Subject: KampalaMeetingLog-2017-10-19.txt
Date: October 19, 2017 at 5:24:02 PM PDT

peterthoeny [2:59 PM] 
hi

himazu [2:59 PM] 
hi

peterthoeny [3:00 PM] 
is there a way to see who is online? e.g. if @haj is here?

himazu [3:01 PM] 
on the Slack app
[3:01] 
at the left bar, you see who is online, I suppose

peterthoeny [3:01 PM] 
i am on the slack desktop app
[3:02] 
ah, yes, haj away

haj [3:06 PM] 
haj is now here :slightly_smiling_face:

peterthoeny [3:07 PM] 
hi @haj

haj [3:07 PM] 
Sorry for being late

peterthoeny [3:07 PM] 
we postponed the meeting by +90 min because @himazu needs to bering his son to school

haj [3:08 PM] 
Ah, ok

peterthoeny [3:11 PM] 
@haj: anything you’d like to discuss?

haj [3:11 PM] 
Just the regular stuff... I've had some time to work on TWiki in the last weeks

peterthoeny [3:11 PM] 
nice
[3:11] 
new perl version?

haj [3:12 PM] 
Not really.  Regularly I'm still on 5.24 (with Debian), but with perlbrew you can run several Perl versions in parallel

peterthoeny [3:13 PM] 
i meant working on twiki for new perl version
[3:13] 
what kind of twiki work did you do?

haj [3:14 PM] 
I think most stuff is covered for 5.24, and 5.26 didn't add much new issues

peterthoeny [3:14 PM] 
ok
[3:14] 
i gave a regex talk at a local conference
[3:15] 
with some twiki plug

haj [3:15 PM] 
I fixed some unit tests so that the suite now almost passes - there's just the bin2hex stuff of SpreadSheetPlugin

peterthoeny [3:15 PM] 
showing sample code of twiki using regex
[3:15] 
what’s the issue with bin2hex?

haj [3:16 PM] 
The bin conversion assumes 32-bit integers
[3:16] 
This can't be covered in the test cases because TWiki apps have no access to Perl's %Config hash

peterthoeny [3:17 PM] 
ah

haj [3:17 PM] 
I doubt the sense of these conversions anyway... TWiki topics live longer than architectures
[3:17] 
Problem only arises when converting negative numbers

peterthoeny [3:17 PM] 
so the conversion code should be fixed?

haj [3:18 PM] 
I don't know... TWiki users can't be assumed to know which architecture their system is running on[3:18] 
Topics live longer than machines
[3:19] 
The conversion is just done with printf, which nowadays assumes 64bit integers

peterthoeny [3:20 PM] 
hmm, i don;t see a bin2hex
[3:20] 
there is a dec2hex

haj [3:21 PM] 
Ok, let me check...
[3:21] 
I'm running the suite as a "conventional" Perl test case with Test::More now

peterthoeny [3:21 PM] 
DEC2HEX used sprintf

haj [3:22 PM] 
DEC2BIN
[3:23] 
Does a sprintf padded with '0' if size is greater than needed - but this only works for positive integers

peterthoeny [3:24 PM] 
ah

haj [3:24 PM] 
Negative integers have always full length of 64 0/1 values

peterthoeny [3:24 PM] 
any use case for converting negative numbers?

haj [3:24 PM] 
I don't have any

peterthoeny [3:25 PM] 
ah, the test case has a negative number
[3:25] 
hence the assumption on integer size

haj [3:26 PM] 
Yes.  DEC2HEX and DEC2BIN are affected
[3:26] 
Expected: $DEC2HEX(-54)=|FFFFFFCA|

peterthoeny [3:26 PM] 
i think it is safe to remove the negative numbers tests

haj [3:26 PM] 
Got: $DEC2HEX(-54)=|FFFFFFFFFFFFFFCA|
[3:27] 
I agree

peterthoeny [3:27 PM] 
in DEC2BIN and DEC2HEX
[3:27] 
i can do that

haj [3:28 PM] 
That's Bugs:Item7640, by the way

peterthoeny [3:28 PM] 
ok
[3:28] 
anything else?
[3:28] 
if not i let you go sleep :slightly_smiling_face:

haj [3:30 PM] 
I can now also run automated tests to check for warnings (like the unescaped { in regexes)

peterthoeny [3:31 PM] 
nice

haj [3:32 PM] 
Plus a general compilation test - makes sense to cycle through Perl versions with perlbrew
[3:33] 
That doesn't catch the gazillion of evals, of cours
[3:33] 
e

peterthoeny [3:33 PM] 
useful stuff

haj [3:34 PM] 
I'll commit that when the fine tuning is done
[3:35] 
I've also closed some bugs, or suggested to close others, in their topics
[3:36] 
Some of them are years old

peterthoeny [3:36 PM] 
thank you @haj!

