search

siemens / codeface

Codeface is a framework for analysing technical and social aspects of software development
siemens.github.io/codeface
GNU General Public License v2.0
67 stars 38 forks source link

README.md: add install wordnet #4

Closed bufferoverflow closed 8 years ago

bufferoverflow commented 10 years ago

Vagrantfile: install graphviz wordnet

Signed-off-by: Roger Meier roger@bufferoverflow.ch

wolfgangmauerer commented 10 years ago

Thanks for the patch! However, I see two issues:

Maybe there are reasons for installing the two packages via the distribution mechanism that I'm missing? Thanks!

bufferoverflow commented 10 years ago

It runs fine for me right now, the one step thing with "vagrant up"

other things:

other ideas?

wolfgangmauerer commented 10 years ago

Hi Roger,

On 03/02/14 23:26, Roger Meier wrote:

It runs fine for me right now, the one step thing with "vagrant up"

other things:

  • removed wordnet from os dependencies.
  • added a homepage attribute to the conf file
  • additional field from DOAP are probably also worth to add thanks! I've merged the changes into branch vagrant on github.

other ideas?

I guess tagging mode is not the best choice for Thrift -- most commits don't seem to carry kernel-style tags. It also seems to me that author and committer are often identical, even if some author != committer actually authored the patch. Is that correct? In that case, proximity mode might be better -- it takes more analysis time, but should produce more accurate results (I'll give the project a shot soon; I'm currently setting up a publically available codeface instance).

I've simplified the stack slightly to avoid introducing a change in one commit, and reverting it in the next. Will merge completely after I've played a bit with the thrift analysis (plus I should also make sure there's some solution for the broken tm-plugin-mail status so that vagrant runs w/o dependency on the state of an external project).

Thanks, Wolfgang

— Reply to this email directly or view it on GitHub https://github.com/siemens/codeface/pull/4#issuecomment-34008067.

bufferoverflow commented 10 years ago

Yes, Apache Thrift does not use kernel-style tags. Usually there is just a comment within commit message with the contributor.

I was not aware of the proximity mode, sounds good.

Thank you! -roger

wolfgangmauerer commented 10 years ago

Hi Roger,

On 05.02.2014 22:06, Roger Meier wrote:

Yes, Apache Thrift does not use kernel-style tags. Usually there is just a comment within commit message with the contributor.

btw., is that a usual convention among Apache projects, or is that specific to Thrift? If it's in wide-spread use, we could adapt the commit parser with comparatively little effort so that the tag-based approach works for this convention.

Thanks, Wolfgang

I was not aware of the proximity mode, sounds good.

Thank you! -roger

— Reply to this email directly or view it on GitHub https://github.com/siemens/codeface/pull/4#issuecomment-34244864.

bufferoverflow commented 10 years ago

Hi Wolfgang This convention is used multiple Apache projects. Would be great to have a commit parser which recognizes this style. Thrift migrated from svn to git and the kernel style tags are not used at the moment. Many Apache projects use svn and such comments with the contributor name. Thanks roger

wolfgangmauerer commented 10 years ago

Quoting Roger Meier notifications@github.com:

Hi Wolfgang This convention is used multiple Apache projects. Would be great to
have a commit parser which recognizes this style. Thrift migrated from svn to git and the kernel style tags are not
used at the moment. Many Apache projects use svn and such comments
with the contributor name.

that's interesting to hear. Mitchell, since you're currently working on analysing some Apache packages anyway, could you please extend the signed-off parser accordingly? We'd loose some disambiguation possibilities because the names are not equipped with email addresses this way, but I expect that to only lead to minor inaccuracies.

