|#251 by yorhel|
2018-03-10 at 07:38
|#252 by shinnew|
2018-04-08 at 08:39
|Binkan Joshikousei, Tsukushitai no! doesn't show up in the edit form for Arai Riko Fujimoto Mihane and Kanbara Mana and I am unable to remove their relation to it. This happened after deleting the Binkan Joshikousei, Tsukushitai no! entry.|
|#253 by yorhel|
2018-04-08 at 09:09
|@252: The relation can be removed by editing the character, then changing something in the VN tab, changing it back to how it was, and then submitting. It's an ugly workaround, but it works.|
What I don't understand is why the relation doesn't show up even after undeleting the VN entry. I'll look into that.
EDIT: Apparently that happens when the releases have been deleted. I'll see if I can fix this.
EDIT#2: Okay, should be fixed now. VNs should now show up in the character edit form even when the VN or its releases have been deleted.Last modified on 2018-04-08 at 09:44
|#254 by leery|
2018-04-16 at 14:37
Read d9#2 for information about how to specify boards.
d9#2 is missing instructions on how to post to multiple boards...
|#255 by leery|
2018-05-10 at 18:08
|The tag browser has malformed HTML.|
All those greater-than symbols are unescaped.
|#256 by yorhel|
2018-05-10 at 18:40
|Malformed XML according to what? Technically, the '>' symbol doesn't have to be escaped, it cannot be misinterpreted as a tag because the '<' is properly escaped.|
|#257 by leery|
2018-05-10 at 18:47
|According to me. I'm the highest authority on XML and HTML -- didn't you know that?|
Unless HTML5 has changed things, <>& always have to be escaped.
Yes, browsers have been "correctly" rendering them for two decades, but malformed is still malformed.
|#258 by yorhel|
2018-05-10 at 19:00
|That's not a compelling enough argument to me. :) I know for sure that XML does not require > to be escaped - I've read the spec enough times while writing yxml. Reason for the change. Best practices be damned when I can't find a good enough reason to follow them.|
If that commit actually breaks stuff, then that changes things, of course.
|#259 by leery|
2018-05-24 at 06:32
|Every tag in tags.json link seems to have meta => true|
|#260 by yorhel|
2018-05-24 at 16:54
|@259: Evidently nobody ever used that field 😕. Fixed.|
|#261 by skorpiondeath|
2018-05-24 at 18:25
|Talking about HTML 5:|
I just tried an HTML 5 validator online and > does not launch a validation error. To be malformed a tag should not be closed but that's not even considered a tag.
Ofcourse < and & should be always escaped.
Talking about XML:
< and & should not be always escaped...they could happily exist inside <![CDATA]>
|#262 by leery|
2018-05-26 at 07:44
|So, does a tag rejection delete a user's tag history for the tag, or just hide it from view?|
I don't consider either one of these options good.
(@261: I'm still at HTML4 land.)
|#263 by yorhel|
2018-05-26 at 07:52
|Votes on deleted tags are deleted. I'm not sure I gave that much thought...|
|#264 by leery|
2018-05-26 at 08:13
|Well, I better start screenshotting my tag proposals then. In case I need to defend them.|
(See also: #244)
|#265 by leery|
2018-07-14 at 06:20
|Why do I get the same results when I search for ドル link as I get when I search for トル link|
|#266 by yorhel|
2018-07-14 at 06:25
Search is hard.
|#267 by leery|
2018-07-14 at 06:28
|Yep, I've seen that you strip accents (and similar) from searches. Which is good.|
I just don't think the dakuten/handakuten marks are an accent.
|#268 by yorhel|
2018-07-14 at 06:37
|You asked a question, I answered. :)|
I didn't even realize that those marks were combining characters in Unicode, so this was an unintended side effect. As you've probably guessed. I wonder if I can exclude kana/kanji from that regex.
|#269 by leery|
2018-07-14 at 06:52
|Add a second regex to recompose stuff?|
use Unicode::Normalize 'compose';
...I think. My ranges may be off by one.
I already had to write a script to re-compose filenames once because Apple's file system only does decomposed unicode and files that went through my Apple system had their filenames broken, sort of... Well, basically, nothing else uses decomposed unicode so searching and copypasting was broken-ish.
I don't think kanji has any composed/decomposed forms btw.
|#270 by leery|
2018-07-14 at 07:11
|While you're at it, I think ～ and 〜 should be normalised to ー|
There are quite a few titles where e.g. すくーる gets written as the cutesy すく〜る
Or すいーと as すい〜と
(Well, it's already cutesy because it's in hiragana but they're making it double cutesy with the tilde.)
(Yes, I know that you're removing 〜 because it's used to delimit subtitles.)
Or you could strip ー out altogether. It's 'wrong', but it has the same effect.
(I think ー is used to delimit subtitles too sometimes.)Last modified on 2018-07-14 at 07:17
|#271 by yorhel|
2018-07-14 at 13:44
|Both implemented: link|
|#272 by rampaa|
2018-08-01 at 04:45
|Not sure if it's important enough for you to care about but I will lay it out in any case. There's this thing called Turkish-İ problem. Basically, the uppercase of i is İ and the lowercase of I is ı in Turkish. See here and here for more detail. So, ideally, I should be able to get the same results when I search for, say, "kazıklı" and "KAZIKLI". But, unfortunately, when I search for "kazıklı" I get this result (as I should) but when I search for "KAZIKLI", I get no result. Same goes for "kıpırtı" and "KIPIRTI". When I search for "kıpırtı" I get this (again, as I should) but when I search for "KIPIRTI" I get no result at all.Last modified on 2018-08-01 at 04:54|
|#273 by yorhel|
2018-08-01 at 15:52
There's this thing called Turkish-İ problem.Fixed.
|#274 by rampaa|
2018-08-01 at 22:22
|Thanks Yorhel, you are awesome.|
|#275 by wakaranai|
2018-08-16 at 06:05
|unable to remove alias from Zeal Blood after making a split into Ooami Gou, server returns 500 error. probably it is still linked to a deleted VN or something.|
You must be logged in to reply to this thread.