'_' should not be used as an identifier, since it is a reserved keyword from source level 1.8 on

[titanseason]

When I use DbSqlBuilder._() method, Eclipse (J2EE) tells me '_' should not be used as an identifier, since it is a reserved keyword from source level 1.8 on.

Any plan to change this method name?

[igr]

yes! :))) I just try to find small english word that would serve as concatenation... any ideas?

• ql()
• sql()
• add()

I believe sql make most sense as it is raw sql. wdyt?

[neroux]

Out of that list I'd vote for add().

However (and there is always a but, isnt there ;-) ) _() is a convenience method for append() and with two very similarly named methods (add/append) doing the identical thing it might become eventually confusing/redundant. As those aliases/shortcuts should usually be as short a possible, the dollar sign (as used in Jerry too) might be a good solution.

Alternatively - and not without potential drawbacks - a Unicode character such as fullwidth low line (＿) might be an option (Java unfortunately does not accept fullwidth plus (＋)). Of course that could create typing - though autocomplete might come to rescue here - and display issues, apart from confused users. I wanted to have it mentioned though :).

[igr]

Good catch @neroux , I forgot there is append which does the same :) In that sense, having another shorter-named method that does the same is... meh :) Then maybe We can simply have one method and not the shortcut (i.e. append -> add).

Wow, unicode character!!! This is mindblowing @neroux!!! Reminds me to:

double Δt = lastPollTime - pollTime;
double α = 1 - Math.exp(-Δt / τ);
average += α * (x - average);

I took _ because its visually small and this append/add is used a lot, so its much cleaner to read. But you are absolutely right; the shortcut must be as short as possible, hence proposal is:

1. Rename append to add (its shorter)
2. Use alternative shortcut, one of:

   int _, ‿, ⁀, ⁔, ︳, ︴, ﹍, ﹎, ﹏, ＿;

Moreover: the following

for (int i = Character.MIN_CODE_POINT; i <= Character.MAX_CODE_POINT; i++)
    if (Character.isJavaIdentifierStart(i) && !Character.isAlphabetic(i))
        System.out.print((char) i + " ");
}

prints

$ _ ¢ £ ¤ ¥ ؋ ৲ ৳ ৻ ૱ ௹ ฿ ៛ ‿ ⁀ ⁔ ₠ ₡ ₢ ₣ ₤ ₥ ₦ ₧ ₨ ₩ ₪ ₫ € ₭ ₮ ₯ ₰ ₱ ₲ ₳ ₴ ₵ ₶ ₷ ₸ ₹ ꠸ ﷼ ︳ ︴ ﹍ ﹎ ﹏ ﹩ ＄ ＿ ￠ ￡ ￥ ￦

[igr]

Now thinking, _ will be probably used in some future version of Java, similar like it is used e.g. in Go language; so having anything similar might not be good idea....

[neroux]

At the moment they do not seem to have big plans for the underscore -> http://git.net/ml/lambda-development-java/2013-07/msg00350.html however yes, the fullwidth low line is visually almost identical to the underscore and that might raise some confusion.

What would you think about $? Its short (even by actual byte count), a popular alias, already used by Jerry, and according to http://mail.openjdk.java.net/pipermail/lambda-dev/2013-August/010676.html safe to use (until they change their mind ;) ).

[igr]

The issue with $ might be that it is used inside of these sql templates a lot. So you might have something like:

...$(" and $table.id=")...

which might be kind of confusion as well, right?

[neroux]

I understand what you mean. I guess in some cases it could confuse some people but generally I'd assume it should be clear what is a command and what data, particularly with syntax highlighting.

My main concern with Unicode characters is about their "typeability" (though that might mitigated by aforementioned autocomplete) and possibly encoding issues.

[igr]

hey man, you are right! in fact, i just realized that example i gave above can't happens. so $ might be nice shortcut.

The template is where $ appears, and those two can't never happen together.

For Jerry it ways easy, since this is what JQuery is using ;) Cool!

[neroux]

Just in time :). With Java 1.9 (currently due in July) _ does not yield a warning anymore but straightforward an error

https://blogs.oracle.com/sundararajan/entry/underscore_is_a_keyword_in

[igr]

Awesome!!!