Enhancements to HttpCache::Transaction layer |
||
Issue descriptionCreating this to track enhancements that can be done in the HttpCache::Transaction layer mostly for making the code simpler: 1. After the cache lock fix is in place( crbug.com/472740 ), it would be better to have LOAD_DISABLE_CACHE transactions also to create an entry and go through Writers::Read. This will simplify HC::T::Read() to always invoke Writers::Read and will remove the need for DoNetworkRead* states. Writers can then reuse network_read_only_ flag to also be set in case of LOAD_DISABLE_CACHE. This will reduce the network_trans_ member of Writers to be only non-null for headers phase and can then be renamed to headers_network_transaction_.
,
Sep 12 2017
Another suggestion that came out of the code review https://chromium-review.googlesource.com/c/chromium/src/+/612400/2 From jkarlin@: "Isn't the same also true of any case where we might restart? E.g., RestartIgnoringLastError, RestartWithCertificate, and RestartWithAuth? It seems like we need to handle every place that the start machine returns an error. Perhaps, if the start machine fails, it should move to a START_MACHINE_FAILED state instead of FINISH_HEADERS. If Read() is called instead of one of the Restart* methods, then remove the transaction from the ActiveEntry and continue reading from the network. WDYT? It also seems like we need to document and DCHECK that consumers shouldn't call the Restart() methods if the start state machine didn't return an error."
,
Sep 13
This issue has been Available for over a year. If it's no longer important or seems unlikely to be fixed, please consider closing it out. If it is important, please re-triage the issue. Sorry for the inconvenience if the bug really should have been left as Available. For more details visit https://www.chromium.org/issue-tracking/autotriage - Your friendly Sheriffbot |
||
►
Sign in to add a comment |
||
Comment 1 by shivanisha@chromium.org
, Jul 11 2017