When going offline, all the rpc calls fail making hystrix open all the
circuits.
This commit changes the behavior so that when we go back online, we
close all the circuits.
We should also avoid making rpc requests while offline, but this should
work for release.
This commits does a few things:
1) Adds cache of token amount to the GetWalletToken endpoint, used by
mobile, in case the user is offline.
2) Split circuits by chain-id (when available) and by host+index when
not
3) It makes GetWalletToken always refresh, as that's directed from an
user action and we want to respect that. A cool down of 10s should be
added in the future to avoid spamming.
- Addressed the issue where remote image URLs failed to be processed due to local file handling restrictions.
- Introduced `FetchAndStoreRemoteImage` to download and save remote images to a temporary file, enabling compatibility with existing processing functions.
- Split the main image handling logic into smaller functions (`OpenAndDecodeImage`, `AdjustImage`) for better modularity and maintainability.
- Updated `OpenAndAdjustImage` to use these new functions, ensuring a streamlined flow for both local and remote images.
- Added cleanup steps for temporary files to prevent resource leaks.
resolves: #15691
* chore_: unused `BuildTx` function removed from the processor interface and types that are implement it
Since the `BuildTx` function is not used anywhere, it's removed from the code.
* fix_: resolving nonce improvements
When the app sends more than a single tx from the same account on the same chain, some
chains do not return appropriate nonce (they do not consider pending txs), because of
that we place more tx with the same nonce, where all but the first one fail.
Changes in this PR keep track of nonces being used in the same sending/bridging flow, which means
for the first tx from the multi txs the app asks the chain for the nonce, and every next nonce is resolved
by incrementing the last used nonce by 1.
Fixes https://github.com/status-im/status-desktop/issues/15930
The issue happened because I received a notification about a contact request, but the request was not associated with a message in the DB. The DB didn't have a message with the required ID in it.
This potential fix moves the creation of the contact request later in the handling code, because the only way this bug can happen as far as I see it is if there is an error later in the handler function and the message is never added to the response, and therefore never saved.