Improve error reporting: provide better literal suffix description

[smoothdeveloper]

What

let foo = 1up

This is not a valid numeric literal. Sample formats include 4, 0x4, 0b0100, 4L, 4UL, 4u, 4s, 4us, 4y, 4uy, 4.0, 4.0f, 4I

Why

It is difficult to remember:

• the valid suffixes
• what exact type they correspond to (despite they are kind of logical, if I made a mistake, better remind me the type as well)
• I was looking for decimal but it isn't there

How

Bonus point if we instead of 4 we show 42 or the actual value.

This is not a valid numeric literal. Valid numeric literals include 1, 0x1, 0b0001 (int), 1u (uint32), 1L (int64), 1UL (uint64), 1s (int16), 1y (sbyte), 1ui (byte), 1.0 (float), 1.0f (float32), 1.0m (decimal), 1I (BigInteger)
#1103

[smoothdeveloper]

note also that let a = 0m works but let a = 0f doesn't which I think is inconsistent.

[smoothdeveloper]

Question: why do UL and L need to be uppercased while others are all lowercase only?

[vasily-kirichenko]

Question: why do UL and L need to be uppercased while others are all lowercase only?

About l vs L, my guess is that's something like 21l is very hard to read. I'm not sure about u vs U though.

[smoothdeveloper]

@vasily-kirichenko I see, what about the one for BigInteger? 21i vs 21I I guess Big Integer = Big I :)

[forki]

We should not question the syntax here. That is pretty much fixed. This
should only be about the error messages
On Apr 28, 2016 12:24 PM, "Gauthier Segay" notifications@github.com wrote:

@vasily-kirichenko https://github.com/vasily-kirichenko I see, what
about the one for BigInteger? 21i vs 21I I guess Big Integer = Big I :)

—
You are receiving this because you are subscribed to this thread.
Reply to this email directly or view it on GitHub
#1129 (comment)

[smoothdeveloper]

@forki: do you think the proposed error message (check on github itself due to edits) is an enhancement?

[dsyme]

Adding the samples is ok, but as a matter of principle please use sentences, rather than phrases:

This is not a valid numeric literal. Valid numeric literals include 1, 0x1, 0b0001 (int), 1u (uint32), 1L (int64), 1UL (uint64), 1s (int16), 1y (sbyte), 1ui (byte), 1.0 (float), 1.0f (float32), 1.0m (decimal), 1I (BigInteger)

It's important to say "include" because the numeric literals are extensible through the QZRING feature.

[smoothdeveloper]

@dsyme updated, thanks for feedback, maybe you should consider putting some guidelines on #1103 regarding how to word those messages.

I still think the final messages should be created/edited by MS with concern at consistency, comprehensibility and conciseness.

[forki]

PR at #1148
Regarding "bonus points": we failed to parse the numeral so we can't say what the "actual value" is here.