Mailing-List: contact user-help@roller.apache.org; run by ezmlm
Precedence: bulk
Reply-To: user@roller.apache.org
Received-SPF: pass (athena.apache.org: local policy)
Content-Type: text/plain; charset=utf-8
Mime-Version: 1.0 (Mac OS X Mail 7.3 \(1878.6\))
Subject: Re: Salt stays same in entryEdit page - is it by design?
From: kyle <kyle@bridge9.sakura.ne.jp>
In-Reply-To: <96136DF5-16B7-440F-8367-0ADFC1A9E1EA@bridge9.sakura.ne.jp>
Date: Sun, 11 Jan 2015 00:36:40 +0900
Content-Transfer-Encoding: quoted-printable
Message-Id: <261C0394-1569-44B4-B4DC-CBE943433853@bridge9.sakura.ne.jp>
References: <80E8DC3B-B23A-4F33-BE4D-E783692F499A@bridge9.sakura.ne.jp>
 <54b127de.105f8c0a.0b38.0c54@mx.google.com>
 <96136DF5-16B7-440F-8367-0ADFC1A9E1EA@bridge9.sakura.ne.jp>
To: user@roller.apache.org

I think I found the cause. UIAction#setSalt(String) is being invoked =
with old salt value (maybe it comes from POST parameter). simply =
deleting UIAction#setSalt(String) solved the problem on my workspace. =
now I'm seeing a new salt value at every hitting of "Save as draft".

2015/01/10 22:49=E3=80=81kyle <kyle@bridge9.sakura.ne.jp> =
=E3=81=AE=E3=83=A1=E3=83=BC=E3=83=AB=EF=BC=9A

> Hello Glen,
>=20
> Thanks for the response.
>=20
> I think it appears that my previous email was not clear. in my =
environment, the clock won't restart when I hit save as draft. the clock =
start ticking at the first time I visit the editing page, it never stops =
so I can't stay in the editing page over 60 minutes. to restart the =
clock (to get a new salt), I need 2 steps - move to other page and get =
back to the editing page. do you mean I should doing so?
>=20
> 2015/01/10 22:22=E3=80=81Glen <glen.mazza@gmail.com> wrote:
>=20
>> Hi Kyle, while it probably could be handled more elegantly, your =
writing would not normally be lost, as you can just hit the back key =
from your browser to get back to your text. 60 minutes is pretty =
generous and the clock restarts everytime you hit save draft as you =
periodically should be doing anyway.  We haven't the resources to try to =
implement google doc style editing within blog entries, at any rate for =
long term editing you can use your favorite desktop editor and then just =
copy and paste the html into the roller blog edit field and then publish =
it.
>>=20
>> Yes this happens with other containers like Tomcat.
>>=20
>> Glen
>>=20
>> -----Original Message-----
>> From: "kyle" <kyle@bridge9.sakura.ne.jp>
>> Sent: =E2=80=8E1/=E2=80=8E10/=E2=80=8E2015 7:57 AM
>> To: "user@roller.apache.org" <user@roller.apache.org>
>> Subject: Salt stays same in entryEdit page - is it by design?
>>=20
>> Hello, I'm running Roller 5.1.1 on WildFly 8.2.0.Final.
>>=20
>> Recently I got into a problem which is getting status code 500
>> by "javax.servlet.ServletException: Security Violation."
>> at editing page 1 or 2 times per day.
>> it occurs when I push save as draft button.
>>=20
>> I'm looking into this problem now,
>> then I found that the salt value in the editing page stays same
>> while I'm writing content and pushing save button sometimes.
>> I mean this element in the form:
>> <input type=3D"hidden" name=3D"salt" value=3D"8l4wTyF0LBKuhJLJ5zvU" =
id=3D"entry_salt"/>
>>=20
>> I turned on DEBUG logs, then I can see the problem more clearly.
>>=20
>> DEBUG 2015-01-10 20:52:28,267 SaltCache:put - PUT =
8l4wTyF0LBKuhJLJ5zvU
>> DEBUG 2015-01-10 20:52:37,924 SaltCache:get - HIT =
8l4wTyF0LBKuhJLJ5zvU
>> DEBUG 2015-01-10 20:52:37,924 SaltCache:get - HIT =
8l4wTyF0LBKuhJLJ5zvU
>> ...snip...
>> DEBUG 2015-01-10 21:51:49,744 SaltCache:get - HIT =
8l4wTyF0LBKuhJLJ5zvU
>> DEBUG 2015-01-10 21:53:14,173 ExpiringLRUCacheImpl:get - EXPIRED =
[8l4wTyF0LBKuhJLJ5zvU]
>> DEBUG 2015-01-10 21:53:14,173 SaltCache:get - MISS =
8l4wTyF0LBKuhJLJ5zvU
>>=20
>> And, after some investigating I found that
>> all salts will be expired in 60 minutes after creation.
>> thus, I can't stay at editing page over 60 minutes.
>> If I overstayed, then I can't save my draft anymore
>> and got "Security Violation".
>>=20
>> Is it expected behavior?
>> or it might be a WildFly specific problem?
>>=20
>> Also I posted about the problem:
>> =
http://www.nailedtothex.org/roller/kyle/entry/debugging-roller-s-salt-proc=
essing
>>=20
>> Thanks.
>=20
>=20