tapset proposal

[Frank Ch. Eigler]

Hi -

Vara wrote:

> Last week i started discussion about tapsets, i am attaching some
> more details.  [...]

As I understand it, your proposal consists of one possible
standardized interface between the C code that the translator emits,
and a library a tapset author may write.  If so, then this scheme fits
into the taxonomy from my other note, the following way.  This is a
special case of the "modified translator" extension path, since the
translator itself needs to emit different C code, in order to call
these tapset library routines.  (Specifically, it would to know to
associate that "sys_read_tapset_func1" function with the probe-point
specification, and to call the associated extraction thingie.)

The degree to which standardizing such an API is beneficial is a
worthwhile question.  One one hand, the tapset author would "own" both
sides of the interface: both the translator changes that call the
functions, and the functions themselves, so he can taylor-design each
little bit.  On the other hand, making similar subsequent tapsets is
made simpler by duplicating an established pattern.

- FChE
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: not available
URL: <http://sourceware.org/pipermail/systemtap/attachments/20050407/5fda984f/attachment.sig>