Date: Sun, 29 Mar 2020 22:17:23 +0200 From: Jan Beich <jbeich@FreeBSD.org> To: "Mikhail T." <mi+t@aldan.algebra.com> Cc: gecko@freebsd.org Subject: Re: Shared libxul.so (was Re: Restoring seamonkey) Message-ID: <imim-c3vw-wny@FreeBSD.org> In-Reply-To: <fae53f0b-0c2c-0a08-d2db-cb40339abc3f@aldan.algebra.com> (Mikhail T.'s message of "Sun, 29 Mar 2020 11:33:11 -0400") References: <857ef528-1dfd-12b6-6579-b03a137ff199@aldan.algebra.com> <wo75-5lf5-wny@FreeBSD.org> <9a797087-e769-3c50-3032-c71b41fab823@aldan.algebra.com> <4ku8-x9zl-wny@FreeBSD.org> <fae53f0b-0c2c-0a08-d2db-cb40339abc3f@aldan.algebra.com>
next in thread | previous in thread | raw e-mail | index | archive | help
"Mikhail T." <mi+t@aldan.algebra.com> writes: > [Forking the Seamonkey thread, because the topic is different] > > On 28.03.20 20:47, Jan Beich wrote: >> libxul.so is no longer the same between various Gecko-based projects. > > It doesn't have to be /the same/ -- just needs to still be compatible > enough for a motivated developer to patch the rest. An argument added > here and there, a method deprecated -- such problems may be solvable > within a port. > > Maybe, libxul -- the best version (from firefox tree, I guess?) can be > moved into a port of its own, for firefox and others to use... Maybe, > a few such ports -- for different major versions, if necessary, the > way we have ruby22 and 24, for example. At best libxul.so from www/firefox-esr and mail/thunderbird can be shared. www/firefox, www/cliqz, www/seamonkey, www/palemoon are based on different Gecko versions. Major versions are released every month, bringing a massive code churn. API changes are probably similar to major LLVM releases except one can't rely on upstream fixing compatibility i.e., low bus factor. See also https://bugzilla.mozilla.org/show_bug.cgi?id=1038639 > And, yes, I'd be willing to do it -- if I can be reasonably assured, > the work will not be rejected off hand on the grounds of "exhausting > maintenance" or some such. > > I've done it before, as a matter of fact -- it was my work, that once > changed firefox and thunderbird to use the common nspr and nss > (r1r141034, r41037), for example. Maintenance is a continuous effort. As you've given up on upstreaming some of your nspr/nss patches complicate updates. I'm guilty of this as well which is why I mainly keep statu quo nowadays. > Back then there was an agreement among port-maintainers, that sharing > components is a good thing -- do we still share this sentiment? It's a trade off. Unbundling has a benefit when it doesn't complicate maintenance much. Sometimes building a bundled dependency is harder than forcing a system version.
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?imim-c3vw-wny>