haj [3:37 PM] 
The notorious "insecure dependency" (Item7721) has vanished with current Perl and Linux versions

peterthoeny [3:37 PM] 
oops, i get a server error in my trunk install after an svn update

haj [3:38 PM] 
I hope I didn't break anything... my own SVN checkout works quite fine
[3:39] 
Perhaps I missed a commit

peterthoeny [3:39 PM] 
no, i suspect a file permission issue in bin

haj [3:40 PM] 
Ah... I finally got rid of that problem by switching to the PSGI engine
[3:40] 
BTW, that's another work in progress: The PSGI stuff allows to automate some of the "manual" tests

peterthoeny [3:41 PM] 
yup, the svn update changed -rwxr--r-- to -rwxrwxr-x
[3:42] 
which broke the script execution
[3:42] 
fixed it back, now it works

haj [3:43 PM] 
Phew!
[3:43] 
Another discovery: For some plugins I'm running Git and Subversion in parallel

peterthoeny [3:44 PM] 
i remember another instance like this, hence fast to fix :slightly_smiling_face:
[3:44] 
apache security thingy

haj [3:44 PM] 
Branching is annoying in SVN, pretty easy and dead fast in Git...

peterthoeny [3:48 PM] 
ok, spreadsheetplugin tests fixed, and new version uploaded

haj [3:49 PM] 
Two weeks ago I forgot to talk to Hideyo-san about Item7796 (Make CGI::Carp work from TWiki's distribution)

peterthoeny [3:49 PM] 
ok, i have to sign off for a while unless you have something

haj [3:49 PM] 
Nope... I have plenty of time right now :slightly_smiling_face:

peterthoeny [3:50 PM] 
ok, ttyl @haj!

peterthoeny [4:33 PM] 
hi @himazu

himazu [4:35 PM] 
Hi @peterthoeny
[4:35] 
I was catching up the conversation
[4:37] 
How should we move forward?

peterthoeny [4:44 PM] 
hi @himazu

himazu [4:44 PM] 
hi

peterthoeny [4:44 PM] 
shall we do informal meeting?

himazu [4:44 PM] 
sure

peterthoeny [4:45 PM] 
i don’t have much on twiki

himazu [4:45 PM] 
I finally switched from iPhone 5s to iPhone 8

peterthoeny [4:45 PM] 
as mentioned to harald, i gave regex talk

himazu [4:46 PM] 
and I got BeatsX BT earbuds

peterthoeny [4:46 PM] 
(oh, i just discovered if you press up key you can fix your latest post) - like that (edited)

himazu [4:47 PM] 
taking train by Apple Pay

peterthoeny [4:47 PM] 
ah
[4:47] 
i am now studying cryptocurrencies
[4:47] 
interesting field
[4:47] 
future

himazu [4:48 PM] 
agreed
[4:48] 
cryptocurrencies are inevitable

peterthoeny [4:48 PM] 
ethereum is interesting

himazu [4:48 PM] 
it sure is

peterthoeny [4:49 PM] 
in bitcoin, a node in a blockchain is only a container for numbers
[4:49] 
in ethereum, a node is a javascript virtual machine

himazu [4:49 PM] 
still not 100% convinced that any of them can work really well at scale

peterthoeny [4:49 PM] 
which makes it very powerful
[4:50] 
bitcoin has an scaling issue: verifying a transaction can take 10-30 min
[4:50] 
and gets worse over time the longer the blockchain gets

himazu [4:51 PM] 
still, you can do transaction in a distributed manner without trusting anybody, is a powerful achievement

