Test failures due to hash randomisation in perl 5.17.6 [rt.cpan.org #81381]

[oalders]

Migrated from rt.cpan.org#81381 (status was 'open')

Requestors:

• 'spro^^%^6ut#@&$%*c

From $_ = 'spro^^%^6ut#@&$�>#!^!#&!pan.org'; y/a-z.@//cd; print on 2012-11-23 16:38:12:

From yves@cpan.org on 2012-11-27 21:35:32:

On Fri Nov 23 11:38:12 2012, SPROUT wrote:
> Perl 5.17.6 randomises hash seeds, exposing an existing bug in
> t/base/ua.t:
> 
> #    Expected: "http://example.com"
> #  base/ua.t line 101 is: ok($ua->proxy('http'),
> "http://example.com");
> # Test 34 got: "http://10.10.10.2:8080" (base/ua.t at line 105)
> #    Expected: "http://example.com"
> #  base/ua.t line 105 is: ok($ua->proxy('http'),
> "http://example.com");
> 
> I happen to have HTTP_PROXY set to http://10.10.10.2:8080 and
> http_proxy unset.
> 
> The test script sets http_proxy.  Which of HTTP_PROXY and http_proxy
> is picked up by sub
> LWP::UserAgent::env_proxy is going to be random, so the test will fail
> half the time.

I just pushed:

https://github.com/demerphq/libwww-
perl/commit/96791d2f72dc20ab6ff3c6ce6442a8db509bcfcc

and submitted a pull request:

https://github.com/gisle/libwww-perl/pull/3

to have it merged to 

https://github.com/gisle/libwww-perl

but i notice that itself is a fork of 

https://github.com/libwww-perl/libwww-perl

which has a lot of pull requests of its own:

https://github.com/libwww-perl/libwww-perl/pulls

So I am unclear if i pushed my request to the right place.

Anyway this is a major blocker for 5.18.

https://github.com/gisle/libwww-perl/pull/3

From $_ = 'spro^^%^6ut#@&$�>#!^!#&!pan.org'; y/a-z.@//cd; print on 2012-11-27 21:51:41:

[genio]

@demerphq I don't think this issue still exists. The PR you submitted was for the old repository and should likely be closed out. I compared your changes against what's currently in LWP/UserAgent.pm and don't seem to see the same problems.

Can you test now and let us know if the problem still exists?

[demerphq]

If you guys have deterministic behaviour under hash randomisation then this ticket can be closec

…

On 1 Apr 2017 1:30 a.m., "Chase Whitener" ***@***.***> wrote: @demerphq <https://github.com/demerphq> I don't think this issue still exists. The PR you submitted <gisle#3> was for the old repository and should likely be closed out. I compared your changes against what's currently in LWP/UserAgent.pm and don't seem to see the same problems. Can you test now and let us know if the problem still exists? — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#230 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AAGex3xQeddIRYZfcl4ZeHBflG5Wo1gfks5rrYx9gaJpZM4MwCn_> .

[genio]

@demerphq Thanks. I'll close out it out for now. Please don't hesitate to yell at us if it turns out I'm wrong.