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
1321
Plugins / Re: Hooking into e-mail templates
« on March 11th, 2013, 03:53 PM »
There are two issues that result.
Firstly, loadEmailTemplate won't load your plugin's language file though you can do that yourself before calling.
Secondly, this gets tricky because the main language editor won't know to exclude your plugin's language entry and will attempt to load it anyway, meaning it would be editable from the main language editor rather than the email templates area. (Or it will when I finish it :P)
What this might mean is that I add the option to splice an existing plugin's language file into the DB, and reuse that option to exclude the same language file from the plugin area of the language editor.
Firstly, loadEmailTemplate won't load your plugin's language file though you can do that yourself before calling.
Secondly, this gets tricky because the main language editor won't know to exclude your plugin's language entry and will attempt to load it anyway, meaning it would be editable from the main language editor rather than the email templates area. (Or it will when I finish it :P)
What this might mean is that I add the option to splice an existing plugin's language file into the DB, and reuse that option to exclude the same language file from the plugin area of the language editor.
1322
Plugins / Re: Hooking into e-mail templates
« on March 11th, 2013, 03:46 PM »
Remember where I said about plugins being able to extend the language files? The exact same process would apply ;)
Though yes there needs to be a hook to extend the actual definitions too.
Though yes there needs to be a hook to extend the actual definitions too.
1323
Features / Re: Mini-skeletons
« on March 11th, 2013, 02:22 PM »at least it'd become a simple boolean test... What do you think?
Haven't had chance to try it yet but I'm excited with the direction things are going :)
1324
Plugins / [Plugin] Re: User Mentions v1.0 (9th March 2013)
« on March 11th, 2013, 02:02 PM »
I only dread how big the index is going to be for that table (since the index size is 3/4 of a KB PER ROW), but it's still relying primarily on the where to sift.
1325
Plugins / [Plugin] Re: User Mentions v1.0 (9th March 2013)
« on March 10th, 2013, 04:26 PM »
If it's working, that's all good :)
1326
Plugins / [Plugin] Re: Notifications system (1.0)
« on March 10th, 2013, 04:25 PM »
What I'd be inclined to do is move the user info bit out of the sidebar and up into the top right under the search area, and put notifications up there.
1327
Plugins / [Plugin] Re: User Mentions v1.0 (9th March 2013)
« on March 10th, 2013, 05:59 AM »
Well, that's where it gets complicated. column IN (value1, value2, value3) ultimately parses as (column = value1 OR column = value2 OR column = value3), which is not efficient.
Now, there is an index on real_name, but I'm not sure how useful the index will be. Create a table with 1000 rows in it and try the 66 permutations on it with an EXPLAIN on the query to see how that ultimately parses out, but I'd be willing to bet it probably will be Using where, not Using index. If it's Using where, it's going to be table scanning anyway just as it would with a LIKE, bearing in mind you only need to do a like against the first part if I've understood you correctly.
If in doubt, try it and benchmark it. I'm only guessing based on what's been described thus far, but I wouldn't expect the index to be that useful.
Now, there is an index on real_name, but I'm not sure how useful the index will be. Create a table with 1000 rows in it and try the 66 permutations on it with an EXPLAIN on the query to see how that ultimately parses out, but I'd be willing to bet it probably will be Using where, not Using index. If it's Using where, it's going to be table scanning anyway just as it would with a LIKE, bearing in mind you only need to do a like against the first part if I've understood you correctly.
If in doubt, try it and benchmark it. I'm only guessing based on what's been described thus far, but I wouldn't expect the index to be that useful.
1328
Other software / Re: Discussing Wedge on simplemachines.org
« on March 10th, 2013, 05:38 AM »Maybe SM Core will fix this
1329
Test board / Re:
« on March 10th, 2013, 01:44 AM »
That would be because we now have a ton of stuff in our DB here that is not properly escaped. Which for them means no practical difference.
1330
Plugins / [Plugin] Re: Improved attachments upload
« on March 9th, 2013, 11:53 PM »
Sure I remember that. I'm still convinced it's ultimately not a blocker for it being a core feature, but simply that the selector is not as selective as it should be about grabbing events.
Mostly I think Nao and I have been quietly hoping the other will have a sudden idea on how to solve the architectural issues that come out of it but I think that in the end I'll just have to go in and do it. There is one aspect to it that is rather awkward to solve but not unsolvable, it's just figuring out how to do it without breaking the crap out of everything.
Mostly I think Nao and I have been quietly hoping the other will have a sudden idea on how to solve the architectural issues that come out of it but I think that in the end I'll just have to go in and do it. There is one aspect to it that is rather awkward to solve but not unsolvable, it's just figuring out how to do it without breaking the crap out of everything.
1331
Archived fixes / Re: Flexbox, meet code
« on March 9th, 2013, 11:17 PM »
Did you specifically see the topic in question?
1332
Archived fixes / Flexbox, meet code
« on March 9th, 2013, 10:55 PM »
As per the attached picture, something's funky with the code tag - the scrollbar appears in the wrong place (behind other text) and doesn't work properly in that it isn't selectable by mouse through no position I can find...
1333
Plugins / [Plugin] Re: User Mentions v1.0 (9th March 2013)
« on March 9th, 2013, 10:53 PM »
Suggestion: take the first part, htmlspecialchars it then query members LIKE 'firstpart%', then sift the results post-query. It may be faster for the worst case scenarios.
Also, I'm thinking there's a bug in the flexbox on Chrome. I'll post a screenshot shortly.
Also, I'm thinking there's a bug in the flexbox on Chrome. I'll post a screenshot shortly.
1334
Plugins / [Plugin] Re: Improved attachments upload
« on March 9th, 2013, 10:48 PM »* Arantor bangs the core feature drum :P
Trouble with making things into core features is that I'm not a fan of 'core plugins' and that prevents easy linkage back to repos.
1335
Plugins / [Plugin] Re: User Mentions v1.0 (9th March 2013)
« on March 9th, 2013, 02:08 PM »
\w+ shouldn't match the ", but if you're doing that on the client side it should be fine.
(It is more of an issue on the server side because all display names and usernames are all pushed through htmlspecialchars first)
Regarding core vs not core, I found through bad experience that 'core plugins' do not work that well from a user perspective. They embody everything that is wrong with Core Features: having to go to an unrelated page that was convenient for the developers to just grab everything together, and it is a proven cause of support requests. Not that there won't be support requests anyway but at least some people will use search and just straight up exploration to find it otherwise.
I'd rather either have it totally as a plugin, or fully core, I don't see that any intermediate measure suits anyone but us from a purely convenient point of view.
I would love to see both this and the notifications core being core features and completely integrated. Seems to me that might be faster?
(It is more of an issue on the server side because all display names and usernames are all pushed through htmlspecialchars first)
Regarding core vs not core, I found through bad experience that 'core plugins' do not work that well from a user perspective. They embody everything that is wrong with Core Features: having to go to an unrelated page that was convenient for the developers to just grab everything together, and it is a proven cause of support requests. Not that there won't be support requests anyway but at least some people will use search and just straight up exploration to find it otherwise.
I'd rather either have it totally as a plugin, or fully core, I don't see that any intermediate measure suits anyone but us from a purely convenient point of view.
I would love to see both this and the notifications core being core features and completely integrated. Seems to me that might be faster?