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 - Nao
2911
Archived fixes / Re: Updated CSS code is buggy
« on June 1st, 2013, 10:03 PM »
Probably best, yeah... I'll try to ftp into this tomorrow.
2912
Features / Re: New revs
« on June 1st, 2013, 12:26 AM »
rev 2144 -- just to get this one out of the way, I guess...
(1 file +2-2, 178kb)

* Updated jQuery to versions 1.10 and 2.0.1, as Google is now hosting them. They're one week late! (Load.php, jquery-1.10.0.min.js, jquery-2.0.1.min.js)
2913
Archived fixes / Re: Updated CSS code is buggy
« on June 1st, 2013, 12:04 AM »
Hmm.
Can you print_r the contents of $this->rules after it's set up..?

Perhaps I should do a double check on indentation at that moment and die if there is an error...

Or, err... just use the bracket format, which is supported already. Then no need for indenting. But, well... it would feel like a defeat, to me. Just like Sass lost to Scss (oddly.)
2914
Archived fixes / Re: Updated CSS code is buggy
« on May 31st, 2013, 12:23 AM »
Reminder ;)

Bed time! I've read about 10% of today's posts... freaky!
2915
Archived fixes / Re: Updated CSS code is buggy
« on May 30th, 2013, 09:14 PM »
Can you try with r2143 then..?

BTW your bugs would imply an indentation error. Did you add any custom CSS files..?
2916
Off-topic / Re: Next System Visitation
« on May 30th, 2013, 08:51 PM »
An entry for fluxbb would be nice, too. It's fast and has very little features (last time I checked, avatars were a plugin), but I've always had a soft spot for it. I think you already looked into it, back in 2010. ;)
2917
Features / Re: New revs
« on May 30th, 2013, 06:38 PM »
rev 2143
(1 sin, 498 bytes) (I think it's the first time I'm getting a commit size in *bytes*, really...)

! Apparently, not fourty-two. Because I love doing meself some last-minute refactoring with an extra layer of bugs. (Subs-Cache.php)
2918
Off-topic / Re: json_decode and arrays
« on May 30th, 2013, 06:26 PM »
Well, apart from the indentation and multiple echos... :^^;:
2919
Plugins / [Plugin] Re: Notifications system (1.0)
« on May 30th, 2013, 06:19 PM »
Nope, haven't moved past this... I'm just taking a break from notifications, after working on them for so long... :P

Anyway!

And yes, the show_ajax is old-school, I even said so in this topic I believe, but what alternative can you offer..? My current plan is to use Zoomedia's ajax stuff as replacement, but... Well, it's not my priority.
2920
Features / Re: New revs
« on May 30th, 2013, 06:11 PM »
rev 2142 -- What is twenty-one times two?
(6 questions, 4 answers)

! Create cache.lock file at install time, so we don't have to deal with it later. Also removed a few unused globals. (install.php)

! dynamic_group_colors function was using MySQL, which killed the install process when showing the HTML header. (Subs-Cache.php)

* Simplified cache_put_data/cache_get_data functions by moving their (similar) init code to a stand-alone function. Harmonized another function to use the same naming scheme. (Subs-Cache.php)

* French translation. (Install, ManageMaintenance, Modlog, Profile)
2921
Archived fixes / Re: Updated CSS code is buggy
« on May 30th, 2013, 05:44 PM »
This is crazy... :-/

First. Blank page... -_-
Traced it manually to the dynamic group color generator. I didn't notice in the first place, but you use wesql in it, and it's a no-no at this point, because obviously the database isn't even created yet, so...
Added a !class_exists('wesql') test at the beginning of the function.
Works.
Now, page 3... Right after entering my database details... Yay, another blank page.
Turned out that this was the same error; looks like, at this point, the wesql object exists, but the database still doesn't work, or something. So, error.
I replaced the test with if (defined('WEDGE_INSTALLER')).
Good, now I can get to complete the install process, eh...

Last of all. Once installed, an intro text says "you'll have to log in before you can use it blah blah..." Well, that's not true. I'm already logged in, at this point...

Worse, on that same page:

Warning: filemtime() [function.filemtime]: stat failed for /cache/cache.lock in /Sources/Subs-Cache.php on line 1628
Looks like the cache.lock file creator at line 1627 doesn't actually work, or isn't called at all because, ahem... the Cache folder already exists at this point, so your file_exists returns true, but you should be testing for the presence of cache.lock, not the cache folder itself...
If I fix line 1626, and reinstall, I'm still getting an error, because this time it's running cache_get_data() with cache.lock not around yet, so blah, it spits out another error.

My suggestion would be to simply force the creation of cache.lock in the installer...

The only errors in my error log are the ones relating to that one error... Uh.

I'm adding these fixes to my local install. I'll probably even commit them by the time you're back, so you can simply rework my changes if anything doesn't work well for you.
2922
Archived fixes / Re: action_list is called too late
« on May 30th, 2013, 09:22 AM »
Given the new position, is it really useful at all..?
One could always use the 'behavior' hook, and simply reset $context['action'] and $context['subaction'] as needed... No..?
2923
Features / Re: New revs
« on May 29th, 2013, 11:15 PM »
rev 2139 -- yay, micro-optimizations..?
(6 files, 4kb)

+ Added support for rect() retro-compatibility to Wess. Specifically, IE 6/7 expect commas not to be used, so you're now free to use the standard commas, and IE 6/7 will remove them. Just, hmm, don't use silly unrelated comma-separated values inside your rect() function, but... There's no reason to, is there..? (Class-CSS.php)

* Increased execution speed for retro-compatibility code. Initially, I was doing a smart implode on all test strings, to run a single preg_replace_callback. After a couple of tests, I realized that doing them individually, whatever the number of tests, was 5 to 10 times faster. Okay, then... I guess sometimes, benchmarking isn't a waste of time. (Class-CSS.php)

* Theoretically, increased execution speed for CSS layout, as I definitely gave up on text-indent hacks to focus on clipping. Costs a few extra bytes, but it's safer, and seems to work just fine in all browsers I tested in. (index.css, sections.css)

- Profile statistics added some redundant HTML that was subsequently hidden by CSS. Yeah, sure. Removed said HTML and its related CSS hack. (Profile.template.php, sections.css)

- Moved profile statistics pie to external. Seriously, there never was a need to embed this one, as it's only seen on one obscure profile page, and it saves a whooping 1.5KB from the final CSS file. Hopefully, I didn't miss anything else... I guess..? (sections.css, extra.rtl.css)

! A quick and easy fix to hide the related notification preview, if opened, if you attempt to mark the notification as read. (script.js)
2924
Bug reports / Re: PM Quoting
« on May 29th, 2013, 04:25 PM »
There are arguments pro and against floating boards, but...
2925
Archived fixes / Re: Subject 0 is treated as empty
« on May 29th, 2013, 04:24 PM »
Yep, the infamous 0 == ''... :-/