internal dep generator sloooow for python

If I use the internal dep generator on rpm-4.4.8, it runs at
a crawl
compared to the external dep generator when finding python
deps.  At least
an order of magnitude slower, probably even 2 orders if not
more.  There's
about .5s delay between exec() of %__python_provides for
each .py file.
Any ideas?

This is on a 4cpu (only 1 used for rpm though) very fast
opteron machine.
Solaris 10.

-frank
_______________________________________________
Rpm-devel mailing list
Rpm-devellists.dulug.duke.edu
https://lists.dulug.duke.edu/mailman/listinfo/rpm-devel

On Mar 6, 2007, at 3:33 AM, Frank Cusack wrote:

> If I use the internal dep generator on rpm-4.4.8, it
runs at a crawl
> compared to the external dep generator when finding
python deps.   
> At least
> an order of magnitude slower, probably even 2 orders if
not more.   
> There's
> about .5s delay between exec() of %__python_provides
for each .py  
> file.
> Any ideas?
>

The internal generator is per-file, not per-build-root.

The .5 second delay smells like the hand-tuned delay in the
select()
loop[ in what used to be known as getOutputFrom(). The delay
was  
calibrated
to give perl dependency extraction pig enough time to wallow
out of  
its sty.

Disable the python generator, it's doing anything useful,
even ever
years of requests for assistance from python script
kiddies.

> This is on a 4cpu (only 1 used for rpm though) very
fast opteron  
> machine.
> Solaris 10.
>

If you want to fix, find where the cycles are disappearing.
There's  
certainly
nothing intrinsically slow about an exec per-file, and the
script is  
easily examined.

73 de Jeff

_______________________________________________
Rpm-devel mailing list
Rpm-devellists.dulug.duke.edu
https://lists.dulug.duke.edu/mailman/listinfo/rpm-devel