Since we mistakenly went ahead and called the last release 2.4.0, we're going to proceed with calling this one 2.4.1, even though it's still a "beta" and we may still need to fix a bug or two (rendering it 2.4.2) before we release it.
This is a larger-than-expected minor update, to fix what in retrospect should have been some rather obvious design flaws with our merged search system's handling of single character searches, flaws which we decided it would be a bad idea to ship the finished update with. Hopefully our fixes won't have made the app buggier or less stable (and in fact we've also fixed a few fairly serious bugs, so overall it should almost certainly be an improvement) but we'd really appreciate it if you'd let us know about any bugs that you do find.
http://cdn.pleco.com/androidapps/plecodroid-121227p.apk
Again, a number of dictionary databases have been updated, as have the main data files (to add a new traditional character frequency table), so you'll want to download the latest versions of all of those when you update.
TTS purchase links: Hui / Lulu
Changes
This is a larger-than-expected minor update, to fix what in retrospect should have been some rather obvious design flaws with our merged search system's handling of single character searches, flaws which we decided it would be a bad idea to ship the finished update with. Hopefully our fixes won't have made the app buggier or less stable (and in fact we've also fixed a few fairly serious bugs, so overall it should almost certainly be an improvement) but we'd really appreciate it if you'd let us know about any bugs that you do find.
http://cdn.pleco.com/androidapps/plecodroid-121227p.apk
Again, a number of dictionary databases have been updated, as have the main data files (to add a new traditional character frequency table), so you'll want to download the latest versions of all of those when you update.
TTS purchase links: Hui / Lulu
Changes
- Redesigned single-character searches so that they will now pay attention to your simplified/traditional character setting and sort by character frequency in your chosen character set - traditional users will no longer have to put up with 几 as the first result for "ji," or with 歷 and 曆 being needlessly grouped together even with 曆 being significantly less common and unrelated to 歷.
- Fixed an issue that could cause alternate character variants that had been merged into a search result to not be visible in the results, so that Pinyin searches for a particular character might not turn up any entries even if the dictionaries do cover that character.
- Made single character Pinyin searches always treat every independent character in the current character set as a separate result, again to aid in finding a particular character from a long result list (and to avoid having a ridiculous case like "hui2" give you a single entry with nine different /-separated characters).
- Separated the speed settings for the Lulu and Hui TTS engines.
- Fixed the rendering of very complicated variants so that if you have a word with multiple variants in both simplified and traditional, cases where more than one variant in one character set maps to a single character in the other will be separated with ;s instead of /s (so that you can tell which character goes with which).
- Fixed a serious bug with full-text searches that could sometimes bring up blank / irrelevant results (or even cause crashes).
- Fixed several major memory leaks that could cause Pleco's performance to significantly degrade over time.
- Greatly reduced the memory requirements of long search result lists, which should help with performance on some devices.
- Sped up searches on some devices with very slow memory allocators by reducing the number of allocations needed to grow the result list.
- Eliminated the option to configure the merge character set separately from the general traditional / simplified option, since it was proving too complicated to juggle conflicting settings for those, and since with multi-character entries we've opted to only merge on simplified (owing to the great variation in traditional versions of words among our mostly-Mainland-developed dictionaries, and the fact that there are only maybe half a dozen cases of different traditional words being converged to the same simplified one and that we already have another system for dealing with those).
- Fixed a couple of rare seemingly-random crashes which we turned up using some more advanced (iOS-only - good thing we're developing on both platforms) memory error finding tools.
- Removed the superfluous / confusing superscripts from some merged search results.
- Fixed a bug in HanDeDict that caused umlauts in definitions to be turned into 'v's, and more minor bugs in several other dictionaries (see the Updates tab).
- Fixed a bug that could cause OCR history entries to show up as "invalid," could cause OCR history views to crash, and could prevent changes to the OCR history from saving correctly.
- Added an alert on startup for users running extremely outdated (to the point of causing very slow searches) dictionaries.
- Started using the built-in font instead of our fallback Song font to display Unicode Extension A characters - if your system starts displaying lots of characters as boxes in this update, you can download Google's Droid Sans Fallback Full font with support for those characters at https://github.com/android/platform_fra ... ckFull.ttf (click on "Raw" at the top right corner of the file box) and install it with the appropriate option in Settings / General.
- Tweaked the merged dictionary icons slightly so that a full-text search of dictionaries in the same language (i.e. an English search of English-Chinese or a Chinese search of Chinese-English dictionaries) will show up with an inverted icon, so that you can easily differentiate these from regular searches.
- Fixed a bug that required you to download the TTS add-ons before you could play their audio samples.
- Applied our code for working around HTC / Sony Ericsson phones' keyboards refusing to acknowledge the standard Android "no suggestions" flag to flashcard fill-in-the-blanks tests that were set to use the system IME, as we had already done for the dictionary search field.
- Fixed a bug that could cause flashcard definition text to appear cut off in the background of stroke order tests.