There is obviously a lot of sadness plown down in this. I will try not to come down on misconceptions, but still address some of the issues I feel are important to note. With my insight into the thoughts of the people over here, I'd at least hope to shorten the distances a bit.
I agree. It happens too often. The main problem of Pike team is
- what they don't need is not going into Pike.
Let me rephrase that a bit, from my own perspective. Something I have little or no experience with, is not something I would hack into Pike myself. Not because I am evil, but because I have other work to do and because the result would be substandard to what a more knowledgeable person would achieve. If such a person would want to volunteer to do the work, though, I would do my share of helping him achieve that. (I believe many of you might recall having a few mail roundtrips with me about ssh keys, account names and the like.)
The same goes for most of the guys here; they all have their own business to attend to, and do what they can with the resources they have and time they can afford to help out. Some take part in the discussions here, some do a lot of silent work behind the scenes, fixing bugs shown by pikefarm or cleaning up code and filling holes in the autodocs. The point here, is that the development is driven by the guy that does something - not by a core team that has an overview of all things people want, and a work alottment to spread over them, in some fashion to their own liking. No community project I am aware of works that way, for that matter.
A lot of excuses given, including "lack of documentation" or "lack of portability" (depsite the fact that some parts of core is not well documented, and some parts of core are extremely ineffective).
For starters, lack of documentation is, I believe, the top most common complaint about Pike, and hence we have clearly marked documentation as mandatory for any and all people who want to commit code to Pike.
As for the core, which still lacks lots of documentation, we try our best to fill the blanks. The extremely ineffective parts in the core would probably benefit from being exposed, if you have such knowledge. For all I know, you might be on to something nobody else is aware of.
The lack of portability has not been an excuse to forbid anything so far. The contributions from Nilsson and Agehall to the recent thread about portability were aiming to help add the scaffoldings to make new additions not break platforms where they don't work. While we all like code that works everywhere, all new pike code does not have to use the smallest common denominator of platform/OS support. We do however make use of autoconf and ifdefs, so pike does not break when run in such environments where the code does not work. This is a rather lax demand when compared to the utopia of complete portability.
[...] I didn't bother to share my opinion with someone. Especially after last experience (related to SSL stuff - I asked a question and it was completely ignored, despite that it is a serious flaw).
Please do not mistake silence for being ignored. When asking questions about a complex subject, chances are nobody knows the answers, nor have the time and knowledge to dig them up. It is unfortunate, but I think this was just what happened in the SSL case.
/ Johan Sundström (folkskådare)
Previous text:
2003-04-24 22:24: Subject: Re: Sad.....
On Thu, Apr 24, 2003 at 09:31:39PM +0200, Xavier Beaudouin wrote:
Why we can't promote Pike ?
- sometimes compatibilty is not really compatibility
IMHO, it is not a good idea to support outdated and rare platforms which are in use by two or three individuals on the planet... But of course it is cool to say that Pike supports 50+ platforms (instead of pulling out support for something that is obsoleted or extremely rare).
- it is NOT embeddable.... (so we can't make a mod_pike for apache).
And it will never be, I believe. At least it will not be done by the Pike team. Why? "Nobody needs it" and "It is more simple to embed Apache into Pike", or even - "Why one needs Apache when there is Roxen?" :)
You can say it is sad, but when I send messages to pike devel ml some of them are thrashed, others are ignored or flammed.
I agree. It happens too often. The main problem of Pike team is - what they don't need is not going into Pike. A lot of excuses given, including "lack of documentation" or "lack of portability" (depsite the fact that some parts of core is not well documented, and some parts of core are extremely ineffective).
This is something that usually called "Not invented here" syndrome. Sad, but this is true.
Actually, what you say - are my thoughs, but it was so obvious that I didn't bother to share my opinion with someone. Especially after last experience (related to SSL stuff - I asked a question and it was completely ignored, despite that it is a serious flaw).
What I do here then? I read. I want to be "in touch" - just in case. Perhaps, one day the Pike team will change their attitude...
Regards, /Al
/ Brevbäraren