Re: patch for a debugger in the making ...

Hi Joe, On Nov 16, 2013 10:02 AM, "Joe Watkins" <krakjoe@php.net> wrote: > > On 11/16/2013 03:01 AM, Andi Gutmans wrote: >> >> Hi Joe, >> >> Congrats on launching this debugger interface. >> I’ve actually been thinking for some time that it’d be incredibly useful to have something like this in the core distribution of PHP. >> >> Is this something that would be interesting for you to work towards? What do others think? >> >> Andi >> >> P.S. - It wasn’t clear to me what the “quiet” flag does but I would air on the side of not showing opcodes by default but have an option where they are shown, and in that case, potentially show even more detail [that can easily be contributed. >> >> >> On Nov 13, 2013, at 1:49 PM, Joe Watkins <krakjoe@php.net> wrote: >> >>> Evening All, >>> >>> A couple of us are working on a debugging platform, in order to deploy on windows we require a few more functions to be declared ZEND_API: >>> >>> https://gist.github.com/krakjoe/7456833 >>> >>> If it's not going to cause any disruption, could we merge that into master please ?? >>> >>> Cheers >>> Joe >>> >>> -- >>> PHP Internals - PHP Runtime Development Mailing List >>> To unsubscribe, visit: http://www.php.net/unsub.php >>> >> > > Morning Andi, > > Thanks :) > > Pending the opinion of everyone else; but YES, definitely interested in working toward that ... > > We do indeed air on the side of caution with quietness enabled by default. The only time quietness is ignored is while stepping, when it's useful to show the oplog regardless. We do intend for the oplog to be a decent representation of the opline, I've had some luck representing op arrays in a human understandable way while working on explain(): > > http://explain.so/explain.php?uri=https://raw.github.com/ircmaxell/PHPPHP/master/php.php > > explain() has a problem in copying [printing] (some) literal constants that I've not yet been able to trace, so didn't want to introduce the same bug here ... > > No RFC exists, I'd be happy to create one at some point if we're discussing inclusion, and it's necessary for one to exist ... Awesome work. There is definitively a need for such tool. One thought, I wonder how hard it is to integrate it in an IDE, if there are specials featured which could make ide integrations smoother or nicer. I would love to have it ad soon as possible on the core. However I'd to be extremely careful for the 1st step. It would be very good to define a set of must have features, make them rock solid, bundle the new sapi, and then continue the development, as usual. The reason to choose this way is to avoid what happened with opcache and 5.5, delayed final release but still unstable opcache. Thanks again for your great work! Cheers, Pierre