linux: improve warnings when NSS/PR shutdowns fail

linux: improve warnings when NSS/PR shutdowns fail

Make one into a LOG(INFO) since it's relatively harmless;
add comments on the other with a link to the bug.

BUG=4609, 18410

[Evan Martin, 2009-08-04 00:45:43]

+willchan in case he has thoughts

[wtc, 2009-08-04 00:48:31]

"They always fail" in the description of this CL is
misleading.  It is possible to make NSS_Shutdown succeed.
It's just that I don't have time to do that.

Can we just LOG the message at the WARN or INFO level?

[Evan Martin, 2009-08-04 00:50:44]

On 2009/08/04 00:48:31, wtc wrote:
> "They always fail" in the description of this CL is
> misleading.  It is possible to make NSS_Shutdown succeed.
> It's just that I don't have time to do that.

Ah, I misunderstood our conversation from the bug, then.

> Can we just LOG the message at the WARN or INFO level?

Ok, will ping you again with a new patch in a sec

[wtc, 2009-08-04 17:21:20]

evan: I added two comments to Patch Set 1, but I got a
"500 Server Error" when I tried to publish them.  Now I
can't publish them again.  Please click inside Patch Set 1
to see my comments.  Sorry about the inconvenience.

[Evan Martin, 2009-08-04 17:29:52]

http://codereview.chromium.org/159836/diff/1/2
File base/nss_init.cc (right):

http://codereview.chromium.org/159836/diff/1/2#newcode137
Line 137: PR_Cleanup();
On 2009/08/04 17:18:56, wtc wrote:
> PR_Cleanup fails for a different reason (only if it's not called by the thread
> that initialized NSPR).  So we should
> log an ERROR message if PR_Cleanup fails.

This also always fails, which is why I was removing these.
I will file another bug on it.

[Evan Martin, 2009-08-04 17:34:42]

ok, new patch up

[wtc, 2009-08-04 17:42:19]

LGTM.  Thanks!

Could you use the shorter http://crbug.com/xxx URLs in comments?

http://codereview.chromium.org/159836/diff/1005/1006
File base/nss_init.cc (right):

http://codereview.chromium.org/159836/diff/1005/1006#newcode146
Line 146: // a potential race.
Nit: "a potential race" => "NSPR isn't initialized and cleaned up on the same
thread"