Thanks to the sterling efforts of Michael Kohler, we now have a page listing the third party code in the Mozilla tree. This is important for licence compliance and so that downstream users of our code know what they are using.
We have attempted to make sure the information is correct, but we would greatly appreciate it if some core hackers could skim down the list with the following questions in mind:
- Have we correctly identified in which product/products each library is being used?
- Are you able to shed light on any of the remaining queries (marked XXX)?
- Is the list complete, to your knowledge?
Thanks!
Freetype is only used in windows mobile and android builds. Not sure if you’re breaking the usage down by OS though.
– Anything listing Firefox and Thunderbird but not SeaMonkey (Growl) looks suspect to me, unless it is only for obsolete versions of Fx+Tb. I’m not saying the “user” attribution is wrong, but maybe it should be double-checked. (The Sm2 about:license mentions the file toolkit/components/alerts/src/mac/growl/ — Sm1 din’t use Toolkit but Sm2 does.)
– “Gecko (up to 3.6)” (libfishsound, liboggz) probably ought to be “Gecko (up to 1.9.2)”.
– Talkback (by Supportsoft), which is closed-source, is still distributed with Thunderbird 2.x (which is almost at EOL but not yet completely — 2.0.0.25pre nightlies are still being built). Shouldn’t it be mentioned with something like “Firefox, Thunderbird (up to 2.x)”? Hm, I see that this component has a bare one-line mention at the bottom of the Tb2 about:license page…
What’s the deal with http://mxr.mozilla.org/mozilla-central/source/netwerk/wifi/src/osx_wifi.h? Is header-only code not counted as third-party, or did someone just miss this (it is in about:license).
Actually, thinking about this more, I wonder if this page is really the right way to go about providing this information.
1) The page lives on wiki.mozilla.org, which is sort-of a random catch-all wiki usually used for planning and thinking aloud, not for documentation or authoritative information. For the information in this new page to be easily accessible, it should be somewhere more prominent (either http://www.mozilla.org/MPL/ or a related section of http://www.mozilla.org, or perhaps on DevMo). This is more-or-less easily fixed by simply moving the page once it actually is in an authoritative state.
2) More serious, however, is that this page is in many ways a needless duplication of the actual authoritative listing of third-party code used by mozilla.org consumer projects: about:license aka toolkit/content/license.html. Worse, this new page is external to, and disconnected from, the code usage, so there is a significantly greater likelihood that the page will fall out of date as people update about:license with new third-party code.
I would suggest that if we wanted to provide an alternate version of the third-party code usage enumerated in about:license, it would be better to do so by having a script that automatically parses toolkit/content/license.html for changes periodically and then updates the alternate version with the new third-party code.
Thanks for the help, everyone!
Tony: Talkback isn’t relevant because a) we aren’t shipping it for much longer, and b) no-one else can ship it anyway because they don’t have the server code. This is why we created Breakpad :-) I’ll let the SeaMonkey folks chime in with knowledge about their situation, rather than make guesses.
Smokey: This is not an alternative to about:licence, because a) about:licence only lists the code we are required to list by the licensing terms, and b) it only lists code used in core products. I maintain about:licence, so they are not going to get out of sync.
Having said that, Michael’s investigations have turned up a few places where about:licence probably needs to be updated.
I’m happy with the page being on wiki.mozilla.org; I’m watching it, so no-one is going to vandalise it without me knowing.
SeaMonkey developer here. We build a standard toolkit same as Firefox and Thunderbird. We may not use all of it but we do ship it.
We also build all of MailNews Core (hg.mozilla.org/comm-central/mailnews/). Again we don’t use all of it (e.g. Gloda) but we do ship it.
So anything in Toolkit and in MailNews core is in SeaMonkey trunk.
Phil
Philip: OK, I updated the page according to your rule. Can you check it and make sure it’s right? Of course, anything marked “Gecko” or “Spidermonkey” implies SeaMonkey.
Looks fine to me Gerv.
By the way, the link to the libical website. Shouldn’t you link directly to http://www.sourceforge.net/projects/freeassociation/ instead of to a website for a product that uses libical and so links to it? Rather like our own ThirdPartyCode page in fact.
Phil
Oops.
“LibIcal is maintained by the Citadel Staff. Its Page-Content is maintained in the Citadel.org libical page”
Hi, anyone can help me hack this site study-world.ir
i get sql user : SUPP0RT pass : 102080
i want delete all, because they are so Fraud
This is good. I’d like to see something like:
http://neugierig.org/software/chromium/notes/2009/12/forking.html
as well, listing differences between bundled code and upstream. Then work towards using as many of these libraries from the system as possible (on Linux anyway).