* In this update, we've replaced the `URLConnection` with the OkHttp library, and now headers are being correctly passed in the request. As a result, the Zim file is successfully downloading.
* To integrate OkHttp into our Gradle project, we've moved the OkHttp dependency to our `buildSrc` folder.
* We have remove the authentication the url to make it normal url because we are adding authentication in headers so does not need to add in url.
* Enhanced getting secret function, if url is encoded then it will fail to get the secret from it so now we are properly decoding the url before extracting the secret key from it.
* To properly test these two functions we have added test cases for it.
* This test case is designed to verify that the functionality works correctly, especially addressing the issue where `LocalFileTransferFragment` would shut down when attempting to search nearby devices for connection.
* In Android 13, we previously used `intent.getParcelableArrayExtra` to retrieve the status of `P2P_DEVICE_CHANGED`. However, this method returned null because it's not actually an array. To resolve this issue, we have updated it to use `intent.getParcelableExtra`.
* For Android 12, we used to check the network state through `ConnectivityManager` every time the `WIFI_P2P_CONNECTION_CHANGED_ACTION` action was triggered. However, this approach wasn't reliable, as the action being triggered didn't guarantee a change in the network state. Instead, it sends an intent that we should now check for network changes. As a result, we have modified our code to utilize this intent within the `KiwixWifiP2pBroadcastReceiver`.
* We are showing a custom error message to the user for this type authentication urls because we does not have any authentication values in our application so it will always fail to download the zim file, so we are showing a error message where we inform to user to reinstall application from the playstore. Since play store application have the zim file in it.
* We have removed the dwds dependent code, now we are using `Regex` expression to match if it is authentication url then we are adding `Authorization` in header.
* To ensure this methods works perfectly in every scenario we have added test cases for it.
* We have removed the hardcoded secret name from the code, now we are extracting secret name from the url.
* Now we are getting userName and password from single secret.
* The issue was that we were not properly uninitialized the service worker, which was sometimes causing problems and preventing us from reading the `Zimit-generated` zim files. Therefore, we have now implemented the proper uninitialization of the service worker before loading the URLs for new zim files.
* Added logging for cases where the `ZimFileReader` failed to find an entry, failed to load an asset, and more.
* Significantly improved the "load more" functionality: Now, if the user is near the end of the list, we start loading new search results. This enhancement enhances the user experience by reducing wait times for new search results.
* Added references in the `KiwixServer` class to help developers understand why we are keeping the library object.
* Implemented test cases to thoroughly test the `getMimeTypeFromUrl()` function in the ZimFileReader class.
* Enhanced the search functionality by consolidating the use of a single viewModelScope instead of multiple coroutines, reducing the potential for future issues.
* Previously our test cases were launching the `KiwixMainActivity` twice (firstly it sets the values in preference and then relaunch the activity) before running the test case which was the cause of slow testing and sometimes it caused to memory leak in some api levels so we have refactored our test cases to launch `KiwixMainTest` once per test case.
* `IntroFragmentTest` is failing on api level 24 due to a memory leak because after clicking on getStarted button is going to the LocalLibrary page but somehow `onStart` and `onStop` methods are calling of the `CoreReaderFragment` which stablish the `serviceConnection` but it was happening very quickly so before attaching the binder to `readAloudService` unbindService method called but at this point service was not created but creating was in progress so after going to library screen it allocate memory to `readAloudService` but `ReaderFragment` is no more visible to the user that's why memory leak happened. We have fixed it by unbinding the service in `onDestroyView` method.
* A bug was introduced after enhancing the search functionality to align with the new `java-libkiwix` wrapper. Initially, when searching for any article in the ZIM file, the loading progress bar was not being displayed. This commit resolves this issue.
* While searching within large ZIM files, the application used to freeze momentarily due to fetching data from the `libkiwix` on the UI thread. We have improved this functionality to provide a seamless user experience.
* Added a loading progress bar at the end of the RecyclerView when loading more search results as the user scrolls to the bottom. This indicator informs users that additional results are being loaded. The progress bar appears if there are more results available for the search term, providing users with visibility into ongoing loading.
* Enhanced the search loading process for larger ZIM files by introducing coroutines. This background threading approach prevents the UI thread from being impacted and ensures a smooth scrolling experience for users.
After addressing the issue documented in https://github.com/kiwix/java-libkiwix/pull/61, we now have the ability to search within zim files that do not have a Xapian index. As a result, we have enhanced our test to utilize pre-existing zim files. This improvement leads to reduced time consumption, improved memory efficiency, and minimized network usage impact.
* Since `Search` is not compatible with those zim files which does not have Xapian index but `SuggestionSearch` have this functionality to search inside those zim files so we have used this.
* Update test cases for test new search functionality.
* We are previously checking `hasEntryByPath`, `hasEntryByTitle`, `mainEntry.isRedirect` which are internally calling the same function as we are calling after checking this condition so it would be better to directly use those function to avoid calling same function twice see more details https://github.com/kiwix/java-libkiwix/pull/60.
* Due to linking errors with libkiwix/libzim functions, direct usage in testing was not possible. To address this, helper classes were created, similar to those in `java-libkiwix`, for testing the search functionality.
* With the integration of libkiwix12, we now receive full results based on the search term. Previously, we were loading the entire list, but we have now implemented pagination to enhance the search functionality.
* The libkiwix provides us with a Search object, which enables us to obtain the suggestion list using start and end indices. Consequently, we have modified our `ZimSearchResultGenerator` code to return the `Search` object instead of a list.
* To accommodate the changes, we have updated the return type of `SearchResultGenerator` to a nullable Search. This change is necessary because we initialize the `SearchState` when the search is initialized in `SearchViewModel`, and initially, we do not have the Search object available. The nullable return type allows us to pass the Search object when it becomes available.
* Since our minimum API level is now 24, we have updated our CI configuration to run on API level 24, which aligns with the minimum API level required for our project.
* Now our minimum SDK version is 24 and we are using some conditions placed on behalf of this api level which are unused now, so we have removed those conditions.
* Removed `CompatV21` file as now it is unused.
* Renamed `CompatV23` to `CompatV24` and refactored the code to support our new minimum api level.
* Refactored `NetworkUtilsTest` to support api level 24.
