[ https://issues.apache.org/jira/browse/CASSANDRA-17617?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17553509#comment-17553509 ]
Berenguer Blasi commented on CASSANDRA-17617: --------------------------------------------- bq. I added your changes to my PR. If you're happy with that you can just move this one on if you please. Thanks [~tanujnay] yes I noticed that. But I have to put up PRs anyway for all maintained versions, add CI and then another committer/reviewer to +1 that, which is the table I just posted. Then I'll merge all that with proper attribution ofc! :-) > CQLSH unicode control character list is too liberal > --------------------------------------------------- > > Key: CASSANDRA-17617 > URL: https://issues.apache.org/jira/browse/CASSANDRA-17617 > Project: Cassandra > Issue Type: Improvement > Components: CQL/Interpreter > Reporter: Tanuj Nayak > Assignee: Tanuj Nayak > Priority: Normal > Fix For: 3.11.x, 4.0.x, 4.1-rc > > > It appears that the list of escaped unicode control characters > [here|https://github.com/apache/cassandra/blob/53a67ff2c36d90d337aba1409498de29931d4279/pylib/cqlshlib/formatting.py#L32] > is a bit too liberal. It seems to include characters such as '1' (0x31) and > '0' (0x30) which do not need to be escaped. It seems that the actual range > should be 0x00 - 0x1F and 0x7F+ as corroborated [by this > page|https://en.wikipedia.org/wiki/Unicode_control_characters]. > > This causes unnecessary escaping and regex substitutions on the CQLSH end > whenever common characters such as any punctuation or a 0 or a 1 appear in > the text column of a table. One might notice that a table with a text column > filled with 2's will take much less time to print than one with all 0's for > this reason. -- This message was sent by Atlassian Jira (v8.20.7#820007) --------------------------------------------------------------------- To unsubscribe, e-mail: commits-unsubscr...@cassandra.apache.org For additional commands, e-mail: commits-h...@cassandra.apache.org