Re: Revisiting WebkitGtk+ in Balsa



Hello Peter,
On 2017.10.26 15:34, Peter Bloomfield wrote:
[snip...]
The warning about webkit1 has been in place since that discussion (in git master, now that development has moved there). To support Debian wheezy we certainly need to keep webkit1 configurable until next May. Ubuntu precise is past its EOL, so no longer an issue. Ubuntu trusty provides the newer webkit2 widget, but only the early webkit2gtk-3.0 API.

Balsa support for webkit2 initially used that API, before upgrading to the webkit2gtk-4.0 API. I looked at the commit that implemented the upgrade, and it's not very extensive; I believe we could build with either API by using the magic of conditional compilation. It would add some minor complication to configure.ac and some #ifdef stuff in one C file (libbalsa/html.c), but that seems like a small cost for enabling Ubuntu trusty users to migrate from the insecure old widget to the newer one, and allowing Balsa to drop webkit1 support next May.

Opinions?
I don't know that my opinion should carry much weight, but even worse, I'm of a really mixed opinion here. Supporting a major distribution is good, given that there is a definite end in sight. However, the other project I'm most heavily involved with (KMyMoney, a personal finance app) seems to have already been dropped by some distros as part of their dropping support for KDE4 libraries. (The KMyMoney team is working on the migration to KDE Frameworks and Qt5, but isn't release ready yet.) That's perhaps worse, in that it makes an app unavailable for a given distro, rather than just restricting users to a perhaps not most recent version.

Given that wheezy is LTS, am I correct that there is not much active development or new versions of things, just bug fixes? If so, might it be possible/reasonable to create a git branch which retains webkit1 support, while dropping it from master. Any significant bug fixes can be back-ported, and a tarball created from that branch for wheezy, if necessary. I suppose it really depends on how much effort it will take to support all versions of webkit until next May. It sounds like adding the IF's for conditional compilation isn't too hard, but will there be an other enhancements that will be harder or even prohibitive to implement in webkit1? Yes, all could be wrapped in conditionals, but is it worth the effort? It's really up to the folks who actuall muck about in the code.


Jack


[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]