[
https://issues.apache.org/jira/browse/CASSANDRA-8638?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14289940#comment-14289940
]
Abhishek Gupta commented on CASSANDRA-8638:
-------------------------------------------
As an alternate solution, here is what I propose to do as opposed to using a
replace:
1. check the first 4 bytes and see if contains a BOM
2. if it contains a BOM, i will lookup the bom marker and further get the
encoding and bom size(from a static dictionary)
3. open the file in the encoding returned in step-2
4. move file pointer ahead by bom size(based on bom size returned in step-2)
5. everything else will remain same
> CQLSH -f option should ignore BOM in files
> ------------------------------------------
>
> Key: CASSANDRA-8638
> URL: https://issues.apache.org/jira/browse/CASSANDRA-8638
> Project: Cassandra
> Issue Type: Improvement
> Components: Tools
> Environment: Red Hat linux
> Reporter: Sotirios Delimanolis
> Priority: Trivial
> Labels: cqlsh, lhf
> Fix For: 2.1.3
>
> Attachments: 0001-bug-CASSANDRA-8638.patch
>
>
> I fell in byte order mark trap trying to execute a CQL script through CQLSH.
> The file contained the simple (plus BOM)
> {noformat}
> CREATE KEYSPACE IF NOT EXISTS xobni WITH replication = {'class':
> 'SimpleStrategy', 'replication_factor': '3'} AND durable_writes = true;
> -- and another "CREATE TABLE bucket_flags" query
> {noformat}
> I executed the script
> {noformat}
> [~]$ cqlsh --file /home/selimanolis/Schema/patches/setup.cql
> /home/selimanolis/Schema/patches/setup.cql:2:Invalid syntax at char 1
> /home/selimanolis/Schema/patches/setup.cql:2: CREATE KEYSPACE IF NOT EXISTS
> test WITH replication = {'class': 'SimpleStrategy', 'replication_factor':
> '3'} AND durable_writes = true;
> /home/selimanolis/Schema/patches/setup.cql:2: ^
> /home/selimanolis/Schema/patches/setup.cql:22:ConfigurationException:
> <ErrorMessage code=2300 [Query invalid because of configuration issue]
> message="Cannot add column family 'bucket_flags' to non existing keyspace
> 'test'.">
> {noformat}
> I realized much later that the file had a BOM which was seemingly screwing
> with how CQLSH parsed the file.
> It would be nice to have CQLSH ignore the BOM when processing files.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
