On Dec 5, 2009, at 12:54 AM, William Stein wrote:

> 2009/12/4 Robert Bradshaw <rober...@math.washington.edu>:
>> On Dec 4, 2009, at 2:27 PM, Dr. David Kirkby wrote:
>>
>>> It would appear when creating a binary for Sage, one would normally
>>> put
>>> something like
>>>
>>> sage -bdist 4.1.2-Solaris
>>>
>>> or similar to create a binary for Solaris of version 4.1.2 of Sage.
>>>
>>> It would be relatively easy to get Sage to report the version
>>> number, so one
>>> does not have to specify that. I'd suggest replacing the
>>>
>>> sage -c "print Sage Works"
>>>
>>> with something that prints the version number, then use that in the
>>> script.
>>>
>>> Not only does it save typing 5 characters, but it is more foolproof,
>>> as one
>>> can't accidentally create a binary with the wrong version number.
>>>
>>> Comments?
>>
>> +1
>
> -1
>
> One drawback is that it makes the input to sdist inconsistent with the
> input to bdist, e.g.,
>
> sage -sdist 4.1.2-solaris
> sage -bdist solaris
>
> The second drawback is that one might want to make a binary
> distribution, and not label it with a version number, or label it with
> a version that properly reflects what you've added to it.  E.g.,:
>
>  (a) I download sage-4.1.2.
>  (b) I change a bunch of crap.
>  (c) I want to make a binary to distribute to my friends, so I do  
> this:
>           sage -bdist 4.1.2mod-solaris
>
> Or, I'm doing something internally and do
>
>          sage -bdist internal_version
>
> By forcing the version number at the front, you're making "sage
> -bdist" significantly less flexible and breaking a nice symmetry
> between bdist and sdist.

Those are all good points. What if the version flag is optional, in  
which case it's generated automatically (detecting the OS as well, or  
failing)?

- Robert


-- 
To post to this group, send an email to sage-devel@googlegroups.com
To unsubscribe from this group, send an email to 
sage-devel-unsubscr...@googlegroups.com
For more options, visit this group at http://groups.google.com/group/sage-devel
URL: http://www.sagemath.org

Reply via email to