Migration Down Fails in 3.3.2 if you don't have have a return char on sql statement.

[h3adache]

CREATE TABLE test.[Migration Failure] (
  ID BIGINT IDENTITY PRIMARY KEY,
  [Test] VARCHAR(255) NOT NULL
);
-- //@UNDO
DROP TABLE test.[Migration Failure];

up + down successfully completes but down is an empty statement.

CREATE TABLE test.[Migration Failure] (
  ID BIGINT IDENTITY PRIMARY KEY,
  [Test] VARCHAR(255) NOT NULL
);
-- //@UNDO
DROP TABLE test.[Migration Failure];

up + down successfully correctly. The only difference is adding a return after the DROP statement.
Probably relates to #101 but need to investigate more.

[harawata]

Thanks @h3adache ,

I thought I added a test for this, but apparently not.
Line 98 must be changed.

-        if (lineBuffer.length() > 0 && !undo && !inUndo) {
+        if (lineBuffer.length() > 0 && (!undo || inUndo)) {

This seems like a bad regression. Should we release 3.3.3 immediately?

[h3adache]

Nice, quick find! Yea, we should do a quick release for this one. Thanks!

[harawata]

3.3.3 is out. Thank you, @h3adache !

[h3adache]

Hi @harawata this bug is still there but on the down stage. It doesn't replace ${changelog} if it's not followed by a return character :P
I can look into it if you don't have time.

[harawata]

@h3adache , XD
Sorry, I should've waited for your verification. It's fixed in the snapshot.

[h3adache]

@harawata thanks! I tried the 3.3.4-SNAPSHOT and it looks great!

[harawata]

Thank you, @h3adache !
Let's wait for a while before releasing 3.3.4 ;D

[h3adache]

lol ok I'll keep testing 3.3.4-snapshot but we should release soon so people can take advantage of the new options and your performance enhancements without snapshot

[harawata]

Hi @h3adache ,

I'll release 3.3.4 this weekend.
Please let me know if you've found something else.