> Am 26.10.2014 um 12:58 schrieb Pierre Joye <pierre....@gmail.com>: > > On Sun, Oct 26, 2014 at 3:24 AM, Joe Watkins <pthre...@pthreads.org> wrote: > >> I'd like to everyone to stay grounded in reality and say that we have >> been checking code into php-src for months, very very few people have >> expressed an interest in what we were actually doing, you aren't one of >> them Stas. > > > As much as I do not like the way this topic was brought to the list. A > point has been raised. > > The chnages done in the latest patches, like the XML based protocol, > are not small changes and will affect us, our users and external tools > for a long time.I understand that discussing things too much prior to > apply it takes time, energy and sometimes could kill the motivation, > but this is how we reach consensus. > > I do not think it is too late to have this discussion, maybe part of > the php specification too.I am a big fan of phpstorm, but it is by far > not the only tool around. The discussions here, if we filter out the > classic bashing and not so well formulated critics, try to figure out > if the choices are actually the best ofr php, now and for the next > years. > > So. Yes, I also have to ask, humbly, both of you to also stay grounded > and accept to participate in a constructive discussion. Maybe push up > a RFC, documenting the current XML protocol, see the pros/cons, etc. > Other can also join this effort. I am sure we can find answers to all > these questions soon, if we leave the sentimental parts of this > discussions out of this, focusing on the actual technical and design > facts. > > Cheers, > -- > Pierre > > @pierrejoye | http://www.libgd.org <http://www.libgd.org/>
I’m fully aware that this is a phpng-style push in a bit smaller scope. And, despite whether that’s in spirit of open source (I’m not so sure if it really isn’t), it was something that had to be done. It’s not yet too late, we can still polish things which aren’t fine in the protocol. Also, it was *necessary* to write the patch first to see how the protocol will also fit a bit into the code. But then I had to restructure the code a lot. And later the thing got so big it got just too much maintenance work to simultaneously develop on the master branch (in krakjoe/phpdbg repo) and xml-protocol branch, so I merged everything. By that time I had a lot of unrelated features and the XML protocol then in master. (At that point I didn’t think it might become an issue). Then more than a week later I found it relatively stable (maybe some polishing left, but ± stable). I considered to have more discussion before, but it’d have blocked all the other improvements pending (would probably have been too much work to separate things again and push). And one really needs a starting base before real discussion can take place (The patch alone IMO is anyway too big to be reviewed as is). So, I pushed. But let us not discuss if it really was a good idea to have it done that way - or not. Please rather discuss where the protocol can be improved, if it’s missing some features etc. There is a Markdown file about the protocol in sapi/phpdbg/xml.md, which is documenting the XML responses one might get. I’m aware that it’s not really well structured, I appreciate any help there. But mainly, please first review if the design of the protocol is okay. Quick info (because it already caused some confusion): input commands are the same for xml protocol than for interactive mode. It’s just the output which differs. Bob p.s.: I don’t regret to have chosen XML - I mean, it’s now to late to discuss whether to choose XML or not, but I don’t regret the decisions I made there.
