Sean Wilcox writes:
> The following bugs fixed here are :
>
> 6302173 startd dies after interaction with inetd / inetd service
> 6764671 Array overrun in librestart
> 6221667 method context errors should be logged in the service's instance log
> 6348085 Spurious "Ignoring duplicate environment variable" message
> 6411391 Empty method_context's don't work
> 5093847 importing a manifest should create method_context properties
> even if not set
> 6517270 partial method context definitions should work
>
> The full webrev is located at :
>
> http://cr.opensolaris.org/~swilcox/full_webrev/
>
> The bfu only webrev is located at :
>
> http://cr.opensolaris.org/~swilcox/bfu_webrev/
Looks good. Just a few comments:
usr/src/cmd/svc/startd/env.c:
209 This problem is not of your making, but I did notice if
asserts are compiled out of this code, that the function
will break.
218 I can't wrap my head around the reason for p != env in this
line. Can you explain it to me?
usr/src/cmd/svc/startd/protocol.c:
341-346 Since startd is multithreaded, is it possible that memory
allocation could fail for delegate_channel_name and succeed
for master_channel_name?
389 Should there be some context in this error message?
usr/src/lib/librestart/common/librestart.c:
164-178 Residual XXXs. Comment is written in the context of fixing
a bug.
tom
