Re: [PATCH v2 0/7] gitweb: fix "make" not including "gitweb" without NOOP run slowdowns

>>   - I wondered if "make NO_PERL=1" would complain about "gitweb" being
>>     in the default targets. It doesn't, but it does actually build
>>     gitweb, which seems a little weird. I don't think we actually rely
>>     on perl during the build (e.g., no "perl -c" checks or anything),
>>     and the t950x tests seem to respect NO_PERL and avoid running the
>>     generated file. So maybe it's OK?
>
> I think it's arguably a bug, but as you note we build/test etc. without
> errors, and I think it's restoring the state before e25c7cc146
> (Makefile: drop dependency between git-instaweb and gitweb, 2015-05-29).
>
> Arguably we should replace with a stub script like git-svn et al, and
> arguably we should leave it, as you're more likely to e.g. run gitweb on
> a webserver, so even if you build a "no perl" package, perhaps it's
> convenient to have "gitweb" part of it, and then on that one box that
> runs it you'll install perl...

>>   - Speaking of backwards compatibility: after this series, "cd gitweb
>>     && make" yields an error. It's got a nice message telling you what
>>     to do, but it's likely breaking distro scripts. Again, I'm not sure
>>     I care, but if the point of the exercise was to avoid breaking
>>     things, well...
>
> I think that's OK, having maintained those sorts of build scripts in a
> past life.
>
> I.e. when you upgrade the package it's a minor hassle, and the error
> tells you exactly what to do, and the fix is a 2-3 lines in your recipe
> at most.