-
csharrison authored
A previous change [1] changed the URL parsing code to be prepared to allocate and reserve size for its own buffers, rather than relying on callers to do so. However, the patch did not address all callers, and regressed performance for callers that *did* reserve their own buffers, by reserving a bit extra space (8 bytes) to account for escaped characters. Instead, if a caller has a pre-reserved buffer that we think can fit the entire URL (minus escaped chars), we should just use it. [1] https://codereview.chromium.org/2617173003/ BUG=657978 Review-Url: https://codereview.chromium.org/2641823004 Cr-Commit-Position: refs/heads/master@{#447372}
60e6ff0e