Is 'purple' secret, too?

[Evan Schoenberg]

On Mar 22, 2007, at 10:40 AM, Ethan Blanton wrote:

> Evan Schoenberg spake unto us the following wisdom:
>> I've been using mtn HEAD for all Libgaim.framework builds since the
>> official switch to monotone.. so Adium 1.0.1 includes a build which
>> is built from the-code-in-monotone-which-wasn't-yet-called-libpurple,
>> as are Adium 1.0.2 betas 1 and 2.  Adium 1.0.2 beta 3, pushed earlier
>> in the week, is using libpurple-as-such, since the s/gaim/purple had
>> occurred.
>
> This is a mistake.  It is unfortuante that we were not consulted on
> this issue.
>
>> I committed monotone HEAD, pre-purple-name-switch, to Adium's Libgaim
>> repository with Adium 1.0.2b2.  That code was therefore immediately
>> available.  I have not committed any changes since then as I didn't
>> want to make the full purpliness public without discussing it here.
>
> This is reasonable, although it would look somewhat suspcious to
> anyone paying attention.

I don't think there's anything wrong with using monotone HEAD before  
the purple name switch.  While it might be 'somewhat suspicious' that  
a number of libgaim changes have been made which to all appearances  
aren't mirrored in the Gaim svn repository, it wouldn't be the first  
time that the libgaim used in Adium has differed from the code in  
Gaim svn.

The use of the purple_* namespaced code, however, is a mistake, and I  
realized this as I was beginning to compose the email which started  
this thread.  I've realized it more in the time between when I sent  
the email and when I returned to my computer just now, and understood  
this fact even more after reading your reply.  Thanks for taking the  
time to reply in such detail, even if your reply has strong tones of  
dislike of me or at least of my actions as regards the issue at hand.

The Libgaim.framework which will be released with Adium 1.0.2 will  
*not* have any mention of purple or pidgin.  I'm currently re- 
renaming all functions, symbols, and defines back to gaim_* and  
GAIM_*.  While there has been an Adium beta which included the  
purple_ symbols (not used in Adium's own code, of course, but present  
in the included binary), that's water under the bridge for which I  
apologize.  You're absolutely right that this is a mistake.

> You are, of course, correct, but I think you understand that you are
> *definitely* splitting semantic hairs here.  Are you saying that you
> had every intention of releasing code which you knew to be part of a
> legal settlement if someone asked, without consulting those involved
> in the settlement itself?  <snip>
> This is not to even mention the shady subversion of the GPL's *intent*
> in requiring particular requests for source code in an age where
> people simply expect the "Download Source" link on a web page to
> provide them with the latest and greatest.  This issue, of course, is
> one of moral character and stomach for dissimilation, so I suppose it
> is out of scope.

No, if someone asked, I would have consulted with this list before  
taking any action; if there were a problem with the naming, I would  
have done what I'm doing right now - reverted the namespacing by hand  
- before distributing.  I suppose that *is* a question of "moral  
character and stomach for dissimilation," as doing so then requires  
weaving a story for our hypothetical individual interested in the  
code such as, "Oh, I renamed the functions because we were talking  
about branching for Adium's own libgaim framework; that was a silly  
idea, though, so they're back to how they were before.  None of the  
actual code was changed in the process."  I read your statement of  
the issue to say that I have low moral character and a strong stomach  
for dissimilation.  I disagree on both counts, if that's the intended  
implication.  It is, in any case, out of scope for the list, as you  
say -- if you would like to continue that aspect of the discussion  
privately off-list you're welcome to do so.

Realizing after releasing a beta using the purple_* namespacing that  
this was a potential problem, I consulted this list before it became  
an issue.  I've never had a 'download source' link on the Adium site  
which led to the libgaim source; instead, it's included as a binary  
in the normal Adium distribution with a separate page explaining how  
to checkout the Libgaim.framework repository if desired.  I certainly  
would not have had a link saying "Download Source" which led to an  
outdated copy of the tree.

> I am extremely disappointed in the judgement of the Adium developers
> in making this decision without consulting us.
This decision was mine and mine alone; no other Adium developer  
manages our use of and interaction with libgaim (libpurple) besides  
in the rare bug fixing foray.  It would be a mistake to direct any  
upset, ire, or disappointment at anyone besides me for this decision.

> Fortunately, it looks like a settlement is at hand (but how many  
> times has that happened?),
> so we may reasonably hope that this lapse of reason will not adversely
> affect it.
As a last moment of defense: the name 'pidgin' has never been part of  
the Adium code base or committed code at any point; only 'purple' has  
been.  The renaming of the library -- fairly recently split off from  
GTK Gaim -- could easily be a matter of internal reorganization  
unrelated to the settlement.

Regardless, it could also easily be read as indicative of a name  
shift already in progress, and I sincerely hope there will be no  
adverse reaction to my lapse of judgement in terms of the settlement.

As mentioned above, I'm correcting this immediately, so all symbols  
from Libgaim in the released version of Adium will be (as they were  
in the previous release) solely namespaced gaim_*, with available,  
current code available under the same condition.

Cheers,
Evan
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://pidgin.im/cgi-bin/mailman/private/cabal/attachments/20070322/67fca381/attachment.htm