The app may maintain PackageInfo for apps already, so there's no need to keep asking the system for it. We could instead just work with what we already have.
Also, we discovered that chunking the package names isn't needed for newer Android versions. This is only relevant for whom has more than 999 apps installed.
We now hang on to the index file while streaming it for repo preview purposes. This avoids having to re-download that file and we can properly add the repo right away.
This then allows us to bring the user to the list of apps in that repository without it being initially empty.
This led to many hard to debug issues in the past. It is easier to always fetch fresh data and not cache it.
Previously, we needed the cache as a search index. Now, search uses all localizations, so the cache isn't needed anymore.
by inserting zero whitespace between their characters to help the existing sqlite FTS tokenizers to split them up.
We have considered splitting them up only at word boundaries, but after consulting native speakers decided to do splitting by chars instead.
Doing this is a hack, but due to the limitations of tokenizers currently available with sqlite, we saw no better solution. While the ICU tokenizer is available as well, it doesn't handle diacritics in other languages.
The zero whitespace is added to zh, ja and ko locales when saving their text to the database. It happens for app names, summaries and descriptions either when loading a full index or when applying diffs. Tests have been added for both cases.
This now also searches in descriptions, author and package names.
The search also considers all languages now and is insensitive to diacritics in most languages.
The AppMetadataFts needed to be re-created for this to work. A migration and a test for this have been added.
With the update to Room 2.7.0 the chunked queries stopped to return distinct values. This is fixed by only processing distinct values when merging the chunked results.
The test case for the chunked quries was also fixed. It did not work properly before because it only tested chunked quries and not the base case which does not need chunked packageInfo
If we are used to update the same repo at almost the same time, a race-condition can happen that tries to apply a diff to an already updated DB. We don't download anything while holding a DB transaction, so the download and check for the repo timestamp happens before we enter the transaction. However, we forgot to re-check the timestamp again within the transaction to be sure the DB state is still as expected.
If an update was available from a disabled repo with a higher priority, the update from an enabled repo with a lower priority would not be offered as an update (i.e. returned by DbUpdateChecker#getUpdatableApps()).
historically, repos were added to the DB without much information and could stay in a broken state until manually removed. If a repo is updating, it should have a cert. So only repos that never did a single index update don't have a cert. Nowadays, this can not happen anymore as we get the repo and its cert before adding it to the DB. So whenever we update a repo, we know its certificate and fingerprint. Thus, this includes a DB migration removing all broken repos and making the certificate for repos in the DB non-null.
On versions before 1.18.0, adding a repo would add it as "empty" to the DB without a certificate right away and if there was an issue with it (e.g. typo in URL), it would stay in the DB until manually removed.
Our migration code did not consider such repos.
The client was already auto-incrementing their weight anyway. But it leaked our internal weight handling into the library consumer which can cause issues like when me are making changes to how we handles repo weights as we are doing now.
so they do not override the information from older repos anymore. This is especially an issue for the official repo which historically had the lowest priority while it should have the highest.
for now only for initial repos, but eventually we should validate for CoreRepository as well to prevent invalid certs from entering the DB by other means.
This is a breaking change.
The new method is in FDroidDatabase and called clearAllAppData(). It now also resets repository timestamps, so we will not try to apply a diff when we should use a full index.
