Show Posts

This section allows you to view all posts made by this member. Note that you can only see posts made in areas you currently have access to.

Messages - Arantor
1771
Archived fixes / Re: Previous/next topics don't work properly
« on January 25th, 2013, 03:38 PM »
Yeah, that's because we've implemented a lot of the fun stuff and bug fixing is the bit no-one likes doing.
1772
Archived fixes / Re: Input boxes not working in Chrome
« on January 25th, 2013, 03:12 PM »
WYSIWTF editor is as stable as ever then - http://wedge.org/pub/bugs/7830/reply-error-with-chrome/
1773
Bug reports / Re: Header headache
« on January 25th, 2013, 03:05 PM »
Quote
CSS files are being generated even when an Atom feed is being requested
That's... interesting. And unexpected.
Quote
add a generic stripos(we::$ua, 'bot') and enforce gzipping *within* CSS and JS only, for these?
Well, we could do that - we do still need to give them some CSS because some of them do do site caching (especially the likes of the Internet Archive, which doesn't have a 'bot' username)

Hmm, I don't know. I'm still a bit spaced out from the visit to the dentist...
1774
Bug reports / Re: Weaving & Warm issues in iOS
« on January 25th, 2013, 02:33 PM »
I'd be inclined to test for 'iPad' in the user agent and be done with it.
1775
Archived fixes / Re: Plugin execution
« on January 25th, 2013, 02:20 PM »
That is truly strange. But it sounds to me more like it's a bug of the plugin than of execution as such. Need to double check - but I wasn't able to get it to misbehave myself locally when I wrote it.
Posted: January 25th, 2013, 02:19 PM

Wait... actually... I can imagine one case where it might - if it ever has to play switcheroo on languages, which it can do under certain circumstances of merging. The plugin is a bit cheeky, it just fudges the language strings.
1776
Bug reports / Re: Minor bug with drafts and entities
« on January 25th, 2013, 02:17 PM »
It should be called for quick reply, but maybe not for entities.

If they're completely empty, they shouldn't be saved, certainly in the original implementation they wouldn't be sent in the first place, but I know it changed since then to be triggered by the events of typing something.
1777
Bug reports / Re: Weaving & Warm issues in iOS
« on January 25th, 2013, 11:43 AM »
No visible difference for me on iOS 5 w/Safari...
1778
Bug reports / Re: Minor bug with drafts and entities
« on January 25th, 2013, 11:42 AM »
The entity conversion happens at save time, normally... See what saveEntities is doing, whether it's being called etc.

As for drafts, I have two drafts only and judging by the content, they'd been saved just as I pressed send, but possibly before it had returned with a cache id.
1779
Features / Re: New revs
« on January 24th, 2013, 11:28 PM »
(4 files, 8KB)

Revision: 1873
Author: arantor
Date: 24 January 2013 22:28:19
Message:
! Banned users could throw an error if they had made a post previously. (Display.template.php, index language file)

! More of the ban UI, it can now pull a ban in the DB into the editing form. Just saving to go, really. (ManageBans.php, ManageBans.template.php)
----
Modified : /trunk/Sources/ManageBans.php
Modified : /trunk/Themes/default/Display.template.php
Modified : /trunk/Themes/default/ManageBans.template.php
Modified : /trunk/Themes/default/languages/index.english.php
1780
And it's one of the reasons Drupal is an architectural mess.

However, it's not that big a deal for us to implement, really. We already have some structure around boards being different types. If we extend that to include albums, a topic in an album board by definition is a topic about a media item. We just need to ensure there is proper logic around allowing/disallowing items to be moved between them inappropriately.

There are a lot more issues that we need to deal with, though it's something that I can do some work on tonight rather than banging my head off the ban UI, I guess.
1781
It is a valid way to go but it is also a source of much complexity :/
1782
Just as a side note, I'd add that no-one who reads the Aeva thread seems to be aware that you're not posting in it. Perhaps next time it comes up I should give them a reminder...
1783
Bug reports / Re : Weaving & Warm issues in iOS
« on January 24th, 2013, 06:04 AM »
Ok, this huge size isn't a quirk just for Retina iPads. I will screencap it later though, time for bed now.
1784
Development blog / Re: It only took two guys two years...
« on January 24th, 2013, 05:36 AM »
There's a lot of change, and there's a lot of similarity too. Hard to say without knowing exactly what joins you're trying to make.

I can guarantee you you're not going to get something in the next two months that would be called stable. It's *possible* you might see a public release of some kind, but it won't be stable if it is. I don't know, really depends on how the next couple of months shapes up as to whether we do any kind of public release in that time, but no chance of being a stable one.
1785
Plugins / Re: Plugin user interfaces
« on January 24th, 2013, 03:46 AM »
Quote from Nao on January 23rd, 2013, 03:13 PM
For mental health safety reasons, I'd strongly suggest going for the post-review process.
i.e.: (1) author gets to upload plugin directly, (2) any users (or maybe a limited group of users, but not 'Customize team' :P) get to try it out and share comments or give a rating, (3) in case any issues (security mostly) are found, report to moderators.

It would be more... realistic. Although in the beginning, I should assume there'll be less plugins to play with.
I should add, I split this off from the plugin interfaces topic, it's more relevant here.

I can see it both ways, but I sort of need to make the decision now because it's all part of the DB structure.

On the one hand, pre-review means there is a better than nothing chance of preventing nasties. It also creates more work, primarily for me. But I know how long it takes to vet plugins, and there's all kinds of neat tricks that can be done to help with that. We can do all kinds of fun like validating plugins on upload, checking the plugin-info.xml file is valid, we can check the hooks that it requires, we can check the files for syntax errors, you name it, it's all doable.

And that's before we even get to human review. In fact, I think I'd be inclined to enforce that even if we go to post review so that nothing appears on the site to regular users if it hasn't at least passed a basic validation.

Pre-review does also put some weight on whoever is reviewing to do a decent job, but it does require also discipline in doing it regularly, not to mention having naturally high skill requirements.

Post-review lowers the workload considerably, of course, plus removes the liability ("but the team should have checked for that") but it raises the chance of someone getting something nasty through the doors.

I'm reasonably amenable to either, though in the case of post-review, I would make it clear that I would be moderating after the fact and reserve the right to remove plugins that are dangerous, and I suspect I will even be doing some post reviews at some point myself, to validate what's going on. Of course, that leads to having some kind of badge related to 'vetted by' status to provide some kind of stability to people who want that kind of thing.

Of course, plugins pushed out under the team banner (i.e. official plugins) will automagically have something special to indicate it.

OK, so let's go with the theory of post-review rather than pre-review. What should happen with beta plugins? Plugins that people want others to test prior to going live, that is - should they automatically just push them to the main repo, but perhaps with a 'beta' tag on them?