peterthoeny [4:51 PM] 
that is a severe limitation for use as a ubiquitous cryptocurrency
[4:51] 
another one is currency fluctuations
[4:52] 
there are ways to fix those issues, which i am currently looking at

himazu [4:52 PM] 
long term, i’m optimistic about somebody somehow will come up with a good enough system

peterthoeny [4:52 PM] 
it will be a big competition to traditional banks
[4:53] 
unless banks get on the cryptocurrency train

himazu [4:53 PM] 
competition to governments
[4:53] 
who are issuing currencies

peterthoeny [4:53 PM] 
yes and no on gov
[4:54] 
yes for fiat currency

himazu [4:54 PM] 
fiat currency is ultimately based on people’s trust

peterthoeny [4:54 PM] 
similar to banks, if gov gets on the cryptocurrency train, it can be beneficial to them
[4:55] 
i think in the future, cryptocurrency will be more trusted than fiat currency

himazu [4:56 PM] 
i wonder if any government can compete with a truly global and well-working digital currency

peterthoeny [4:56 PM] 
not with global, but yes if the cryptocurrency is tied to local currency

himazu [4:57 PM] 
think about places like Zimbawe
[4:57] 
people don’t trust local currency
[4:57] 
and Brazil while it was in a bad inflation

peterthoeny [4:58 PM] 
yes, there us$ or another foreign currency and also cryptocurrency is seen as more trustworthy

himazu [4:59 PM] 
I suppose it’s possible for a digital currency to become more trusted than US$

peterthoeny [4:59 PM] 
indeed

himazu [4:59 PM] 
or any other government backed currency

peterthoeny [4:59 PM] 
that is because the us$ is a fiat currency not much backed by gold

himazu [4:59 PM] 
right
[5:00] 
and it’s not practical to go back to the gold standard

peterthoeny [5:00 PM] 
yes
[5:01] 
if interested i recommend looking into the tech details of ethereum
[5:01] 
i watched a few videos, very informative

himazu [5:02 PM] 
I hope I have time to do that
[5:02] 
will try

peterthoeny [5:04 PM] 
i will participate in an art show in a gallery the whole november
[5:04] 
on sat i will install the prints

himazu [5:05 PM] 
how many will you install?

peterthoeny [5:05 PM] 
3 large prints and 6 smaller ones

himazu [5:06 PM] 
how do you print those?

peterthoeny [5:06 PM] 
online order
[5:06] 
metal prints and acrylic glass prints
[5:06] 
more expensive, but look much nicer
[5:06] 
the large ones are 1m wide, smaller ones 50 cm

himazu [5:07 PM] 
have you tried Fracture?

peterthoeny [5:07 PM] 
no
[5:07] 
what is that?

himazu [5:07 PM] 
it’s sponsoring many podcasts
[5:08] 
https://fractureme.com
fractureme.com
Fracture | Glass Photos
Fracture allows you to print your pictures directly on pure glass.

peterthoeny [5:08 PM] 
looking…
[5:08] 
that looks interesting
[5:08] 
seems similar to acrylic glass print
[5:09] 
also printed on backside?

himazu [5:09 PM] 
is “acrylic glass” plastic or glass?

peterthoeny [5:10 PM] 
ah, $85 for 20 inches print, pretty much the same as acrylic glass
[5:10] 
acrylic is a type of plastic
[5:10] 
looks and feels like glass but is plastic, lighter than glass
[5:11] 
when you knock on it you know it is plastic

himazu [5:11 PM] 
ok. Fracture seems to be real glass

peterthoeny [5:11 PM] 
yes, just looked at the faq
[5:12] 
hm, maybe i should order one and compare to acrylic
[5:13] 
maybe the name fracture is an indication that the glass fractures easily :wink:

himazu [5:13 PM] 
I hope not

peterthoeny [5:13 PM] 
do you have anything twiki related?

himazu [5:13 PM] 
not really

peterthoeny [5:14 PM] 
neither me
[5:14] 
i have some new twiki consulting work coming up, not sure yet if this involves code enhancements, or just tml work
[5:14] 
if nothing else shall we close the meeting?

himazu [5:15 PM] 
yep

peterthoeny [5:15 PM] 
ok
[5:16] 
i’ll post the logs and minutes as usual
[5:16] 
ttyl @himazu -san

himazu [5:17 PM] 
ttyl