Author: Kyle M
Revision Date: 2012-01-12 17:45:37 UTC

* common.py: use json falling back to simplejson
* dsa.php: fix on 64bit archs, php's gmp bindings suck for random generation so switch to openssl for now
* dsa.py, dsa.php, demos/dsa.php/key_basics.php:
 1) in generate_pq/V creation be mod 2**seedlen as well as in validation, that was fixed in the final FIPS-186-3
 2) fix generation/validation of g, index and count were supposed to be binary representations after-all (test vectors at http://csrc.nist.gov/groups/STM/cavp/documents/dss/186-3dsatestvectors.zip confirmed it)

Author: Kyle M
Revision Date: 2018-03-13 23:23:12 UTC

Rioch format 0.2, just adds padding in the header for future fields

Author: Kyle M
Revision Date: 2018-02-12 10:56:59 UTC

Initial import, this is some circa-late-2016 code for parsing the blif output from yosys.

You'll want to use lineparser/bline_read as a starting point, the code parsley generates for the ometa grammar is a bit too slow, faster just to use regular expressions to parse.

Author: kyle
Revision Date: 2013-04-14 08:51:48 UTC

* tfplib/scripts/edisp: use python2 instead of python
* tfplib.fileutils.Searcher:
 * Handle include lines in ld.so.conf (TODO: might need to ignore them on non-modern-linux systems if it's linux/glibc/binutils specific)
 * Be a bit saner with cache entries (previously if you used Abs with cache=True or AbsAll with cache=True and didn't iterate over all the results in the future you wouldn't get any new entries)
  * Cache entries from ldconfig don't have .done set currently so iterating over AbsAll will have to check each path referenced in `sections` again before StopIteration'ing.
  * It might still be somewhat suboptimal, if you don't finish iterating on the next run after yielding the cached entries it will still check to see if the file exists in each path referenced by `sections` that isn't cached. Keeping a set of directories/paths checked might make sense.

Author: Kyle M
Revision Date: 2018-02-08 20:58:17 UTC

finally comitting some older code, urn split exporting, basic evdev input code (not wired in yet), and a simple split subtractor

Author: Kyle M
Revision Date: 2013-06-17 10:12:34 UTC

Add server_error_detach and server_error_reschedule_timeout config options, What happens when you get a server error during a reforward is now configurable:

* If server_error_detach is true the Reforwarder will be detached from the mainloop and the mapping the reforwarder was trying to forward will be scheduled to run whenever the Reforwarder is attached again.
* Otherwise the Reforwarder won't be detached and instead the mapping will be scheduled for another reforward attempt in server_error_reschedule_timeout seconds

Previously the Reforwarder was always detached when a server error happened, the problem with that is unless we eventually get a service-proxy-available signal for the Service Proxy the Reforwarder was registered for it would never get reattached.

(In the case of pulling the ethernet cable GUPnP/GSSDP will time out eventually and send a service-proxy-unavailable signal (which will also detach it), when that happens after replugging you will eventually get a service-proxy-available signal, but if you replugged after the first server-error but before the service-proxy-unavailable you'd never get a service-proxy-available)

Detaching is probably still the safer option in that you won't keep sending unfulfillable requests to the server, but it could stop anything from being reforwarded until a program restart which isn't great if it's supposed to be long-running.

More elegant solutions would be nice, maybe a way to signal to GUPnP/GSSDP to verify a resource still exists and update itself if not.

Author: Kyle M
Revision Date: 2011-10-16 20:59:37 UTC

oops, clear the info in .open as well

Author: Kyle M
Revision Date: 2011-10-16 18:37:11 UTC

* Add .info attribute to nodes, and a few useful bits of info on ELF and PE nodes
* Add Info Tab in gspec, still start on the Imported tab though