github.com/moarvm/moarvm | IRC logs at colabti.org/irclogger/irclogger_logs/moarvm
Set by AlexDaniel on 12 June 2018.
vrurg MasterDuke++ 00:43
Geth MoarVM/unicode-13-update: 2168824d12 | (Timo Paulssen)++ | tools/UCD-download.p6
ucd-download: fix folder names (don't drop .0 from path)
 02:37
MoarVM/unicode-13-update: 3b7a04af3d | (Timo Paulssen)++ | tools/ucd2c.pl
add support for unicode 13 "Tangut Ideograph Supplement"

and the CJK Ideograph Extension G, which is also new
MoarVM/unicode-13-update: 73c314127f | (Timo Paulssen)++ | 2 files
attempt at unicode 13 data in the unicode db

the automated scripts have to be changed a little bit more, because emoji-data.txt isn't in emoji-13/ any more.
Also, I haven't checked if everything is still correct, especially the tangut ideograph supplement block support could possibly be wrong?
timotimo i think nobody started on this yet, so i went ahead and did the simplest pieces
timotimo samcv: you may want to look over my changes, and perhaps you also know why the downloader had to be changed for the folder names on the ftp that apparently recently changed from "5" to "5.0" and such? and why ever the emoji-data.txt went from emoji-13 to ucd.zip or something? 02:38
i manually copied the emoji-data.txt file over into the emoji-13 folder to get ucd2c.pl to finish, but there's surely a more general approach here 02:39
samcv timotimo, hmm. i'll look at it before lunch today 06:52
samcv timotimo, i will check it out now 19:23
samcv timotimo, oh the ftp names sometimes change... at least it has happened before. I'm going to run the ucd2c.pl and see what crops up (before your changes) just to double check things. maybe add more error checking depending 19:30
samcv I am going to check out Unicode spec chapter 4 as the "die" message my past self wrote: 19:36
processing UnicodeData.txt...<Tangut Ideograph Supplement> encountered. Make sure to check www.unicode.org/versions/Unicode10.0.0/ch04.pdf for Name Derivation Rule Prefix Strings
Also you will likely have to make a change to MVM_unicode_get_name() and add a test to nqp at ./tools/ucd2c.pl line 1916
samcv timotimo, ok, i've made some improvements to the script. I am writing into the branch you pushed. I removed your changes to the script, and i want to make sure that the error that happens when it's run is clear enough for you to make the right choice 20:35
Geth MoarVM/unicode-13-update: a088b2e041 | (Samantha McVey)++ | tools/ucd2c.pl
unicode: Add a new guard for prefix string's

When new prefix strings or extensions are added we should make sure that we die if detect an unknown one.
In addition, ensure that there is much more robust explaination about Name Derivation Rule Prefix Strings. This will add a clearer error message when detected, as well as a bunch of comments to hopefully make it easier for others to deal with making changes to the script in future unicode versions.		 20:36
MoarVM/unicode-13-update: 8e20187be5 | (Samantha McVey)++ | tools/ucd2c.pl
unicode: update path of emoji-data file

In version 13 of unicode the emoji-data.txt file is in a new location.
samcv tldr, your changes were incorrect, but it's mostly because the script wasn't clear enough and not enough comments in this section. Please run it again and I am hoping I have fixed the situation (goal being to make it easier for people not me to update unicode versions) 20:37
lizmat samcv++ 20:39
moon-child a while ago, I started working on a library to embed rakudo. It mostly worked fine, although it wasn't very featureful. Today, I tried running its tests, and I get a '*** stack smashing detected ***: terminated' 22:48
basically the way it worked is there was a raku function that evaluated strings passed to it, constructed a native c object based on that, and returned it. I get that error after returning the object
jnthn Probably somewhere there's a buffer overflow. I suggest using either valgrind or ASAN (on a debug build) to get more information on where it happens. 22:54
moon-child should I recompile my moar with debug syms then? 22:55
jnthn Yes
moon-child valgrind says nothing 23:13
timotimo could be a security feature that's not happy about Nativeχall 23:14
moon-child will try rebuilding with asan
timotimo: could be...I doubt it, though 23:15
only thing I can think of is w^x, but I don't think I have that on
timotimo stack smashing would probably be stack cookies or something 23:16
jnthn I think (though hazy memory) it's something that gets compiled in 23:17
moon-child one thing I was thinking was that it doesn't know the abi convention for returning a struct of that size. But that seems like an odd thing to regress on
jnthn stackoverflow.com/questions/134567...d/50690687 23:18
timotimo i don't think we have any support for returning structs ATM 23:19
moon-child wait, no, nvm, it was a pointer to a struct
so it couldn't be that anyway
jnthn Maybe you got a compiler upgrade, or something changed such that the stack protector is turned on
moon-child but if the stack protector is triggering, the code is probably doing something wrong anyway 23:20
jnthn Indeed. 23:21
Are you running MoarVM HEAD, btw?
moon-child whatever is the most recent thing that nqp gen-moar pulls 23:22
jnthn I sw this go by recently: github.com/MoarVM/MoarVM/commit/22...0e55380bab

Powered by Cro and the Raku® Programming Language.

Please report any issues / comments / feature requests as an issue on App::Raku::Log.

Thank you!