--- Log opened Thu Oct 23 10:31:59 2008

10:33 < ohshima> Hi, (this is Yoshiki).  For those who are interested in, I uploaded my authorship code to squeaksource.
10:33 -!- Craig_ (n=craig@c-71-204-145-57.hsd1.ca.comcast.net) has joined #squeak
10:33 < matthewfu> ohshima: thanks
10:33 -!- Craig_ (n=craig@c-71-204-145-57.hsd1.ca.comcast.net) has quit [Remote closed the connection]

10:34 < matthewfu> ohshima: RandalSchwartz mentioned you had a .cs that could be filed into 3.8 to make it license clean
10:35 < matthewfu> that was backported from etoys 4
10:35 < ohshima> The "relicensed" version should have condensed sources and be all clean, so having a new major version number makes sense.
10:35 < ohshima> http://tinlizzie.org/updates/etoys/updates.
10:35 < ohshima> and look for changes for "etoys3.2".
10:36 < ohshima> I guess the email to squeak-dev is still irregular.
10:36 < ohshima> http://lists.squeakfoundation.org/pipermail/squeak-dev/2008-September/131586.html
10:37 < matthewfu> I assume not gap-filler.cs
10:37 < ohshima> and
10:37 < ohshima> http://tinlizzie.org:8080/seaside/examples/authorship2
10:37 < ohshima> are other interesting URLs.
10:37 < kencausey> ohshima: Thank you!
10:37 < ohshima> that is right.  Don't worry about gapFillers, of course.
10:38 < matthewfu> so, even though those are for etoys 3.2, you suspect they'll be fine to patch to Squeak 3.8.1?
10:39 < matthewfu> or will I still need to get your tools and look for things wrong with 3.8.1?
10:39 < matthewfu> etoys has a lot of removals from 3.8.1, I think
10:40 < ohshima> Not so much from 3.8.
10:40 < ohshima> Most of the code that the removal changes deal with are not so different in Etoys image and 3.8.
10:40 < matthewfu> ok
10:43 < matthewfu> so, MethodAuthorship is the tool to calculate all the history for an image and search for unlicensed code?
10:44 < ohshima> yes.
10:44 < ohshima> Load MethodAuthorship-yo.1.
10:45  * matthewfu grabs a 3.10 image and tries
10:46 -!- matthewf (n=tapple@ip24-251-149-228.ph.ph.cox.net) has quit [Read error: 110 (Connection timed out)]
10:46 < ohshima> and evaluate:
10:46 < ohshima> Authorship initialize.
10:46 < ohshima> missing := authors missingAuthorsWithMethods.
10:46 < ohshima> authors := Authorship new.
10:46 < ohshima> missing := authors missingAuthorsWithMethods.
10:46 < ohshima> notAccounted := authors notAccounted.
10:46 < ohshima> Oops insert: authors := Authorship new. at the second line. 
10:46 < ohshima> I'll be back later.
10:47 -!- ohshima (n=ohshima@76.214.15.121) has quit ["Riece/4.0.0 Emacs/21.4 (windows-nt)"]
10:48 -!- You're now known as matthewf

10:51 < matthewf> authors notAccounted returns an empty ordered collection

10:55 < matthewf> neato
10:56 < matthewf> I'll look at this soon
10:58 < kencausey> It needs to have who and who has not agreed to MIT hooked into ti
10:59 < matthewf> oh. I assumed it did that as part of the initialize
10:59 < kencausey> Maybe I'm wrong, I don't see it in the web interface
10:59 < matthewf> questions for yoshi


