You are not logged in.
Addendum: since there is no obvious reason why the filter function would crash (though it might be, or incorrectly returning false), I'd also like to know:
There is a custom publish hook for this module. Does it need to clear cache explicitly?
Tim, I will take a look and let you know. Thanks!
doon.mok, did not mean to ignore you. Sorry about that, and thanks for the suggestion. Truncating worked, but only once.
Tushar
Hello,
I recently changed the filter function of a module view, and now the view seems to be stuck in a single state. I.e., publishing new items doesn't change the view.
(The new published items do pass the filter; they should show up.)
I am trying to debug this and wonder if there is some obvious error in the filter function?
If I change the filter function to something else and change it back, the view is refreshed and shows exactly the expected items. But publishing an item never updates the view. Fwiw there is a publish hook as well, but it has not changed for a long time and hasn't been a problem.
Any pointers?
Best,
Tushar
We recently had a problem with page revisions, which for now has been solved by simply creating a fresh revisions table.
This brought up a new problem, which I am mentioning in case it is a considered a bug:
It happens when you save the current revision of a page whose last_edited_by is a user who has been deleted. The revision is saved, but does not display on the admin screen.
A reasonable compromise in this case could be to outer join on the users table, and show a missing author as "unknown" or "former account" or similar. I think having all revisions available is more important than showing accurate authorship. Can this be done?
Thanks,
Tushar
Hi, Tim,
I wasn't quickly able to get a fresh install up to compare. But it seems from the code that at least one column needs a new name?
currently we have: bigtree_page_revisions.resource_allocation
new: bigtree_page_revisions.resources
I am not sure if this is the only change. Would like to apply a migration query if there is one specifically for this table. Please let me know what's a good way to do this.
Thanks very much,
Tushar
Tim, thanks, this is fantastic news!
I now see from our DB that previous admins had saved revisions of certain pages. Unfortunately only one Published Revision ever shows up on the revisions tab of any page. No saved revisions. Trying to save a new revision from current has no effect as far as we can see. Is there a step I am missing?
This is Bigtree 4.4.13.
Thanks again,
Tushar
I feel I may be missing something fundamental. But here goes:
What's a good way to achieve "versioning" of Bigtree pages? Our user base is expanding and they seem to mean at least a couple of different things by this:
1. Modify page content for a limited time, then roll it back to a previous version.
2. Switch between multiple stored versions of content.
Any procedures that use standard admin features are of course preferable. Don't want to (badly) customize admin if I can avoid it.
Thanks!
Tushar
Hello,
After a long time I got another chance to try enabling cloudfront, and this time I was lucky.
The distribution is now set in Bigtree. But none of the URLs produced by Bigtree have changed. I expected all S3 URLs to change automatically, but is there a catchup to be done?
Bigtree went through a long cycle of recaching when I set the Cloudfront distribution. I don't understand what it did, but maybe it's related.
The Cloudfront URLs themselves are loading fine.
Best,
Tushar
Composer update is failing with out of memory on our test server, unfortunately. AWS SDK is at 3.52.30. If that helps.
There's CA certs in the cache directory.
Best,
Tushar
I just tried 4.4.14. Bigtree went through an upgrade sequence for file metadata, which may be unrelated.
After that, the cloud storage behavior has not changed. All I can configure is AWS Region, Access Key ID, and Secret Access Key, as before.
Tushar
(Continuing a previous conversation here)
I am trying to set up Cloudfront, already have S3. Bigtree is on 4.4.12.
I thought maybe resaving the S3 configuration in the cloud storage tab would prompt me for Cloudfront configuration. I was wrong. What am I missing?
I have done nothing in AWS's own admin, by the way.
Thanks,
Tushar
Thanks! For now, an ad hoc query to blow away empty redirect_urls will do for us. Absolute broken_urls as well.
Our bigtree_404s table is getting quite bulky. I am seeing entries in there that don't immediately make sense, and am thinking of cleaning them up. Questions:
1. Does an absolute URL as broken_url have any effect at all? There's a lot of them in there, somehow. Not sure how they got in.
2. Does an external link in redirect_url serve any purpose? Does bigtree redirect to external sites?
3. If redirect_url is empty, would deleting that row break any functionality? (Assuming we are not interested in the number of requests to the broken URL.)
