Hi,
I know what the problem is. readLong will first try to convert
9223372036854775808
to a long (that doesn't work), and then negate it (that would work, but
it's too late). One solution is to first negate it, and then convert to a
long. See below for a possible patch (not tested).
> It seems any negative number causes the same problem. .. CREATE TABLE
connection (id BIGINT AUTO_INCREMENT(-1, 1) PRIMARY KEY)
I think that's not the problem. Your statement (with -1) works for me.
Patch (it also renames getInt to readInt):
### Eclipse Workspace Patch 1.0
#P h2
Index: src/main/org/h2/command/Parser.java
===================================================================
--- src/main/org/h2/command/Parser.java (revision 5847)
+++ src/main/org/h2/command/Parser.java (working copy)
@@ -536,7 +536,7 @@
private Prepared parseAnalyze() {
Analyze command = new Analyze(session);
if (readIf("SAMPLE_SIZE")) {
- command.setTop(getPositiveInt());
+ command.setTop(readPositiveInt());
}
return command;
}
@@ -2966,15 +2966,15 @@
return function;
}
- private int getPositiveInt() {
- int v = getInt();
+ private int readPositiveInt() {
+ int v = readInt();
if (v < 0) {
throw DbException.getInvalidValueException("positive integer",
v);
}
return v;
}
- private int getInt() {
+ private int readInt() {
boolean minus = false;
if (currentTokenType == MINUS) {
minus = true;
@@ -2982,12 +2982,16 @@
} else if (currentTokenType == PLUS) {
read();
}
- if (currentTokenType != VALUE || currentValue.getType() !=
Value.INT) {
+ if (currentTokenType != VALUE) {
throw DbException.getSyntaxError(sqlCommand, parseIndex,
"integer");
}
+ if (minus) {
+ // must do that now, otherwise Integer.MIN_VALUE wouldn't work
+ currentValue = currentValue.negate();
+ }
int i = currentValue.getInt();
read();
- return minus ? -i : i;
+ return i;
}
private long readLong() {
@@ -2995,14 +2999,19 @@
if (currentTokenType == MINUS) {
minus = true;
read();
+ } else if (currentTokenType == PLUS) {
+ read();
}
- if (currentTokenType != VALUE ||
- (currentValue.getType() != Value.INT &&
currentValue.getType() != Value.LONG)) {
+ if (currentTokenType != VALUE) {
throw DbException.getSyntaxError(sqlCommand, parseIndex,
"long");
}
+ if (minus) {
+ // must do that now, otherwise Long.MIN_VALUE wouldn't work
+ currentValue = currentValue.negate();
+ }
long i = currentValue.getLong();
read();
- return minus ? -i : i;
+ return i;
}
private boolean readBooleanSetting() {
@@ -3901,7 +3910,7 @@
column.setSequence(sequence);
}
if (readIf("SELECTIVITY")) {
- int value = getPositiveInt();
+ int value = readPositiveInt();
column.setSelectivity(value);
}
String comment = readCommentIf();
@@ -4005,7 +4014,7 @@
readIf("CHAR");
if (dataType.supportsScale) {
if (readIf(",")) {
- scale = getInt();
+ scale = readInt();
original += ", " + scale;
} else {
// special case: TIMESTAMP(5) actually means
@@ -4027,7 +4036,7 @@
} else if (readIf("(")) {
// Support for MySQL: INT(11), MEDIUMINT(8) and so on.
// Just ignore the precision.
