Re: [Swig-devel] The swig hate webpage

[Marcelo M. <mm...@ac...>]

David Beazley wrote:

>
> On Dec 7, 2005, at 7:58 PM, Marcelo Matus wrote:
>
>> Ah!, finally we have an official (sort of) I-hate-swig web page!
>>
>> For your enjoy:
>>
>>   http://bob.pythonmac.org/archives/2005/03/11/swig-hate/
>>
>
> Ah yes.... This reminds me of a quote I heard from John Ousterhout at  
> a Tcl conference several years back:
>
>    "There are two types of programming languages---those that  
> everyone hates and
>     those that nobody uses."  - J. Ousterhout.
>
> So, in that vein, I'm not so concerned about it;-).
>
> The hate page does have some valid points though. I think we do tend  
> to change things a lot---maybe even more than we should.  I often  
> wonder whether the 1.3.x branch will ever stabilize. I also sometimes  
> wonder whether my absence from the scene is a factor (as most people  
> know I've been missing in action for awhile). On the other hand, I  
> never intended SWIG to be used in a way where you would be writing  
> tons and tons of typemaps and all sorts of low-level API code.   So,  
> if people are using it in that way and stuff is breaking, they've  
> really missed the whole point.


well, I hope with the new Unified Typemap Library, at least at the 
typemap side, the things get much more
stable and consistent between languages (perl, python, tcl and ruby at 
least).

I think the stablity will be always a haunting problem since even we 
have a very large test-suite to
ensure backward compatibility, people use swig in unexpected ways. And 
of course, even we
have a bug track system, mailing list, etc, the problems are not always 
reported, and users
keep using the last version that works for them, or just rant.

Strangely enough, ranting usually takes more time than filling an error 
report.... but well,
the human nature.

But for the same, given  the last large changes, and that most of the 
users never test the
CVS version, I was looking for large projects using Swig and try the CVS 
version
with them. I already tested Subversion, PyOpenGL, PyICU, PyOgre (with 
the help
of Clay Culver), and in the process I found the swig-hate page too :).

I hope this release to be very very very compatible. But well, even when 
I already found/fixed
some bugs in Swig, I also found some  new behaviors in G++ (oh my god!) 
that break
some interfaces as G++  get closer to the C++ standard.

Aaah!, time to start another I-hate-C++ web page!

Also, with the new Unified Typemap Library, I hope we can document and make
public some of its internal functions/macros, and then reduce the 
unexpected ways users
are now using swig.

Marcelo


>
> I have to say though, I've gotten pretty used to complaints about  
> SWIG.  It's easy for someone to come along and say that it "sucks"  
> along with some comment like "one of these days someone is going to  
> come along and do C/C++ parsing right."   I just have one response to  
> that---do it then.  If this is so easy and you know how to do it  
> "right", then stop whining, put your money where your mouth is, and  
> make something that's better.   I'll be more than happy to tell  
> people about it and might even become your biggest supporter.
>
> For what it's worth, I've always had a fairly neutral attitude about  
> SWIG. If someone were to come up with an alternative that completely  
> blew it out of the water, that would be fantastic.  As it is, the  
> SWIG web page has always tried to point people to alternative tools  
> and approaches. I've certainly never made any claims that SWIG is the  
> best, the easiest, or only way to solve the wrapping problem.   
> Frankly, I'm flattered that people are using it at all.
>
> Cheers,
> Dave
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
> -------------------------------------------------------
> This SF.net email is sponsored by: Splunk Inc. Do you grep through log 
> files
> for problems?  Stop!  Download the new AJAX search engine that makes
> searching your log files as easy as surfing the  web.  DOWNLOAD SPLUNK!
> http://ads.osdn.com/?ad_id=7637&alloc_id=16865&op=click
> _______________________________________________
> Swig-devel mailing list
> Swi...@li...
> https://lists.sourceforge.net/lists/listinfo/swig-devel