3.2.50 Beta (Android Oreo + enhanced TTS)

[mikelove]

On Play now. (enroll in betas at https://play.google.com/apps/testing/com.pleco.chinesesystem)

Changes:

• Added preliminary support for Android Oreo and two interesting new features from it:
  • Screen Reader now supports Android's new precise text location API, so recognized Chinese characters in some UI controls (including much of WeChat, thankfully) should now appear precisely aligned with the location of the original character.
  • System text-to-speech playback now supports word-by-word highlighting, but this won't be usable until Google releases their Android Oreo update to Google TTS.
  • We also added an Adaptive Icon so our app will look up-to-date on Oreo devices.
• Updated to modern Android-7-style notifications for Clipboard Monitor + for controlling Screen Reader, on devices running Android 7/8. (though we're not sure how useful the latter is now that it can also be invoked by a Quick Settings Tile)
• Added a startup alert when our built-in PLC dictionary is missing.
• Made Chinese text centered a bit more nicely in the search box.
• Made the 'edit entry' button appear in definition screens even if you haven't created a user dictionary yet (but own the flashcard module that would allow you to do so). If you don't like this change, you can disable it in Settings / Definition Screen.
• Removed non-functional audio play icons from Card Info.
• Fixed a bug that could cause crashes when auto-generating pinyin for cards that have no Chinese characters in their headwords.
• Fixed a bug that could keep Clipboard Monitor from triggering if copying new spans of text too quickly.
• Fixed a bug that prevented some text from extracting correctly on 'share'd web pages.
• Fixed buggy placement of highlights in Text File Reader.
• Fixed a bug that could prevent some checkbox preferences from actually saving their values correctly.
• Fixed a bug that could make our TTS stop playback when it encountered a rare character.
• Fixed a crashing bug when playing audio with TTS for flashcards that lack a Pinyin/Cantonese reading.
• Fixed a bug that could make CC-CEDICT stay visible as an available download in "Add-ons" even when it's already installed.

That Screen Reader API is a bit buggy, but works really well when it does work - once they get around to supporting it in web pages instead of just TextViews it could be a real game-changer (would definitively establish Reader > OCR at any rate).

Not a massive update, but on the plus side we were able to turn it around quickly while on iOS it's going to take us months to properly support ARKit + iPad drag-and-drop. And the Adaptive Icon should finally address the several years of complaints about our icon not looking sufficiently Android-y.

Also, some advance warning: we're going to be dropping support for Android 4.x pretty soon, so if you're still running some flavor of that I'd recommend you start researching your options to upgrade to Android 5.

 

[mikelove]

Released another beta, now called 3.2.50 (have gone through a couple of staged rollouts in the interim since we weren't getting enough Android O users with beta alone), with three major changes:

• Added support for new "enhanced" versions of our Qiang, Hong, Kayan and Kaho TTS voices, which you can download for free in "Add-ons" if you already own (or purchase) one of those voices. Can also visit their add-ons pages to compare low and high quality samples (which have also gotten longer in general) - the data files are extremely large (1+ GB) but the audio quality is really pretty amazing, especially if you use headphones or Chromecast or some other better-than-a-built-in-phone-speaker system.
• Added the oft-requested option to choose a specific TTS voice for each gender for Mandarin headwords in Settings / Audio.
• Brought back the 'open directly in app' option for Clipboard Monitor that we had been forced to disable in Android Oreo - it still puts up a notification (has to) but now that notification just tells you it'll open the app when the clipboard changes.

Also a modest update to PLC fixing a few hundred buggy Pinyin / traditional mappings (latter mostly in words with commas) and removing a reference to the Republic of China ending in 1949 that somehow slipped through our propaganda filter.

(enhanced audio + TTS voice selection for headwords also coming in a just-submitted iOS update, which also finally makes New PLC built-in on iOS now that Apple has bumped the OTA limit to 150 MB)

 

[澳]

Any possibility about Yafang TTS getting the high quality audio as well?

 

[mikelove]

That’s up to the people we licensed it from - at the moment they don’t offer it in a high quality version (I believe it pre-dates the other 4 new voices by a bit) but if that changes we’d certainly try to get it in Pleco.

 

[Fox]

My build is listed as 3.2.51, so I am uncertain as to whether you want this feedback here in this thread or not, but in my clipboard reader, the pop up definitions are appearing incorrectly. They always show up in the upper-left corner rather than near the word being defined, and the word being defined is not highlighted while the definition is being displayed. If I alter the settings to make the bubble appear at the bottom of the screen, it does, though the word remains unhighlighted. The same problem does not manifest in the pdf reader: the word being defined is highlighted, and the definition bubble floats near it.

