Public bug reported:
The language L10n translated language fields that are marked
__like_this__
Under certain conditions where database output happens to contain double
underscores in certain contexts, the parser will get confused, and end
up not replacing the translated terms.
It's quite possible for some site database output to accidentally match
the rule '/\_\_(.*)\_\_/U'
In fact it's happening on Open Font Library!
To reproduce this error:
* Create a fresh aiki site
* Create a new user in the database: INSERT INTO aiki_users (username,
full_name) values ('example__', 'example__')
* Create a new widget
** In the SQL query: select * from aiki_users
** In the content: <p><a href="((full_name))">((username))</a></p>
The important output will be:
<p><a href="example__">example__</a></p>
The effect is that L10n() seems to have trouble parsing this page.
If you view source on this page you can see that other __aiki_terms__
are not getting replaced:
<meta http-equiv="Content-Type" content="text/html;
charset=__encoding__" />
** Affects: aikiframework
Importance: Critical
Status: New
** Affects: openfontlibrary
Importance: Critical
Assignee: Christopher Adams (christopheradams)
Status: Confirmed
** Changed in: aikiframework
Importance: Undecided => Critical
** Changed in: aikiframework
Milestone: None => 0.9
** Also affects: openfontlibrary
Importance: Undecided
Status: New
** Changed in: openfontlibrary
Status: New => Confirmed
** Changed in: openfontlibrary
Importance: Undecided => Critical
** Changed in: openfontlibrary
Assignee: (unassigned) => Christopher Adams (christopheradams)
** Changed in: openfontlibrary
Milestone: None => 0.3
--
You received this bug notification because you are a member of Aiki
Framework Admins, which is subscribed to aikiframework.
https://bugs.launchpad.net/bugs/794093
Title:
languages L10n is brittle on database fields containing __
Status in Aiki Framework:
New
Status in Open Font Library:
Confirmed
Bug description:
The language L10n translated language fields that are marked
__like_this__
Under certain conditions where database output happens to contain
double underscores in certain contexts, the parser will get confused,
and end up not replacing the translated terms.
It's quite possible for some site database output to accidentally
match the rule '/\_\_(.*)\_\_/U'
In fact it's happening on Open Font Library!
To reproduce this error:
* Create a fresh aiki site
* Create a new user in the database: INSERT INTO aiki_users (username,
full_name) values ('example__', 'example__')
* Create a new widget
** In the SQL query: select * from aiki_users
** In the content: <p><a href="((full_name))">((username))</a></p>
The important output will be:
<p><a href="example__">example__</a></p>
The effect is that L10n() seems to have trouble parsing this page.
If you view source on this page you can see that other __aiki_terms__
are not getting replaced:
<meta http-equiv="Content-Type" content="text/html;
charset=__encoding__" />
_______________________________________________
Mailing list: https://launchpad.net/~aikiframework.admins
Post to : aikiframework.admins@lists.launchpad.net
Unsubscribe : https://launchpad.net/~aikiframework.admins
More help : https://help.launchpad.net/ListHelp