Thanks,
Tushar
Sometime in 2018, the implementation of replaceRelativeRoots() changed from str_replace to strtr. We upgraded to a post-2018 bigtree version very recently.
It doesn't play well with the version of Form Builder that we have, which sanitizes form submissions with replaceRelativeRoots. (It used to recursively sanitize before, now it warns and returns null or an empty string.)
Due to some urgency over here, I am going to remove replaceRelativeRoots calls in Form Builder. But I was wondering if there is an officially blessed fix for this.
Thanks,
Tushar
Hi, Tim,
Any word on this? It seems the admin backend and frontend might have different ideas about what the file browser ajax endpoint should be, admin/ajax/files vs admin/ajax/file-browser . Wondering if there is an upgrade step I have missed?
Tushar
Following up to the upgrade saga:
The upgrade took about 95 minutes and went well. No major problems to the front end that we can see so far. Some questions:
1. There is a files tab in admin, new to us. It looks fairly intuitive; however, the "File Browser" that we have and use for attaching images to tours, buildings etc is now empty. How do we reuse our S3 images?
2. Google Analytics: we want to disable this in admin. How to do that?
3. Should custom/json-db/, vendor/, and cron-run.php be put in .gitignore?
4. There are various minor warnings in admin. An example:
Warning: strtr() expects parameter 1 to be string, array given in /var/www/caf/bigtree/core/inc/bigtree/cms.php on line 1754
Is there some obvious thing I have overlooked that is causing this?
We have no set date for upgrading, so no rush.
Thanks,
Tushar
This is fantastic. Thanks again Tim!
Tushar
Tim, thank you! I am going to start fresh and see how it goes. I have two questions:
1. Will a 4.2-compatible database and 4.4 BigTree trigger this upgrade, just by loading the developer tab?
2. Does the upgrade modify or delete cloud images? Almost all our images, on dev or prod, are served from the same S3 bucket. (File manager being unusable for a while is okay, but user-facing images need to keep serving).
Thanks again!
Tushar
I am trying to upgrade from 4.2 to 4.4 and have a few immediate problems: e.g. custom classes not found, routed templates searched for in the wrong place (basic) etc.
I did the "upgrade" with git pull in the bigtree directory, and now I wonder if that's the right way. Especially since the admin developer tab began to go through an upgrade process a little later (which I stopped).
So, since it's early for specific questions:
1. What's the recommended way to go from php5.6 -> 7.2 AND bigtree 4.2 -> 4.4 ?
2. Are there some changes to custom code that are a must? (We have a custom router among other things.)
3. Can the process of converting file browser data be speeded up or done from cli.
4. (This is documented but unfortunately can't find it) What PHP libraries are required? E.g. curl.
Thank you,
Tushar
Sounds good, Tim. I will see if I can extend the class independently of the service provider.
To clarify, we want this for ticket purchase confirmations. It's proven useful to attach a PDF ticket to the message itself. (We use a different provider for newsletters etc, which doesn't involve BigTree.)
Thanks!
Tushar
We will be needing an email service provider in the near future, and it will probably be one supported by BigTree.
BigTreeEmailService provides one method, sendEmail, and it's not obvious from the docs how to send a message with an attachment.
Any ideas? A typical email in our situation is HTML with multiple PDF attachments.
Thanks!
Tushar
That's the best news ever! Thanks for keeping things, as they say, loosely coupled. Tushar
Hello,
To recall, I am trying to "unveil" a large number of content changes (to pages and tree structure).
At this point, dump-from-stage/undump-to-live seems to be the only option.
But there are other parts of the live DB that really can't be overwritten (e.g. calendar). So I am trying to get a minimum list of tables that need to be synced.
There are four page operations being performed on stage:
1. Create new
2. Edit
3. Move to another parent
4. Delete
So far, I believe these have effects on: bigtree_pages, bigtree_route_history, bigtree_audit_trail.
So I think that dump/undump on these three will do the trick. Is there anything I am missing? Thanks! Tushar
Clear and simple. Thanks, Tim!
I'm back.
I might have to reorganize our existing set of pages—which amounts to scripting a large number of "move page" operations.
Is it correct to say that a single updatePageParent() call will accomplish a move-page? Also, I assume this method takes page ids?
Thanks,
Tushar