Pleco - 'Common Mode'

Shun

状元
Looks fine on an LG Nexus 5X running Marshmallow 6.0.1: (using the same pqb file)

tmp_15819-Screenshot_20160316-20030922523619.png
 

mikelove

皇帝
Staff member
@Shun - are you using our latest 3.2.26 update? This looks like an issue brought out by that - we changed the way a couple of formatting tags work in that update and it seems like this database was using those tags in a nonstandard way. (main tags affected were the 'block' tag U+EAC3, which as far as I know we've never documented anywhere, and the 'color' tag U+EAC1, which I'm not aware of any widely-circulated user dictionaries employing) We never embed those tags directly in any of our dictionary databases (they're all generated by another, completely undocumented part of our database system) so we thought it would be a totally invisible change but it seems like there may have been a few user dictionaries using those tags after all.

Anyway, fixing it should just be a matter of replacing every string of two @@s with a string of four @@@@s - do a find-and-replace on the .txt file to make that change, then import it and it should work fine.

(re that color tag, if anybody does want to use that, now, instead of putting the color in the lower 15 bits of two characters, you put it in the lower 8 bits of 4 characters, with EC00 as the top 8 bits; we made a terrible decision years ago in allowing our numeric format tags to potentially overlap with valid characters and now we've finally fixed that by confining them all to the PUA)
 
Last edited:

Shun

状元
Oh, I was using version 3.2.25 from the Play store. Thanks for the heads-up!

Edit: I now have 3.2.26 & your fix worked.
 
Last edited:

Peter

榜眼
@Shun - are you using our latest 3.2.26 update? This looks like an issue brought out by that - we changed the way a couple of formatting tags work in that update and it seems like this database was using those tags in a nonstandard way. (main tags affected were the 'block' tag U+EAC3, which as far as I know we've never documented anywhere, and the 'color' tag U+EAC1, which I'm not aware of any widely-circulated user dictionaries employing) We never embed those tags directly in any of our dictionary databases (they're all generated by another, completely undocumented part of our database system) so we thought it would be a totally invisible change but it seems like there may have been a few user dictionaries using those tags after all.

My bad. That block tag is great because it uses less vertical space than "double carriage return". I found that tag by walking the U+EAxx parameter space, so you were right in thinking no one would be crazy enough to use it!

(re that color tag, if anybody does want to use that, now, instead of putting the color in the lower 15 bits of two characters, you put it in the lower 8 bits of 4 characters, with EC00 as the top 8 bits; we made a terrible decision years ago in allowing our numeric format tags to potentially overlap with valid characters and now we've finally fixed that by confining them all to the PUA)

Awesome, now all we need is a "dictionary of colors" using the solid block character (█).
 

Shun

状元
Can I ask, which text editor works best for you with special Unicode characters on a Mac? Is it vim? For long text files, TextWrangler is very slow, I have to wait 20 seconds between each step. Thanks!
 
Last edited:

mikelove

皇帝
Staff member
@Shun - TextWrangler's paid sibling BBEdit is one of the faster ones, actually. Ultra-Edit and Sublime Text are good bets too. None of them have stellar Chinese support, though, or are as fast as the best of the Windows text editors like EmEditor, so despite being Mac shop for really big files we usually end up using EmEditor in an emulator.
 

Shun

状元
Great information, thanks! Linux also seems to have some fast editors, now trying out vim on Manjaro Linux. :)
 

Peter

榜眼
Returning to the original topic... I think the 'chars' page would also benefit from some kind of a common mode.

Pleco currently displays all related characters from its characters database, even ones for it has no pinyin reading. Although sorted by frequency, IMHO it would be useful to further filter/colourise the results using a "common mode" list.
 

mikelove

皇帝
Staff member
Turn on "common compounds" in Settings / Definition Screen / CHARS tab - that will limit it to characters for which it has a reading + definition.
 

Peter

榜眼
Even though I turned on the common compounds option, Pleco still shows chars without reading & definition. Test case: search for the sick radical (疒) and go to the chars page.

I also tried disabling and then uninstalling the extended unihan character dictionary, but that made no difference.
 
Top