Thanks, Wolfgang

Thanks roger

Reply to this email directly or view it on GitHub: https://github.com/siemens/codeface/pull/4#issuecomment-34449629

mitchell-joblin commented 10 years ago

On Wed, Feb 12, 2014 at 3:40 PM, Wolfgang Mauerer notifications@github.com wrote:

Quoting Roger Meier notifications@github.com:

Hi Wolfgang This convention is used multiple Apache projects. Would be great to have a commit parser which recognizes this style. Thrift migrated from svn to git and the kernel style tags are not used at the moment. Many Apache projects use svn and such comments with the contributor name.

that's interesting to hear. Mitchell, since you're currently working on analysing some Apache packages anyway, could you please extend the signed-off parser accordingly? We'd loose some disambiguation possibilities because the names are not equipped with email addresses this way, but I expect that to only lead to minor inaccuracies.

Sure, I will add it sometime in the next couple days. I don't imagine that it should be too much effort.

Kind regards,

Mitchell

Thanks, Wolfgang

Thanks roger

Reply to this email directly or view it on GitHub: https://github.com/siemens/codeface/pull/4#issuecomment-34449629

Reply to this email directly or view it on GitHub.

mitchell-joblin commented 10 years ago

Hi Roger,

On Fri, Feb 7, 2014 at 4:30 PM, Roger Meier notifications@github.comwrote:

Hi Wolfgang This convention is used multiple Apache projects. Would be great to have a commit parser which recognizes this style.

I was planning on extending the parser to include these alternate tags but then I looked in the httpd project and did see any reference to a tag. I understand that Thrift uses them but do you have a couple of other examples aswell?

Thanks,

Mitchell

Thrift migrated from svn to git and the kernel style tags are not used at the moment. Many Apache projects use svn and such comments with the contributor name. Thanks roger

Reply to this email directly or view it on GitHubhttps://github.com/siemens/codeface/pull/4#issuecomment-34449629 .

bufferoverflow commented 10 years ago

Hi Mitchell

please have a look at this example commit message from Apache Cassandra:

Lost row marker after TTL expires
patch by slebresne; reviewed by iamaleksy for CASSANDRA-5762

you will find similar issue identifiers within ml communication and commit messages on several Apache projects. The issue identifier (e.g. CASSANDRA-5762) is probably a better key to be used for analytics in combination with the name of the user commenting on a JIRA issue.

best! -roger

bufferoverflow commented 10 years ago

Apache Thrift might become more interesting to analyse. facebook worked offline for a while and republished their changes: https://code.facebook.com/posts/1468950976659943/under-the-hood-building-and-open-sourcing-fbthrift/ http://arstechnica.com/information-technology/2014/02/facebook-open-sources-thrift-again-with-fbthrift-overhaul/

the code is here https://github.com/facebook/fbthrift and it will hopefully merged back piece by piece to Apache Thrift.

bufferoverflow commented 9 years ago

@wolfgangmauerer @MitchellJ any news on this?

wolfgangmauerer commented 9 years ago

Unfortunately not yet, but we did not forget about this. It's just that we're way slower than we would like to be...

mitchell-joblin commented 9 years ago

@bufferoverflow @wolfgangmauerer Actually we are not slow on this, maybe just forgetful. I added this feature back in the first week of April in commit ed57c250. I also sent the patch to the mailing list before it was merged.

wolfgangmauerer commented 9 years ago

Am 25/11/2014 15:02, schrieb MitchellJ:

@bufferoverflow https://github.com/bufferoverflow @wolfgangmauerer https://github.com/wolfgangmauerer Actually we are not slow on this, maybe just forgetful. I added this feature back in the first week of April in commit ed57c25 https://github.com/siemens/codeface/commit/ed57c25047c36e7bce6ca21dcdf006bc7cc64c4f. I also sent the patch to the mailing list before it was merged.

okay. Then I likely reviewed these changes, but forgot about them. Guess I need to catch a break :)

— Reply to this email directly or view it on GitHub https://github.com/siemens/codeface/pull/4#issuecomment-64403223.

bufferoverflow commented 9 years ago

Thanks @MitchellJ ! What about the changes from this pull request?

wolfgangmauerer commented 9 years ago

FYI: Matthias Gemmer started to work on this issue recently, he'll give the Vagrant (or possible also a docker) approach some testing, and then it will go into master.