I did not put any distributable elements in web.xml files. So people still needs to follow the advice in catalina-container element in ofbiz-containers.xml
git-svn-id: https://svn.apache.org/repos/asf/ofbiz/trunk@1334336 13f79535-47bb-0310-9956-ffa450edef68
cases, this will always be configured, so might as well do it early.
This really should be final, but that change will come later.
git-svn-id: https://svn.apache.org/repos/asf/ofbiz/trunk@1334256 13f79535-47bb-0310-9956-ffa450edef68
base-64 crypted({hashType}$text), key values are base64 encoded,
encrypted row values are $salt$encoded. Note, however, that this code
is not yet active, as it hasn't been added to the list of key handlers.
This new handler is *not* active yet.
git-svn-id: https://svn.apache.org/repos/asf/ofbiz/trunk@1334255 13f79535-47bb-0310-9956-ffa450edef68
classes and concrete instances. This allows for adding new types of key
management blocks, so that enhanced features can be added. No
functional changes have actually been implemented yet; all existing
systems should behave *exactly* the same.
git-svn-id: https://svn.apache.org/repos/asf/ofbiz/trunk@1334254 13f79535-47bb-0310-9956-ffa450edef68
longer a need to pre-populate the EntityKeyStore table with 20 random
keys; in reality, these 20 keys were enought to expand the keyMap
HashMap's internal structure, to a size that wouldn't normally be
auto-expanded during production. And it was this lack of auto-expanding
that was keeping the map from becoming corrupted in very rare
circumstances.
git-svn-id: https://svn.apache.org/repos/asf/ofbiz/trunk@1334252 13f79535-47bb-0310-9956-ffa450edef68
rest of the code. The key stored in the database is not only just for
decrypting or encrypting, it's just a key.
git-svn-id: https://svn.apache.org/repos/asf/ofbiz/trunk@1334214 13f79535-47bb-0310-9956-ffa450edef68
into two separate methods; a find for an old-style hashed key will no
longer create the old-style key. This commit does not fix the existing
issue when multiplle threads try to encrypt a value with a key that
hasn't been created or loaded from the database yet; that is still to
come.
git-svn-id: https://svn.apache.org/repos/asf/ofbiz/trunk@1334212 13f79535-47bb-0310-9956-ffa450edef68
idea where those bad digest password values came from. Really, I have
no clue. Everything worked in testing. I'm very confused. I have
verified that I can log in now with the demo accounts. Sorry about the
breakage.
git-svn-id: https://svn.apache.org/repos/asf/ofbiz/trunk@1334173 13f79535-47bb-0310-9956-ffa450edef68
now follow the same form: (String hashType, ..., String valueToHash).
Some methods also follow a polish notation suffix.
git-svn-id: https://svn.apache.org/repos/asf/ofbiz/trunk@1334172 13f79535-47bb-0310-9956-ffa450edef68
it would call String.getBytes(), without specifying an encoding. That
meant that the platform default encoding would be used, which might
change between different installs(if someone changed an environment
variable, for instance). Now, there are explicit variants to be called.
When creating a new value, use cryptValueUTF8. If, for some reason,
calling code is doing it's own comparison, it must continue to use
cryptValueBytes(password.getBytes()).
git-svn-id: https://svn.apache.org/repos/asf/ofbiz/trunk@1334170 13f79535-47bb-0310-9956-ffa450edef68
This is a first pass - there are more things that need to be fixed, but I didn't want to rewrite the whole thing in one commit.
git-svn-id: https://svn.apache.org/repos/asf/ofbiz/trunk@1333960 13f79535-47bb-0310-9956-ffa450edef68
go thru a deprecation cycle for this, as this method was only recently
added, and has never been officially released.
git-svn-id: https://svn.apache.org/repos/asf/ofbiz/trunk@1333664 13f79535-47bb-0310-9956-ffa450edef68
String.getBytes(Charset) to fetch the bytes. Also switches the one case
of getBytes("UTF-8") to the Charset version.
git-svn-id: https://svn.apache.org/repos/asf/ofbiz/trunk@1333654 13f79535-47bb-0310-9956-ffa450edef68
upstream. However, I decided that it might be generally useful, so
committed it anyways. I forgot that I needed to add the copyright
header to it. I suppose I should add a pre-commit hook somehow either
in svn or git.
git-svn-id: https://svn.apache.org/repos/asf/ofbiz/trunk@1333619 13f79535-47bb-0310-9956-ffa450edef68
This is not a complete solution and induces more problem than it solves (notably in RuntimeData for jobs), to be continued...
git-svn-id: https://svn.apache.org/repos/asf/ofbiz/trunk@1331996 13f79535-47bb-0310-9956-ffa450edef68
This was due to MacroScreenRenderer.renderScreenletSubWidget() calling HtmlFormRenderer instead of MacroFormRenderer. So there was still a call to the old, now inexistent, call_cal js function.
So this not only fixes this case but also all other similar cases, when a navigation-form-name screenlet attribute was used and a date field used in the related form.
I have also checked there are no other uses of HtmlFormRenderer in MacroScreenRenderer. So I think w are safe from other side effects...
git-svn-id: https://svn.apache.org/repos/asf/ofbiz/trunk@1331898 13f79535-47bb-0310-9956-ffa450edef68