 

[mikelove]

We should have a fix out for this in a few days, in the meantime if you turn off the ‘paginate’ option in Settings / Reader that should work around this issue.

 

[mikelove]

Fix is on beta channel now (signup link in first post), sorry about that.

 

[Bvo]

There is a bug, steps to reproduce:

1) type qiri
2) change q to c
3) there will be the only one entry.
Adding space (ciri -> ci ri) will show several entries.

After that even deleting the space (ci ri -> ciri) will not hide entries.

 

[mikelove]

Hmm, not seeing this here - could you email or PM me a backup of your settings? (Settings screen / Backup)

 

[mikelove]

Looks like this most likely relates to your user dictionary - is the dictionary switch icon changing when you do some of these searches? (going into / out of that dictionary because the system only finds results in it)

 

[Bvo]

@mikelove ,
yes, the icon is really changing.
But to Adsotrans dict, not to an user dict.
I think it's a wrong behaviour.

 

[mikelove]

Ah, that explains it - it's coming up in Adso because 'ciri' also matches a full-text English result there ('Cirillo'). It's switching to an English search because when you delete the first letter of 'qiri' you end up with 'iri' which is not valid Pinyin and hence only finds matches in an English dictionary.

 

[Bvo]

@mikelove ,
yes, it works this way, but it's surely not an expected result...
Why it switches to Adso only? It's even C-dict, not an English one...

 

[mikelove]

Because when it sees 'iri' it tries to get you the best match in that, which happens to be a result from Adso.

If you'd rather it not switch dictionaries on you unless you manually tap the dictionary switch button, go into Settings / Search Engine and set 'dict search order' to 'current only.' However, with that option configured that way you'll also have to tap on that button to switch between English and Chinese searches.

 

[Bvo]

No, it swithes not on 'iri', but on 'ciri'.
On 'iri' it just switches to E-dicts. On 'Ciri' - to Adso (Chinese(!) dict).
Anyway, do you agree that if just adding and removing a space changes results - it's not correct? It's not intuitive and not logical for an average user.

The option helps, thanks - for my use-cases manual switching is even more convenient than auto.

 

[mikelove]

No, the space is useful because it provides disambiguation. For example if you type "xian" and get a lot of results for the single syllable 'xian' before you get any for 'xi'an', it's easier to type a space between 'xi' and 'an' than to type an apostrophe, and unintuitive to still get results for 'xian' when you've typed 'xi an'. There are a lot of places in Pleco where an imperfect solution is nevertheless the least bad one, and I think this is one such place.

"Ciri" is searching Adso for English words because we do that for all C-E dictionaries; if you disable the option in Settings / Search Engine to include fulltext in E-C searches and also disable the "full-text, C dicts" dictionary group in Settings / Manage Dictionary Groups / English, that will prevent Pleco from searching those.

 

[Bvo]

"Ciri" is searching Adso for English words because we do that for all C-E dictionaries

It's OK (and actually is great!), the question is why it searches one C-dict for English words only and doesn't search other C-dicts for Chinese words?

About spaces - your example shows a good decision.
But I'm talking about another case: adding+removing a space returns different results, not about differences between "with" and "without".
In my example both searches are 'ciri' and between them was just one operation "adding+deleting a space". But results are absolutely different.
If you think it's correct - OK.
Anyway, my question is answered

 

[mikelove]

It's searching for English words first because deleting the 'q' put something that looked like an English word in the search field, so it assumed it should switch languages. Putting a 'c' in left it with something that could be English or Chinese, and since at that point it was already in English mode it stayed there.

 

[HW60]

In 3.2.52 I sometimes find cards in the word list marked as flashcards, which do not belong to my flashcards. In the screenshot, only the second 尚未 is one of my custom flashcards, the first with english definition is not. It may be that it was a dictionary related flashcard before I changed it into a custom card very long ago.

Pleco Reader shows only the first page of a file when Paginate text is not set. If set, all pages can be seen.

Edit: Pleco Reader now working again with Paginate text off (3.2.52 of Nov 2)

 

[Bvo]

@mikelove ,
I still cannot get all results in a predictable way...
The example we discussed above returns Adso entry 'Cirillo', but in manual mode (neither auto too!) I cannot get that entry, even if I switch to full-search.
How it works and how I can get that Adso entry with 'Cirillo' without the trick above?

PS: in Settings 'Integrate full-text search' is on for both C and E.