The earliest prototype of Xacerbate was written in perl. There were two problems with this:
- The standard CPAN libraries didn't offer support for implementing features such as document cacheing, which seemed important
to us to maximize efficiency
- We encountered some LibXML2/LibXSLT memory issues that seg-faulted perl, and trying to identify - let alone - fix those issues looked like a long and uncertain journey
Our best guess was that the memory issues were due to conflicts between the C XML libraries and perl regarding the meaning of 'freed memory'. The best
way to avoid those issues was to take control of the memory management. In other words, the low-level nature of C was for once a feature! Also, Tony
had experience of using the LibXSLT library to build C applications.
Xacerbate is not intended to grow into an ever more complex application. It needs to do a relatively simple job as efficiently as possible, to support
the interesting stuff at XSLT application level. For that brief, C looked like a fairly good choice.
0 comment(s) so far, not open for comments