- getPositiveInt();
+ readPositiveInt();
read(")");
}
if (readIf("FOR")) {
@@ -4532,7 +4541,7 @@
command.setRowBased(false);
}
if (readIf("QUEUE")) {
- command.setQueueSize(getPositiveInt());
+ command.setQueueSize(readPositiveInt());
}
command.setNoWait(readIf("NOWAIT"));
read("CALL");
@@ -4940,7 +4949,7 @@
} else if (readIf("NUMBERS")) {
command.setInt(Constants.ALLOW_LITERALS_NUMBERS);
} else {
- command.setInt(getPositiveInt());
+ command.setInt(readPositiveInt());
}
return command;
} else if (readIf("DEFAULT_TABLE_TYPE")) {
@@ -4951,7 +4960,7 @@
} else if (readIf("CACHED")) {
command.setInt(Table.TYPE_CACHED);
} else {
- command.setInt(getPositiveInt());
+ command.setInt(readPositiveInt());
}
return command;
} else if (readIf("CREATE")) {
Regards,
Thomas
On Sat, Aug 30, 2014 at 1:31 PM, Steve McLeod <[email protected]>
wrote:
> It is not your use of Long.MIN_VALUE that is a problem. It seems any
> negative number causes the same problem. For instance, this fails with a
> syntax exception:
>
> CREATE TABLE connection (id BIGINT AUTO_INCREMENT(-1, 1) PRIMARY KEY);
>
>
> On Saturday, 30 August 2014 08:21:54 UTC+2, Gili wrote:
>>
>> Noel,
>>
>> Both the H2 documentation and java.lang.Long.MIN_VALUE indicate that
>> -9223372036854775808 is a legal value, yet H2 rejects it. Isn't this a bug?
>>
>> Gili
>>
>> On Saturday, August 30, 2014 1:36:15 AM UTC-4, Noel Grandin wrote:
>>>
>>> Our internal implementation of sequences uses longs for everything, so
>>> you will need to use values for start/stop/cycle/increment that fit
>>> into a long value.
>>>
>>> On Fri, Aug 29, 2014 at 11:04 PM, Gili <[email protected]> wrote:
>>> > Two clarifications:
>>> >
>>> > I am using version 1.3.176.
>>> > This behavior contradicts the documentation:
>>> > http://www.h2database.com/html/datatypes.html#bigint_type
>>> >
>>> > Please also take this opportunity to clarify the difference between
>>> IDENTITY
>>> > and BIGINT in the documentation. As far as I understand it, IDENTITY
>>> is
>>> > equivalent to "BIGINT AUTO_INCREMENT PRIMARY KEY" but it's not clear
>>> if
>>> > there are any other differences. For example, the documentation reads
>>> "Used
>>> > values are never re-used, even when the transaction is rolled back"
>>> for
>>> > IDENTITY but not for BIGINT AUTO_INCREMENT so it's not clear if there
>>> is
>>> > some magic under the hood.
>>> >
>>> > Gili
>>> >
>>> > On Friday, August 29, 2014 4:53:16 PM UTC-4, Gili wrote:
>>> >>
>>> >> Hi,
>>> >>
>>> >> If I run:
>>> >>
>>> >> CREATE TABLE connection (id BIGINT AUTO_INCREMENT(-9223372036854775808,
>>>
>>> >> 1) PRIMARY KEY);
>>> >>
>>> >> I get:
>>> >>
>>> >> Syntax error in SQL statement "CREATE TABLE CONNECTION (ID BIGINT
>>> >> AUTO_INCREMENT(-9223372036854775808[*], 1) PRIMARY KEY)"; expected
>>> "long"
>>> >>
>>> >> I've successfully initialized all other types at their minimum
>>> values.
>>> >> BIGINT seems to be the only type that has a problem with this.
>>> >>
>>> >> Gili
>>> >
>>> > --
>>> > You received this message because you are subscribed to the Google
>>> Groups
>>> > "H2 Database" group.
>>> > To unsubscribe from this group and stop receiving emails from it, send
>>> an
>>> > email to [email protected].
>>> > To post to this group, send email to [email protected].
>>> > Visit this group at http://groups.google.com/group/h2-database.
>>> > For more options, visit https://groups.google.com/d/optout.
>>>
>> --
> You received this message because you are subscribed to the Google Groups
> "H2 Database" group.
> To unsubscribe from this group and stop receiving emails from it, send an
> email to [email protected].
> To post to this group, send email to [email protected].
> Visit this group at http://groups.google.com/group/h2-database.
> For more options, visit https://groups.google.com/d/optout.
>
--
You received this message because you are subscribed to the Google Groups "H2
Database" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to [email protected].
To post to this group, send email to [email protected].
Visit this group at http://groups.google.com/group/h2-database.
For more options, visit https://groups.google.com/d/optout.
