Signed-off-by: Rahix <rahix@rahix.de>

Signed-off-by: Rahix <rahix@rahix.de>

Use RTC_GetTime() instead of RTC_GetSecond() because otherwise a read
immediately following the RTC alarm will return the previous value.
RTC_GetTime() will return E_BUSY a number of times before finally being
successful and returning a timestamp.  To keep the system load minimal
during this waiting, the task is put to sleep for 4ms everytime the read
fails.  This value might need further adjustment.

Signed-off-by: Rahix <rahix@rahix.de>

Signed-off-by: Rahix <rahix@rahix.de>

Signed-off-by: Rahix <rahix@rahix.de>

Signed-off-by: Rahix <rahix@rahix.de>

Signed-off-by: Rahix <rahix@rahix.de>

Signed-off-by: Rahix <rahix@rahix.de>

Signed-off-by: Rahix <rahix@rahix.de>

Signed-off-by: Rahix <rahix@rahix.de>

Signed-off-by: Rahix <rahix@rahix.de>

Signed-off-by: Rahix <rahix@rahix.de>

Signed-off-by: Rahix <rahix@rahix.de>

Signed-off-by: Rahix <rahix@rahix.de>

ld emits relocations for weak undefined symbols. We want to safely skip
them.

This lets us decide where the stack live, and place it at the best
possible spot for it to grow towards the image.

This is the first step towards having a heap in l0dables.

Signed-off-by: Rahix <rahix@rahix.de>

We want to run a linter on CI submission.

We also want to display nice error messages to the user. Thus,
annoyatron is born.

The flow is as follows:
 - a MR is submitted and triggers the 'lint' pipeline
 - annoyatron gets pinged over https that an MR requires attention
 - annoyatron starts running and waits until the MR's pipeline run
   succeeds or fails
 - the lint jobs either succeeds or fails
 - annoyatron notices the pipeline passed, and inspects its results to
   see if the lint pipeline failed or not.
 - annoyatron posts a comment, if necessary