11:35 -!- ohshima (n=ohshima@76.214.15.121) has joined #squeak
11:35 < ohshima> Here I am.
11:35 < ohshima> notAccounted may be a misnomer, but it means that there is no authorinitilas that is not in our database.  So that is good.
11:35 < ohshima> The database is in the "data" method category of Authorship.
11:36 < ohshima> #returnedSignatories are the ones who actually returned the signed agreements.
11:37 < ohshima> #agreedContributors are the ones who said they will do, but haven't turned the actual written agreement yet.
11:37 < ohshima> #newContributors are people whose code was never under SqueakLicense.
11:38 < ohshima> #missingSignatories are the ones who haven't turned in, or initials we don't know who they are.
11:41 < matthewf> ohshima: does this package automatically fetch the SqueakV1.sources and SqueakV2.sources and the signers list, and compute the version info from those?
11:41 < matthewf> I see the class comment for Authorship has lots of complicated instructions
11:43 < ohshima> No for automatically fetching.  That class comment is the one you have to execute.
11:45 < Craig> why aren't you using the Squeak project at http://netjam.org/squeak/contributors ?
11:45 < ohshima> The files are in http://tinlizzie.org/~ohshima/3.10.2-history.tar.gz.
11:47 < ohshima> Craig: well, creating data from flat text file was more convenient, but there was no strong reason.
11:47 < ohshima> I need to drop out again.
11:48 < Craig> ? that data was already in that project, as a Squeak collection object, which I would think would be more convenient
11:49 < Craig> (the collection contains instances of Contributor, which has behavior one would use when doing searches via select: etc.)
11:49 < Craig> all the textual lists at http://netjam.org/squeak/contributors were generated from it
11:50 < Craig> oh well
11:50 < Craig> it sounds like you entered all that data in again from text, for no good reason
11:51 < Craig> and then you make me modify those Contributor instances again


12:28 -!- ohshima (n=ohshima@76.214.15.121) has quit ["Riece/4.0.0 Emacs/21.4 (windows-nt)"]


12:57 -!- ohshima (n=ohshima@adsl-76-214-15-121.dsl.lsan03.sbcglobal.net) has joined #squeak
12:58 < ohshima> Craig: Ok... then, Craig, there were reasons, some of which you were aware of.
12:58 < ohshima> Just loading the project doesn't set up the class variable, etc., so it doesn't make a workable image by itself.
12:58 < ohshima> Your data had fair number of missing entries and errors, so fresh look was needed.
12:58 < ohshima> A collection of Contributor objects stored in projects doesn't let us compare different rivisions.  (That is a biggie).
12:58 < ohshima> A Contributor object holds just one initial, but in the real world, many contributors have more than one initials so the design wasn't right.
12:58 < ohshima> For various purposes, the collection of methods and its history by an author should be accessible, but yours doesn't have it.
12:58 < ohshima> I didn't type all data in by myself, I started from one of available packages.
12:58 < ohshima> All in all, why don't you just write a short code snippet with #pairsDo: to update your data from my data?  That is the most straight forward thing to do.

13:01 < Craig> ohshima: the class variable was irrelevant when you had an inspector on the collection sitting there in front of you when you loaded the project. You could then set the variable if you wanted, or get rid of it (I got rid of it)
13:01 < Craig> I updated the data, as you asked.
13:01 < Craig> don't know what you mean by "A collection of Contributor objects stored in projects doesn't let us compare different rivisions.  (That is a biggie)."
13:02 < Craig> you could add the collection of methods and its history by an author
13:02 < Craig> I don't want to bother keeping multiple copies of the same data. If you want to keep the master database, fine. I will delete http://netjam.org/squeak/contributors
13:03 -!- toytoy (n=dindin@unaffiliated/toytoy) has quit [SendQ exceeded]
13:03 < Craig> however, I think having actual objects for the contributors is the most useful thing to do
13:07 < Craig> I guess we'll take this up at the next meeting
13:21 < ohshima> Having an inspector there wouldn't get you go too far, though.  You may accidentally close it, and that is it.  "comparing revisions" means the revision of data itself.  How do you know that which entries were updated in last month, for example?  I do have actual objects, too.  These can be generated as you wish.  You don't necessarily have to delete the page, but if it is hard for you to keep it up to date, I can put the late
13:25 < Craig> as I said, you can set whatever variables you like to that collection
13:25 < Craig> you can easily add the code to do the other features you mentioned
13:26 < Craig> I think you're just being difficult at this point.
13:26 < Craig> you weren't using objects the last time you pointed me at your code. It was just methods with a bunch of literals in them
13:26 < Craig> I'm done with this.
13:26 -!- Craig (n=craig@c-71-204-145-57.hsd1.ca.comcast.net) has quit []

13:28 -!- ohshima (n=ohshima@adsl-76-214-15-121.dsl.lsan03.sbcglobal.net) has quit ["Riece/4.0.0 Emacs/21.4 (windows-nt)"]


--- Log closed Fri Oct 24 00:00:33 2008