Re: a2ps download site

["Max Bowsher" <maxb at ukf dot net>]

Igor Pechtchanski wrote:
> On Tue, 15 Apr 2003, Max Bowsher wrote:
>
>> Igor Pechtchanski wrote:
>>> David,
>>>
>>> Looks much better now.
>>> A few more nits:
>>> - The patch still has CRLF line endings.
>>> - The .tar.gz file extracted from the source archive has executable bits
>>> set.
>>
>> This hasn't been rigourously checked for in the past.
>> Does it matter? If so, someone needs to patch the "Contributing" webpage.
>
> It probably doesn't matter, which is why I called them "nits" rather than
> "problems".  I'm "nitpicking" :-D
>
>>> - Why does the "all" step in the script perform the "spkg" and "pkg"
>>> steps?  They are not needed for the build.  Neither is "finish", for
that
>>> matter.  (That may be an artifact of the build script, if so, ignore
this).
>>
>> That's from the build script, so leave it as is.
>>
>> If you think about it, "all" from the perspective of a packager does
>> logically include spkg, pkg and finish.
>
> Yes, but then the build instructions in the README should not say
> "./a2ps-4.12-1.sh all", but rather "./a2ps-4.12-1.sh prep &&
> ./a2ps-4.12-1.sh conf && ./a2ps-4.12-1.sh build && ./a2ps-4.12-1.sh
> install".

Depends how you define "build". Anyway, existing packages just say
"./buildscript.sh all".

>>> - Is there a particular reason why the build is performed in-place,
rather
>>> than in a .build directory (which is created, BTW)?
>
> Is this one usually checked for?  Does it matter?

Well, its not ideal, but personally, I think its better to build in place
than to make extensive modifications to the sources to build out of tree.
(Unless the maintainer is going to push the modifications upstream).


Max.