[ https://issues.apache.org/jira/browse/OAK-2389?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14291517#comment-14291517 ]
Thomas Mueller commented on OAK-2389: ------------------------------------- > Can we please document what encode() is for? It's not for JSON, as far as > can tell. JsopBuilder.encode is documented ("Convert a string to a quoted Json literal using the correct escape sequences..."). It is for JSON, see usages of that method. I'm not sure if you mean another encode? > I don't see "case" in the source JsopBuilder.escape, line 289. > I prefer the constants as they generate both more readable and more compact > output I think the current "switch / case" actually generated more readable output than the array of substitution strings. But it is debatable, I understand. > a test case that checks for round-tripping of broken surrogate pairs OK, I will try that. > potentially a fix for the encode() method That's fixed now. It should be possible to write a test, I will try. > does our standard build cover the persistent cache? Not yet, I think. I opened OAK-2444 for that. > if we did a test on the JCR level, do we have a way to disable a memory cache > (which might make problems in the persisted value invisible?) Sorry I don't understand. Do you mean it prevents "real" tests for the persistent cache? Yes that's true. Probably testing with a small in-memory cache would help there. > issues with JsopBuilder.encode and .escape > ------------------------------------------ > > Key: OAK-2389 > URL: https://issues.apache.org/jira/browse/OAK-2389 > Project: Jackrabbit Oak > Issue Type: Bug > Reporter: Julian Reschke > Attachments: OAK-2389-b.diff, OAK-2389.diff > > > 1) escape() escapes many characters that do not need to be escaped (>127) > 2) encode() does not encode many control characters that would need to be > escaped when read through a JSON parser. -- This message was sent by Atlassian JIRA (v6.3.4#6332)