28 Apr 2022
[Coke] ... golfing not easy, there's a lot in that test that seems to rely on things already done. 13:22
4 May 2022
Xliff m: role FontWidget { }; multi sub trait_mod:<is> (Attribute \a, :$font!) { a does FontWidget }; class A { has $.aa }; trait_mod:<is>(A.^attributes[0], :font); 12:58
Oops.
5 May 2022
Geth cro-http/cookie-tweaks: 4 commits pushed by (Jonathan Worthington)++ 13:46
cro-http: jnthn++ created pull request #176:
Assorted cookie improvements
13:47
cro-http/cookie-tweaks: 2dca3f556d | (Jonathan Worthington)++ | 2 files
Survive illegal cookie values in parser

Only extract the legal part, but quietly drop the rest. Closes
  github.com/croservices/cro-http/issues/171.
14:11
cro-http: a3065c932d | (Brian Duggan)++ | 2 files
Fix percent-encoding

from datatracker.ietf.org/doc/html/rfc3986#page-12
   2.1. Percent-Encoding
   ... A percent-encoded octet is encoded as a character
... (8 more lines)
14:16
cro-http: 56e47973f7 | (Jonathan Worthington)++ (committed using GitHub Web editor) | 2 files
Merge pull request #172 from bduggan/fix-encoding

Fix percent-encoding
[Coke] Haven't forgotten the golf for that windows failure, but it's challening. :| 15:23
aha: gist.github.com/coke/0c6518edb9d82...247a7a7bbf 15:41
there's a golfed error on windows for the output-socket.t test in Log::Timeline; the last line dies with "Cannot write to a closed socket" 15:42
if I output .WHERE on the $conn object before and after the last assignment to $conn, it changes. 15:44
updated, got it down to 8 lines of code, 2 debug statements and some comments 15:53
Will update ticket 15:54
done 15:55
jnthnwrthngtn [Coke]: Thanks, will try and get that to that soon. Putting in a bit of time on Cro things to get a release out in the coming weeks. 16:11
[Coke] Thanks. I'm wondering if closing the client connection here is somehow triggering a close on the server. 16:48
Geth cro-core: d48d216161 | (Jonathan Worthington)++ | 3 files
Fix IRI to URI conversion

If we have a valid IRI then it already has all percent encoding that a URI would need for things in ASCII range, and atop of that needs anything outside of ASCII range to be percent-encoded in order to form a valid URI. Thus explicitly pick out the non-ASCII range when deciding what to percent encode. Also, since this function is perhaps not very ... (5 more lines)
17:02
10 May 2022
lizmat clickbaits stackoverflow.com/questions/721809...memory-use 08:12
11 May 2022
Xliff m: my token aaa { <( 'aaa' )> }; say "baaab" ~~ /'b' <aaa> 'b'/ 21:30
19 May 2022
Geth cro-webapp/form-custom-components: 42f3cc2108 | (Jonathan Worthington)++ | 5 files
Trait and rendering for custom form components
20:00
cro-webapp/form-custom-components: 53727e9701 | (Jonathan Worthington)++ | 3 files
Support custom parse/serialize of custom component
21:01
20 May 2022
cro-webapp/form-custom-components: f97276f551 | (Jonathan Worthington)++ | 3 files
Pass further data for component renderer to use
20:57