Re: [PATCH 0/4] yacc.c: provide the Bison version as an integral macro

[Akim Demaille]

Hi Uxio,

> Le 19 nov. 2020 à 18:55, uxio prego <uxio.prego@gmail.com> a écrit :
> 
> Hi,
> 
>> On 11 Nov 2020, at 09:05, Akim Demaille <akim.demaille@gmail.com> wrote:
>> 
>> I used a scale of 100, which should not be a problem.  We could use a
>> scale of 1000
> 
> Do you still use a count starting in 90 for release candidates?

Yes, I usually do.

> How safe is a limit of ten release candidates?

Given the history, I think it is safe enough: 

$ git tag -l | fgrep .9
v2.6.90
v2.7.90
v2.7.91
v3.1.90
v3.1.91
v3.2.90
v3.2.91
v3.3.90
v3.3.91
v3.4.90
v3.4.91
v3.4.92
v3.5.90
v3.5.91
v3.5.92
v3.5.93
v3.5.94
v3.6.90
v3.6.91
v3.6.92
v3.6.93


>> , or even something not regular as Boost does (107400 is
>> 1.74).
> 
> Hm.
> 
> The obvious: there is no ‘1.74’, there is ‘1.74.0’.

Yeah, ok.

> The sneaky: there is a ‘1.65.1’ listed.
> 
> https://github.com/boostorg/config/commit/4239074552cb233c09f344d5b76c80b8c22f22c8

Which is indeed `#define BOOST_VERSION 106501`.  I'm not sure I saw
what you wanted me to see here.

Cheers!