Re: [lmi] Proposed workflow for proprietary repository

[Vadim Zeitlin]

On Sat, 11 Nov 2017 00:33:07 +0000 Greg Chicares <address@hidden> wrote:

GC> First let me ask whether you received the message that you replied
GC> to here in normal email, because I did not, not even at my corporate
GC> email account that is also subscribed to this mailing list.

 I received it from the mailing list and I see it in the archives (see
http://lists.nongnu.org/archive/html/lmi/2017-11/msg00017.html), so I
don't see any problems from here.

GC> What are the exact consequences, in a world where:
GC>  - Kim uses msw, and I do not
GC>  - Kim sets 'core.filemode' to false, and I do not
GC>  - Kim adds or modifies nonexecutable files often, but executable ones never
GC> ? AFAICT, under those assumptions, any file she touches is mode 644, but all
GC> the files she ever touches should be 644, so this really will just work.

 Yes, this is what I believe too.

GC> OTOH, looking for an anticipated "Okay!" message is not as good as full
GC> automation; and if it is not observed, then manual cleanup is necessary.
GC> Instead, because we really want a cut-and-pastable series of maximally
GC> robust commands, I think I'll amend the instructions thus:
GC> 
GC>    # When everything is ready to share...
GC>  
GC>  cd /opt/lmi/proprietary
GC> +printf "forcing correct permissions "
GC> + for d in src data test tables; do (\
GC> +      printf "$d..." \
GC> +   && find ./$d -maxdepth 1 -type f -not -name '*.sh' -not -name '*.sed' | 
xargs chmod -x \
GC> + )done
GC>  git bundle create YourBundleName origin/master..HEAD --branches
GC> 
GC> I tested that, and it seems to do the right thing, quickly enough...

 Looks good to me too.

 Regards,
VZ