| //! during `NORMAL`) impossible. |
There was a problem hiding this comment.
This should be part of the crate-level documentation and link to it.
There was a problem hiding this comment.
Yes, actually, this should be part of the a653rs documentation.
| anyhow!("At least one periodic or aperiodic process is expected to exist"), | ||
| )); | ||
| let part_name = self.partition.base.name(); | ||
| warn!("partition {part_name}: no process is scheduled") |
There was a problem hiding this comment.
Rather we should check at the end of cold_start if there is no process. If the partition is run multiple times a second, this produces too much log-spam on a relatively high log-level.
There was a problem hiding this comment.
This then would be a check inside the partition, AFAICT. I'm not an expert on this however.
There was a problem hiding this comment.
I thought further about this. I think it is completely fine to spam the log if there is a partition that at runtime creates an invalid configurration (no processes...). If anything, this should grab the attention of the developer, so spamming is justified. We could more clever logging (akin "20 identical messages dropped") but I don't see the justification for making this, considering that this only ever occurs when there is a low-key screw up by the developer to begin with.
No description provided.