Relax whitespace handling in SG_ line parser (Vector CANdb++ accepts tighter formatting)

[ChrisFromNI]

Hi,
some DBC files generated or edited outside CANdb++ contain SG_ signal definitions with no or reduced whitespace between tokens. While this still follows the Vector DBC grammar, DbcParser currently expects strict spacing and fails to parse these lines.
Example of a minimally‑spaced but valid SG_ line (accepted by CANdb++):
SG_ SignalName optionalMultiplexerIdentifier :0|16@1+(0.1,0)[0|100]"rpm"Vector__XXX

Request:
Please consider relaxing the SG_ line parser to allow token adjacency where unambiguous (e.g., no required spaces around :, |, @, +/-, ( ), [ ], units, and receiver lists). This would improve compatibility with real‑world DBC files encountered in industry.

Thanks for maintaining this excellent library!

[Uight]

Parsing currenlty is regex based; some currenlty requried spaces (/s+) can safely be replaced with (/s*) but for the UnitGroup and the ReceiverGroup this can cause an overmathcing error;

However the following regex (generated by AI) should work:

private readonly string m_signalRegex = $@"\s*SG_\s+(?<{NameGroup}>[\w]+)\s*(?<{MultiplexerGroup}>[Mm\d]*)\s*:\s*(?<{StartBitGroup}>\d+)\|(?<{SizeGroup}>\d+)@(?<{ByteOrderGroup}>[01])" +
                                $@"(?<{ValueTypeGroup}>[+-])\s*\((?<{FactorGroup}>[\d\+\-eE.]+),(?<{OffsetGroup}>[\d\+\-eE.]+)\)\s*\[(?<{MinGroup}>[\d\+\-eE.]+)\|(?<{MaxGroup}>[\d\+\-eE.]+)\]" +
                                $@"\s*""(?<{UnitGroup}>[^""]*)""\s*(?<{ReceiverGroup}>[\w,]+(?:\s+[\w,]+)*)";

tested with this using test

        [TestCase("SG_ SignalName:0|16@1+(0.1,0)[0|100]\"rpm\"Vector__XXX")]
        [TestCase("SG_ SignalName m0:0|16@1+(0.1,0)[0|100]\"rpm\"Vector__XXX")]
        public void SignalSyntaxWithMinimalWhitespacing(string line)
        {
            var dbcBuilderMock = m_repository.Create<IDbcBuilder>();

            dbcBuilderMock.Setup(mock => mock.AddSignal(It.IsAny<Signal>()))
                .Callback<Signal>(signal =>
                {
                    Assert.Multiple(() =>
                    {
                        Assert.That(signal.Name, Is.EqualTo("SignalName"));
                        Assert.That(signal.StartBit, Is.EqualTo(0));
                        Assert.That(signal.Length, Is.EqualTo(16));
                        Assert.That(signal.ByteOrder, Is.EqualTo(1));
                        Assert.That(signal.ValueType, Is.EqualTo(DbcValueType.Unsigned));
                        Assert.That(signal.Factor, Is.EqualTo(1E-1));
                        Assert.That(signal.Offset, Is.EqualTo(0));
                        Assert.That(signal.Minimum, Is.EqualTo(0));
                        Assert.That(signal.Maximum, Is.EqualTo(100));
                        Assert.That(signal.Unit, Is.EqualTo("rpm"));
                        Assert.That(signal.Receiver.FirstOrDefault(), Is.EqualTo("Vector__XXX"));
                    });
                });

            var signalLineParser = CreateParser();
            var nextLineProviderMock = m_repository.Create<INextLineProvider>();

            Assert.That(signalLineParser.TryParse(line, dbcBuilderMock.Object, nextLineProviderMock.Object), Is.True);
        }

However the AI always tries to keep the old behavior which is why it changed the receiver group parsing tp avoid parsing this testcase:
[TestCase("SG_ qGearboxOilMin : 0|16@1+ (0.1,0) [0|6553.5] "l/min" "NATEC"")]

However it also parses this successfully:

@" SG_ MCU_longitude m7 : 28|29@1- (1E-006,0) [-10|35.6] ""deg"" NEO WHEEL TOP"

It parses that with three receivers;

And actually the old parsing also parses that which is probably wrong because in the spec it says the , is requried as far as i get it

However for that i ddint find a quick fix as the parsn gthen always stops at the first receiver leaving left over parts and i wasnt sure if i should fail aprsing if not the full line is matched