Name | Modified | Size | Downloads / Week |
---|---|---|---|
Parent folder | |||
README.txt | 2014-02-14 | 1.4 kB | |
dbvolution-0.9.8.pom | 2014-02-12 | 9.2 kB | |
dbvolution-0.9.8.jar | 2014-02-12 | 374.5 kB | |
dbvolution-0.9.8-sources.jar | 2014-02-12 | 244.6 kB | |
dbvolution-0.9.8-javadoc.jar | 2014-02-12 | 1.1 MB | |
Totals: 5 Items | 1.7 MB | 0 |
0.9.8 is out and a good thing it is too. The permittedPattern() method of QueryableDatatype has managed to get thru innumerable versions without a problem. But strangely didn't function correctly within my JSP page at work. After much angst and worry I managed to find and fix the bug in a, supposedly, core part of DBV. For at least 70 revisions there has been a bug in QueryableDatatype where the permitted pattern was set to the QDT itself and not the actual pattern. Obviously this is wrong and any amount of testing would have spotted it. But QDT.permittedPattern() wasn't tested. It seems that my early qualms about including pattern matching that WASN'T Java regular expressions prevented me from writing tests. I planned to fix that at some point by writing a miracle procedure that made Java reg-exp just work on all the databases. As time went by I got involved in the other more immediately important tasks and ideologically pure tasks like Java Reg-exp and Full Outer Joins on all databases got put on the rainy day shelf. Now that is OK for full outer joins as they're hard to do and rarely used. But the LIKE comparison is a big part of SQL. I really should have had tests that automatically indicated that a big function of DBV was broken. So, long story short, permittedPattern() was broken, it's now fixed, and there are tests to keep it fixed.