The simple cache plugin uses t() for the '#type' of granularity
So by default no granularity will be saved so panels with arguments has a wrong caching by default so this panels display white sometimes when they should not :(
| Comment | File | Size | Author |
|---|---|---|---|
| #4 | panels_fix_css_cleanstring.patch | 1.72 KB | jcmarco |
| #1 | 920266-panels-simple_cache-t.patch | 682 bytes | dawehner |
Comments
Comment #1
dawehnerHere is the patch. What a facepalm
Comment #2
tobiasbComment #3
jcmarco commentedIn the last commit: http://drupal.org/cvs?commit=437926
There is a big problem with the new cleanstring in the css field.
Now the pane custom css values are all linked instead respecting the configured ones, due to the new:
$content->css_id = ctools_cleanstring($pane->css['css_id']);
I mean, if I have set up the CSS class in the CSS properties of a pane as:
default_title dark title-F64B00 view-top-hot inner-lightblue-white
Now I get this in the html code:
default-title-dark-title-F64B00-view-top-hot-inner-lightblue-white
Comment #4
jcmarco commentedReading the documentation in ctools/includes/cleanstring.inc
I think that there is no way to have a field with custom css values and apply it the cleanstring,
because by default it changes spaces and untranslatable characters to '-' or whatever else:
* - 'separator': Change spaces and untranslatable characters to
* this character. Defaults to '-'.
So I would recommend to roll back part of the changes in this commit.
Comment #5
merlinofchaos commentedUm? jcmarco, you seem to have taken over an unrelated issue.
Comment #6
merlinofchaos commentedFormally: Ignore #3 and #4. Those belong in a new issue, not taking over a completely unrelated issue.
jcmarco: I assume this is just an error. The only reason you don't get an ear-full is because you've been around before and have earned some karma. =)
Comment #7
merlinofchaos commentedOh I see now. I was confused, because I had forgotten that I had committed this patch, and that cleanstring patch went in as part of this accidentally.
Comment #8
merlinofchaos commentedOk, the cleanstring patch wasn't meant to be committed, so removed. That allows this to be marked fixed.
Comment #9
jcmarco commentedI realized that the issue subject was not related with the new issue, but the problem was created by a wrong commit on this issue. So I considered the best to attach it to the problem root, to ease the analysis.
Sorry for any inconvenience.