Maksim Ivanov

BTW, in case anyone would like to contribute the disconnect/disable workaround proposed in this issue (as on our side working on this issue isn't prioritized yet), the implementation could follow...

Some pointers: * Documentation regarding the dialog and the "force_allowed_client_app_ids" policy: [docs/connector-app-api.md#permissions-for-the-pcsc-api-consumers](https://github.com/GoogleChromeLabs/chromeos_smart_card_connector/blob/master/docs/connector-app-api.md#permissions-for-the-pcsc-api-consumers) * Code that decides whether to show the dialog or use the policy is at //third_party/pcsc-lite/naclport/server_clients_management/src/permissions_checking/checker.js (specifically, see...

We might want to make progress on this one, since we're getting a user report about this repeated crash, for which we don't have specific repro steps.

As a quick update, the improvement made in #156 reduces the severity of this issue: we now stop self-reloading in case we detect crash loops, which means that we do...

The minified example is: ``` goog.require('goog.testing.jsunit'); goog.global['testMy'] = { 'setUp': function() { console.error('setUp called'); }, 'tearDown': function() { console.error('tearDown called'); }, 'test': function() { console.error('test called'); }, }; ``` It...

Some potential directions to investigate for this bug: - It might be some deadlock in the CCID Driver code; - It might be a bug in our libusb-to-chrome.usb bridge implementation;...

And one more idea: * It might be a bug in the middleware app, combined with the insufficient sanity checks in the Smart Card Connector app (#136).

@hcldan Do you mean this particular bug report (delayed NaCl initialization in tests) or the Smart Card Connector in general? For the former, we're not actively investing into NaCl support...

@hcldan I see, thanks for explaining the use case. Based on what we just discussed, there doesn't seem to be a way to write such a single implementation that would...

The tests hang on bots. It seems the existing "SCardConnectProtocolMismatch" test should also simulate an (empty) policy receipt, so that the tweak in SCardConnect doesn't wait for the policy infinitely...