Re: Re: RFC::Error Handling Guidelines for PHP5 packages

On Tue, 24 Aug 2004, Alexey Borzov wrote: > Hi, > > Davey wrote: > > As I said, this is my final attempt to inject sanity into this... > > although I may contact the PEAR Group on this matter, only they can > > really interject now that the voting process has started. > > If PEAR group turns down the community approved proposal after failing > to participate in community process in any way, this will have very > unfortunate consequences. I have just "returned" from a period where I had absolutely no time for PHP/PEAR, to see that PEAR seems to have absolutely no direction wrt. to exceptions and PHP 5 migration (I have missed out a few thousand emails mind you). This proposal does not go into the nitty-gritty of the real error handling-related issue PEAR is facing: how to migrate to exceptions without too much pain for developers and users? Yes, this is beyond the scope of this RFC, so then my claim is that this RFC is premature. I've suggested to Sergio that we can extend it to include a migration plan _and_ these guidelines, pull this RFC back and then publish a new and comprehensive one. My impression (feel free to correct me if I've misunderstood you Sergio) is that he would rather just push it through now since it's in the pipeline and he (and probably others) have been working on it for so long. If the RFC is premature, it doesn't help how long it's been in the pipe, it's still premature. Does the fact that I've been away for several months force me to shut up when I see that a bad move is about to be made? I've learnt this the hard way: when PEAR was in its infancy, there were bi-monthly "what is PEAR?" flamewars. The main reason for this was that I had not done a good enough job communicating my ideas. But I did communicate parts of it, enough for people to get part of the picture and misunderstand, and then you have it going. Don't do that mistake. I'm seeing the same patterns here and I'm pretty sure the same thing will happen with error handling if we're not careful. If this RFC goes out there, people are going to think this is the rule book for dealing with PEAR errors in PHP 5. It's bad enough that they have to rewrite their code once, if not several times because they keep getting new information about the "latest and greatest" way of doing error handling. Sergio is probably pissed off because I'm trying to make him drop/freeze something he has invested a lot of time in, but that is nothing compared to the collective amount of time all of us would have to spend reimplementing error handling in some package for the Nth time (to make it compatible with another package that just did the same thing). There is an RFC that needs to come before this one, and it is the contents of that pear-dev needs to focus on: * how to maintain compatibility with existing packages until then, both for package maintainers and users * bloat control: PEAR_Error, PEAR_Exception and PEAR_ErrorStack total over 2000 lines of code, error handling must not become even more cumbersome than it is today with PEAR_Error * the end goal, when PHP 5 is the primary development platform I suggest we start there, but in a new thread please. - Stig