Re: Attachment downloader?

[Trevor Daniels]

Assaf, you wrote Sunday, March 15, 2015 3:53 AM
>
> On 03/14/2015 01:57 PM, Trevor Daniels wrote:
>> I confirm this works, at least for the one instance I tried.
>> I guess now we need to write a script which parses the JSON
>> file to find the attachment parameters and then to wget them
>> using the assembled url.
> 
> I think I understand now what type information is available form google-code,
> now I'll need to go look into the savannah code.
> 
> If I understand correctly, the timeline allows projects to keep using 
> google-code issues until August 2015, at which point it will be come 
> read-only?

Correct

> This doesn't leave us a lot of time (especially that there aren't many 
> volunteers working on savannah code...).
> I'll try to work on it (among all the other things), but it will be good to 
> also look at fall-back options.
> if you do examine some other known bug-trackers solutions, would you mind 
> sharing your options with savannah people?

In my opinion, especially following the recent experience of Google, we should 
move to an open-source system.  There are few others apart from Savannah 
(Gerrit is one) but I'm not aware that anyone in the LilyPond project is 
exploring any of them at present.  As our source code is based at Savannah that 
seems an obvious place to keep our issues DB, but the decision will be on how 
well our needs and desires can be accommodated.  Your response means we're off 
to a good start!
 
> As a side note, to better understand the exported JSON format,
> I wrote a small script which generate a static HTML from the JSON and also 
> download the attachments - producing a static local copy.
> It's ugly but it (mostly) works.
> 
> If you're interested, you can see the results here:
>  http://gcissues.housegordon.org/p/lilypond/
> 
> Or download all the files here:
>   http://gcissues.housegordon.org/

This is excellent progress!  Thanks for taking the trouble to do this.  We can 
now confidently assume all the data and all the images can be extracted, at 
least in a usable read-only form.  

I see the size of the tarball of the issues DB is 111Mb.  Not sure how big the 
DB would be when extracted, but it will be large compared to a text-only issues 
DB.  Would the size be a problem for Savannah?
 
> if there are LilyPond developers willing to help with hacking the savannah 
> php code, please send an email to  address@hidden and we'll continue the 
> discussion there.

I'm not really a hacker (my main contribution is LilyPond documentation), so 
I'm hoping some of the others on the LilyPond developers list will pick up on 
this.  There is a lot to do to either tailor the Savannah interface to 
accommodate our established scripts and workflow, or develop a new one based on 
the Savannah facilities.  

To what extent are the various fields in the Savannah issues interface 
tailorable on a project-specific basis?  I explored the website for details, 
but failed to find any.  Perhaps I looked in the wrong places.

Thanks again for your help!
Trevor