TMSR Log Summary - 9/18/2018

- Discussion on diana_coman’s options with using either ave1’s ADA implementation of UDP sockets with ASM inline 2 or asciilifeform’s light UDP sockets ib with C code and ADA wrapper
http://btcbase.org/log/2018-09-18#1851101

- Discussion on Asciilifeform’ s lib not supporting any options to be set for the UDP socket because the want is short code and without the unix the api should continue to make the same amount of sense as before
http://btcbase.org/log/2018-09-18#1851106

- Diana_coman had trouble coming up with reliable sets of errors that the UDP ops in linux may produce and received "all errors from IP may be returned by recv /send" from linux pages
http://btcbase.org/log/2018-09-18#1851126

- Discussion on how to approach UDP and eggog
http://btcbase.org/log/2018-09-18#1851134

- Raw packet support would be beneficial for the fyootoor - when the adaizing kernel is started
http://btcbase.org/log/2018-09-18#1851161

- Mircea_popescu made that point that, “if large bulks of memory have to cross call boundaries, something else is almost certainly fucking broken”
http://btcbase.org/log/2018-09-18#1851181

- Discussion on if varying the packet size and the packet count, rather than just the packet count is the best approach
http://btcbase.org/log/2018-09-18#1851191

- Mircea_popescu requested that diana_coman set up a testing harness to send a soup of all packets lengths from 1 to 65536 bytes each hour back and forth
http://btcbase.org/log/2018-09-18#1851201

- Discussion regarding mircea_popescu’s lulzfarm
http://btcbase.org/log/2018-09-18#1851224

- Discussion on what data details should result from diana_coman’s testing harness/ the minimum being survival rate, and latency
http://btcbase.org/log/2018-09-18#1851268

- Asciilifeform is open to approaches that will simplify the api
http://btcbase.org/log/2018-09-18#1851286

- Ave1’s work is appreciated to get rid of the C mess and for the ada inline assembler
http://btcbase.org/log/2018-09-18#1851297

- Diana_coman is going to set up the test harness and use asciilifeform's lib in smg.comms
http://btcbase.org/log/2018-09-18#1851313

- Asciilifeform made the point that - tcp, unixsockets, etc. should live in their own separate lib
http://btcbase.org/log/2018-09-18#1851315

- Asciilifeform did not close the socket in the demo and will correct this for next time
http://btcbase.org/log/2018-09-18#1851320

- There is a missing restrict pragma in the lib, pragma Restrictions(No_Implicit_Conditionals) that was not reinserted
http://btcbase.org/log/2018-09-18#1851323

- Asciilifeform made the point that, the Socket can be made a 'controlled type’ and it is possible to make the lib a 'generic' with created udptrons of different packet length runti
http://btcbase.org/log/2018-09-18#1851325

- Lobbes updated auctionbot eta 'steps to fruition' and is on track for October 31st delivery of the underlying table and field structure
http://btcbase.org/log/2018-09-18#1851330

- Discussion on ave1’s process of using the list of system calls (with more than 3 parameters), the man pages, and linux kernel code
http://btcbase.org/log/2018-09-18#1851332

- Discussion on ave1’s method to support different platforms by implementing the same module in different source directories and then selecting the directory to use in the project file and the use of python
http://btcbase.org/log/2018-09-18#1851342

- Mircea_popescu can't think of a thing that would need/want to use both tcp and udp as a toplevel - in response to ave1 working towards having different modules on top of a common base to support tcp / unix sockets
http://btcbase.org/log/2018-09-18#1851382

- Discussion on trad irc client that lobbes is building atop trinque 's design
http://btcbase.org/log/2018-09-18#1851398

Leave a Reply