Re: [geogig-dev] fixing headers

[Jody Garnett <jgarnett@xxxxxxxxxxxxxxxx>]

Yep, remember to leave out the _javascript_ code that we recycle from elsewhere.

On Fri, Aug 22, 2014 at 6:37 AM, Gabriel Roldan <groldan@xxxxxxxxxxxxxxxx> wrote:

Thanks Jody,

reviewing and compiling the new contribution zip file now.

On Thu, Aug 21, 2014 at 10:57 PM, Jody Garnett <jgarnett@xxxxxxxxxxxxxxxx> wrote:

Headers are fixed, Gabriel if you could kindly review and submit to IP review when ready.

I focused on Java files, doing a second pass to find reference to either OpenPlans or BSD license in our codebase. I did not introduce headers into any pom.xml, xml or property files.

For reference:

* https://github.com/boundlessgeo/GeoGig/issues/799 Update Header with EPL license and correct date 

* https://github.com/boundlessgeo/GeoGig/issues/803 File public class diff_match_patch.java does not follow class naming convention

Jody

On Tue, Aug 19, 2014 at 7:47 PM, Jody Garnett <jgarnett@xxxxxxxxxxxxxxxx> wrote:

So some good news and some bad news ...

* Good news: The Releng project has a copyright header fixing tool that knows how to talk to git hub and figure out file creation and last modification date.

Looking at geogit graph (yes I ment geogit) we can see a range of project creation times.

* Bad news: Commit 49e63f05d67262934c043f8c7a8560be8e2bf243 refactored org.geogit as org.locationtech.geogig did not manage to capture everything as a "rename" so a lot of the files in our repository are only five weeks old.

Looking at the geogig graph we can see this as a massive spike of created files. 

Reviewing the refactor commit shows many files (such as CacheFactory) were actually recognized as a rename. Checking the generated header for these files I initially thought that the header tools was not smart enough to follow renames. However even github history is confused (for example here is the history of a CacheFactory ).

--

Jody

On Tue, Aug 19, 2014 at 12:16 PM, Jody Garnett <jgarnett@xxxxxxxxxxxxxxxx> wrote:

Gabriel asked me to look at fixing up the headers (to correctly list which authors/organisations) worked on each file.

Stack exchange has a partial answer - short listing which file each user has worked on. Since we have a small number of committers this will probably be fine?

git log --no-merges --stat --author="Pattern" --name-only --pretty=format:"" | sort -u

My rough strategy is to bulk replace the headers, and then use queries like the above to fill in the appropriate organisation.

I probably need to do a second query to figure out when files were added, according to stack exchange the following will work:

for file in `git ls-files`

do

    HASH=`git rev-list HEAD $file | tail -n 1`

    DATE=`git show -s --format="%ci" $HASH --`

    printf "%-35s %s\n  %s\n" $file $HASH: "$DATE"

done

I will try these two automated approaches, but may need to ask the dev list here to untangle which github ID belongs to which individual (and thus what organisation to list in the header).

This is not my first time to the rodeo - here are the notes from updating geoserver headers.

--
Jody Garnett
Senior Software Engineer | Boundless
jgarnett@xxxxxxxxxxxxxxxx
877-673-6436
@jodygarnett

--

--
Jody Garnett
Senior Software Engineer | Boundless
jgarnett@xxxxxxxxxxxxxxxx
877-673-6436
@jodygarnett

--

--
Jody Garnett
Senior Software Engineer | Boundless
jgarnett@xxxxxxxxxxxxxxxx
877-673-6436
@jodygarnett

--

Gabriel Roldán
Software Developer | Boundless
groldan@xxxxxxxxxxxxxxxx
@boundlessgeo

--

--
Jody Garnett
Senior Software Engineer | Boundless
jgarnett@xxxxxxxxxxxxxxxx
877-673-6436
@jodygarnett