Parsing of scripts containing the syllables "end" or "begin" is broken

[etluchs]

A script containing the syllable "end" (or, I suppose "begin") outside a comment or a quote will prevent the code from finding any further line-breaks, thus, usually breaking the script.
the offending line is

testcontainers-java/modules/database-commons/src/main/java/org/testcontainers/ext/ScriptUtils.java

Line 168 in b6de1ac

if (!inLiteral && !inComment && containsSubstringAtOffset(script, "END", i)) {

An example script would be

create tablespace DEV datafile 'dev.dbf' size 256M autoextend on;
CREATE USER bla IDENTIFIED BY testuser DEFAULT TABLESPACE DEV;
CREATE USER blub IDENTIFIED BY testuser DEFAULT TABLESPACE DEV;
CREATE USER bloh IDENTIFIED BY testuser DEFAULT TABLESPACE DEV;

[kiview]

Thanks for spotting this @etluchs.
Can you already propose a better solution?

[etluchs]

Hm, one could hack it even more by only considering begin/end when found as whole words surrounded by whitespace equivalent.
But I think, it might be worth biting the bullet and use some simple parser (generated with antlr or something).

[kalexmills]

This is affecting my project, since I'm trying to produce a schema which contains the word end in one of its columns... I'm moving back to 1.10.3 until this is fixed.

EDIT: Confirmed that using test containers at 1.10.3 fixes the issue for me. Others might use the same workaround.

[rnorth]

For now please work around by using the older version. This issue annoys me and is top of my list of bugs to fix. Apologies.

_{Sent with GitHawk}

[rrhartjr]

Any updates on this? Looks like any substring of end causes a failure.. including 'gender' :(

[ivankovbas]

I've just checked on version 1.12.2 and it seems the issue is already fixed, so it could be closed. @rnorth isn't it?

[rrhartjr]

I confirmed it fails on 1.11.4 and works on 1.12.2, we have upgraded our dependency to 1.12.2, so I'm satisfied :)