Re: Website upload

[Urs Liska]

Am 07.03.2017 um 15:34 schrieb Phil Holmes:
> ----- Original Message ----- From: "David Kastrup" <address@hidden>
> To: "Phil Holmes" <address@hidden>
> Cc: "Urs Liska" <address@hidden>; <address@hidden>
> Sent: Tuesday, March 07, 2017 1:56 PM
> Subject: Re: Website upload
>
>
>> "Phil Holmes" <address@hidden> writes:
>>
>>> Sorry for top post - as you see my email client doesn't properly
>>> indent your reply.
>>>
>>> Anyway, I've now done the following command:
>>>
>>> rsync -n -aO -d --no-r --delete -i $BUILD/out-website/website/
>>> $DEST/website/ &>rs.txt
>>>
>>> Which gives the output in the attached file.  I guess this means it
>>> would only delete the filenames preceded by "*deleting" - I assume the
>>> other files named are just for information?
>>
>> Or update.  But I am afraid that for looking at a dry-run, I'd leave off
>> the --no-r (no recursive directory descent) option.  The output will be
>> humongous but complete.
>>
>> -- 
>> David Kastrup
>
>
> I did run it to start with without the ---no-r and without redirecting
> the output.  The output I got on my terminal was a load of lines like:
>
> .f...p..... pictures/ties-scoring-example.png
>
> and I was concerned that this meant that those files would be deleted,
> so I thought it better to restrict it to the root directory.  I now
> understand the output a bit better.  I've since repeated it with the
> recurse option with the result attached.  As you see, it would delete
> a lot of files we want to keep (examples) but it could be possible to
> use its output as a guide for manual deletion?

I did *some* random checks by searching the source text of
http://lilypond.org/examples.html for links to those files listed as
to-be-deleted, and it seems none of them is actually used anymore in the
page. I did *not* try to download the whole thing and grep for it, but I
have the strong suspicion that files like
ly-examples/tab-example.preview.eps actually *are* the kind of orphaned
files we are talking about (victims of the latest review of the examples
page?)

So ideally someone (as said I can't build currently) could grep through
the whole built website, taking the attached file (produced by a
combination of grep and sed) as the basis and this line:

for f in $(cat to-be-deleted.txt); do git grep -n $f; done

This should return all places in the website where the to-be-deleted
files are referenced - and my hypothesis it would return nothing.

Urs

>
> -- 
> Phil Holmes 

-- 
address@hidden
https://openlilylib.org
http://lilypondblog.org

to-be-deleted.txt
Description: Text document