RFR - JDK-8202442 - String::unescape (Code Review)

[Jim Laskey]

revised webrev: http://cr.openjdk.java.net/~jlaskey/8202442/webrev-02/index.html

> On Sep 19, 2018, at 11:57 AM, Jim Laskey <james.laskey at oracle.com> wrote:
> 
> Will do.
> 
> 
>> On Sep 19, 2018, at 11:42 AM, Jonathan Gibbons <Jonathan.Gibbons at oracle.com> wrote:
>> 
>> If you're making the change, then you might add a test case based on Jan's suggestion, testing
>> 
>>    "\u005ct".equals(`\u005ct`.unescape())
>> 
>> -- Jon
>> 
>> 
>> On 9/19/18 7:40 AM, Jim Laskey wrote:
>>> I see and good point.  I think Jon is correct.  If you are using the JLS as the basis of your argument, then you better damn well stick to the JLS.
>>> 
>>> I’ll make the changes.
>>> 
>>> Cheers,
>>> 
>>> — Jim
>>> 
>>> 
>>>> On Sep 19, 2018, at 11:10 AM, Jan Lahoda <jan.lahoda at oracle.com> wrote:
>>>> 
>>>> I guess Jon's comment was that (per JLS) the outcome of unicode unescapes can then participate in the escape sequences in String literals. So, this:
>>>> "\u005ct"
>>>> 
>>>> is (as far as I know) a single character-literal (a tab), while it seems that
>>>> `\u005ct`.unescape()
>>>> 
>>>> is two characters:
>>>> \t
>>>> 
>>>> Not sure if that's an intent or not.
>>>> 
>>>> Jan
>>>> 
>>>> On 18.9.2018 20:55, Jim Laskey wrote:
>>>>> The intent, of course, is to offset the raw string literals non-translation of Unicode escapes and escape sequences. That is, have the multi-line cake and eat the escapes too.
>>>>> 
>>>>> So a developer could have
>>>>> 
>>>>> 	String s = `
>>>>>                   \t\tTitle
>>>>>                   \t\t\tbody
>>>>> 		   ...
>>>>> 
>>>>>                   `.align().escape();
>>>>> 
>>>>> to have tabs inserted in the string.
>>>>> 
>>>>> "\\" "\u005c\u005c" and `\` all translate to the same string. `\u005c` translates to "\\u005c”. `\u005c`.unescape() thustranslates to be the same as "\\”, "\u005c\u005c" and `\`.
>>>>> 
>>>>> Cheers,
>>>>> 
>>>>> — Jim
>>>>> 
>>>>> 
>>>>> 
>>>>>> On Sep 18, 2018, at 3:33 PM, Jonathan Gibbons <jonathan.gibbons at oracle.com> wrote:
>>>>>> 
>>>>>> Jim,
>>>>>> 
>>>>>> In JLS, and hence javac, Unicode escape handling happens early on at a low level, before string escape handling. This means that it is technically possible to write string escape sequences in terms of Unicode escapes.
>>>>>> 
>>>>>> I'm not suggesting you should do the same here, but you should be aware of the difference, compared to javac behavior.
>>>>>> 
>>>>>> -- Jon
>>>>>> 
>>>>>> 
>>>>>> On 9/18/18 10:51 AM, Jim Laskey wrote:
>>>>>>> Please review the code for String::unescape. Used to translate escape sequences in a string, typically in a raw string literal, into characters represented by those escapes.
>>>>>>> 
>>>>>>> webrev: http://cr.openjdk.java.net/~jlaskey/8202442/webrev/index.html
>>>>>>> jbs: https://bugs.openjdk.java.net/browse/JDK-8202442
>>>>>>> csr: https://bugs.openjdk.java.net/browse/JDK-8202443
>>>>>>> 
>>>>>>> Cheers,
>>>>>>> 
>>>>>>> — Jim
>>>>>>> 
>> 
>