On Jun 25, 2013, at 8:01 AM, Igor Galić <[email protected]> wrote:
> > > ----- Original Message ----- >> On Tue, Jun 25, 2013 at 8:41 AM, Leif Hedstrom <[email protected]> >> wrote: >> >>> On 6/25/13 3:44 AM, Igor Galić wrote: >>> >>>> Hi folks, >>>> >>>> I'd like to discuss if we can and should sync the plugins >>>> tree including experimental plugins with 3.2.x and any >>>> other "stable" branch, for that matter. >>>> >>> >>> I'm +0 on this, i.e. I have no problem with it, but I personally >>> wouldn't >>> need it. >>> >>> One caveat: It's possible that plugins that works on "master" would >>> not >>> work on stable branches, without also promoting other commits to >>> those >>> stable branches. E.g. fixes / additions to APIs, or bug fixes in >>> the core. >>> >>> >> Yes, I think this will bite us more than once. Also a few plugins >> have made >> the jump from experimental in master. > > Right now, that's two of the main reasons I'm starting this discussion > in the first place: > > Plugins from master that used to *just* work on 3.2.x no longer compile > because of the changes we have made to the header files (ink_defs.h > et al). On the other hand, IMO stable plugins (gzip) that work very > well with 3.2.x now simply don't compile. You mean that they don't compile if you do it manually using tsxs? I guess that I don't see why that is so important, seeing as they get built by the build tree. Mixing plugins from different ATS versions seems like a fairly esoteric use case. If someone (else) wants to to the work, I'm ok with that. > > Given our extensive code for version-checking in the plugins (TS-1953) > I don't think we should *allow* for breakage between versions. > > > i > > -- > Igor Galić > > Tel: +43 (0) 664 886 22 883 > Mail: [email protected] > URL: http://brainsware.org/ > GPG: 6880 4155 74BD FD7C B515 2EA5 4B1D 9E08 A097 C9AE
