[PATCH] Teach llvm::format_hex() to support formatting without a prefix '0x'

[Sean Silva]

On Sat, Jan 24, 2015 at 5:09 PM, Zachary Turner <zturner at google.com> wrote:

> Fwiw, the default value of 4th arg is true already, so it's unnecessary.
> As of my latest patch i moved no prefix mode to a separate function, so
> printing with/without prefix is now
>
> OS << format_hex(N, 16) // 0x12345
> vs.
> OS << format_hex_no_prefix(N, 16) // 12345
>
> It does seem a bit clearer if we change it, but i was worried that it
> would break external users in ways that would only be visible at runtime,
> which is much worse than breaking them at compile time


If that is a concern, you can change the name to formatHex (no clue why
format_hex was chosen instead, as that would comply with the coding
standard better). Comments in raw_ostream.h even refers to it as formatHex.

-- Sean Silva


>
> On Sat, Jan 24, 2015 at 8:17 AM Sean Silva <chisophugis at gmail.com> wrote:
>
>> On Fri, Jan 23, 2015 at 10:00 PM, Chandler Carruth <chandlerc at google.com>
>> wrote:
>>
>>>
>>> On Fri, Jan 23, 2015 at 1:58 PM, Rui Ueyama <ruiu at google.com> wrote:
>>>
>>>> Why does format_hex prints out "0x" prefix in the first place? If it
>>>> doesn't, like printf %x, we don't need an extra parameter or a separate
>>>> function. It may be too late, though.
>>>
>>>
>>> I would be fine to change all users to just write '0x' into the stream
>>> first.
>>>
>>
>> +1. Compare:
>>
>> OS << format_hex(N, 16, false, true);
>> vs.
>> OS << "0x" << format_hex(N, 16);
>>
>> -- Sean Silva
>>
>>
>>>
>>> _______________________________________________
>>> llvm-commits mailing list
>>> llvm-commits at cs.uiuc.edu
>>> http://lists.cs.uiuc.edu/mailman/listinfo/llvm-commits
>>>
>>> _______________________________________________
>> llvm-commits mailing list
>> llvm-commits at cs.uiuc.edu
>> http://lists.cs.uiuc.edu/mailman/listinfo/llvm-commits
>>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20150124/33b5b99b/attachment.html>