[Bug] MSSQL support bugs

Hello!

I'm using HeidiSQL with MS SQL Server. Here I will report bugs, I've found.

First one: invalid generated alter syntax.
Explanation:

CREATE TABLE "test" (
"col1" INT NOT NULL,
"col2" VARCHAR(4) NULL,
PRIMARY KEY ("col1")
);
// I want to make this table m2m link: I add col2 to the PK via GUI
// These are generated statements:
ALTER TABLE ."test"
ALTER COLUMN "col2" VARCHAR(4) NOT NULL;
ALTER TABLE ."test"
DROP PRIMARY KEY,
ADD PRIMARY KEY ("col1", "col2", ""); // "" -- WTF?
// SQL Server reports syntax errors on them
// How should it look like:
ALTER TABLE ."test"
ALTER "col2" VARCHAR(4) NOT NULL;
ALTER TABLE ."test"
DROP CONSTRAINT "PK_test",
ADD CONSTRAINT "PK_test" PRIMARY KEY ("col1", "col2");

Sorry, lost COLUMN

// How should it look like:
ALTER TABLE ."test"
ALTER COLUMN "col2" VARCHAR(4) NOT NULL;

BUG: Invalid column default value in create statement.

CREATE TABLE "accounts" (
"owner" INT NULL DEFAULT CURRENT_TIMESTAMP,
"code" VARCHAR(64) NULL DEFAULT CURRENT_TIMESTAMP(64),
"name" NVARCHAR(128) NULL DEFAULT CURRENT_TIMESTAMP(128),
"version" TIMESTAMP NULL DEFAULT NULL
);

All columns are nullable:

SELECT COLUMN_NAME,COLUMN_DEFAULT FROM INFORMATION_SCHEMA.COLUMNS WHERE TABLE_NAME = 'accounts';
COLUMN_NAME,COLUMN_DEFAULT
owner       (null)
code        (null)
name        (null)
version     (NULL)

BUG: default constraint not dropped before "drop column" from gui.

Trying to drop column, part of log after pressing "save" button:

ALTER TABLE ."accounts"
DROP COLUMN "is_read";
/* Ошибка SQL (4922): The object 'DF__accounts__is_rea__1975C517' is dependent on column 'is_read'
ALTER TABLE DROP COLUMN is_read failed because one or more objects access this column. */

I get the idea the developer is ignoring all MS SQL bugs.

Well, the Subversion log is publicly available and it contains several revisions regarding SQL Server bug fixes. Of course, you can never get those lazy open source programmers work for you when you need it ;-)

Come on, relax. I understand the non-MySQL engines support is still pretty rough but that's what experimental means.

There is 1 bad thing in this project: bizarre development tools used.

I could make patches for better mssql support, but I won't be able to test if my changes even compile.

Well, the Subversion log is publicly available and it contains several revisions regarding SQL Server bug fixes. Of course, you can never get those lazy open source programmers work for you when you need it ;-)

Come on, relax. I understand the non-MySQL engines support is still pretty rough but that's what experimental means.

Yes, I should relax. I seemed a bit uptight there. I don't expect bugs to be fixed immediately. I'm involved with enough open source projects myself to know that. I just expect some kind of reply or response on a bug report.

Honestly, I feel your pain. I have myself reported fairly obvious and 100% reproducible bugs in several open source projects and response time normally ranges from three weeks to never. And I'm not talking about bugfixing, just a simple acknowledgement. Funny thing is that project team size does not seem to matter. (In this case, though, it's basically a one man project and I am not that man.)