Re: Feature idea: required version flag

[Garrett Cooper]

On Mon, Nov 3, 2008 at 9:22 PM, Stephan Beal <address@hidden> wrote:
> On Mon, Nov 3, 2008 at 4:24 PM, Greg Chicares <address@hidden> wrote:
>> On 2008-11-03 13:46Z, Stephan Beal wrote:
>>>
>>> i know we've discussed the topic of "how to check if the user is
>>> running Make version XYZ or higher" before, but i just had another
>>> idea for helping to solve this...
>>
>> What's wrong with the solutions already discussed?
>
> They're quite verbose.
>
>> If you don't want to write the test in a makefile, then run
>> `make --version` and parse the output; you could write a wrapper
>> script to do that.
>
> The output isn't in a uniform format, so the parsing isn't as trivial.
> A wrapper script would be platform-specific,
>
>>> In this invokation, make would refuse to run if its version is less
>>> than 3.80. If its an older make, it would probably puke on the unknown
>>> -R flag and die anyway, so we're protected either way.
>>
>> How would you distinguish that mode of failure from other modes?
>
> i see no need to differentiate. ?

I definitely see a benefit to a compatibility requirement, or at least
some degree of error'ing out if a certain feature doesn't exist within
Make.

One issue I've dealt with in the past, which may be confusing to some
folks is the use of $(abspath ..). The errors returned through the use
of $(abspath ..) prior to v3.81 are confusing.

-Garrett