27 Jun 2022 |
timo |
but i may not have time to measure things properly today |
08:37 |
|
nine |
482 spectest files passing on RakuAST. 302 without doing a non-RakuAST spectest run first |
09:10 |
|
timo |
\o/ |
|
|
lizmat |
nine++ |
11:15 |
|
|
And yet another Rakudo Weekly News hits the Net: rakudoweekly.blog/2022/06/27/2022-...-seasoned/ |
14:21 |
|
vrurg |
I wonder if valgrind is capable of helping me to trace down likely memory issues. In my work project I get errors like Buf[int8] not matching Blob, VMNull from symbol lookups, etc. Could be caused by double-free, wrong pointers, whatever. But for now I have no clue what to look at. |
14:43 |
|
timo |
it's worth a try at least, but i can't make any guarantees |
15:22 |
|
28 Jun 2022 |
Nicholas |
Good *, * |
05:36 |
|
japhb |
Good *, Nicholas |
05:37 |
|
29 Jun 2022 |
Nicholas |
Good *, * |
05:56 |
|
nine |
It's 18 degrees outside and raining. That _is_ a good morning :) |
06:20 |
|
Nicholas |
It's >25 and not. |
06:22 |
|
|
The cellar is cool, but the rest of the house is about 25 too. |
|
|
|
I have plenty of coffee, but I'm running low on tea. (Shops are open) |
06:23 |
|
Voldenet |
be glad you're not in the US, 60 degrees is considered normal there |
06:35 |
|
|
absolute madness |
06:36 |
|
[Coke] |
... of all the madness going on here ATM, I can live with farenheit. :) |
13:29 |
|
nine |
That's now 484 spec test files passing and 342 without doing a non-RakuAST spectest run first. Btw. to me that second number is the more important one as I'm working mostly on getting precompilation of those test dependencies working right. |
14:55 |
|
[Coke] |
I have a moar/openssl/libcrypto bt at github.com/jnthn/p6-io-socket-asyn...issues/55; please let me know if I can get more information to go with the bt there to be helful. |
19:48 |
|
|
*helpful\ |
|
|
30 Jun 2022 |
Nicholas |
good *, * |
06:22 |
|
timo |
[Coke]: i tend to like compiling moarvm with -O 0, but here the trace goes so far into the native library that perahps just getting the actual arguments passed to libssl are a good starting point? can you get the debug symbols for libssl and libcrypto? |
09:15 |
|
|
can you go to the last MVM_ function and ask for `call MVM_dump_backtrace(tc)`? |
09:19 |
|
[Coke] |
I will try this evening, thanks for the suggestion. doubt I can easily get debug versions of the installed ports. |
12:17 |
|
timo |
ports isn't a thing that compiles stuff itself, right? |
12:19 |
|
[Coke] |
If I understand correctly, it can be binaries, but it doesn't need to be binaries. I could dig and see if I can re-install/build with debug |
12:21 |
|
1 Jul 2022 |
Nicholas |
[* GOOD *] |
05:35 |
|
lizmat |
indeed! |
05:42 |
|
timo |
my recommendation for the day: when you copy your root filesystem, perhaps make extra sure selinux attributes are kept, or at the very least turn enforcing off at boot if you'd like to be able to log in |
23:47 |
|
japhb |
My recommendation: Never copy your root filesystem, aside from maybe system backups. :-) |
23:54 |
|
2 Jul 2022 |
timo |
would have been no problem if btrfs send had worked :) |
00:18 |
|
|
adding the new device to the btrfs volume, then removing the old device has been advised against for some reason |
|
|
|
i'm running restorecon now, but setroubleshootd is flipping its absolute shit |
00:19 |
|
|
i have been wanting to get rid of setroubleshootd for a longlong time |
00:24 |
|
nine |
timo: I have always migrated to new (local) storage by adding the new device and removing the old. Worked flawlessly every time. |
07:29 |
|
|
We are now at 500 passing spec test files on RakuAST :) |
16:19 |
|
|
And the latest batch of commits fixed those annoying infinite loops in spec tests |
16:20 |
|
|
358 without doing a non-RakuAST spectest run first! |
17:01 |
|
|
So the difference went down from 180 to 142. |
17:02 |
|
|
and 503 |
17:18 |
|
japhb |
Go nine, go nine! |
19:44 |
|
timo |
super cool |
21:06 |
|
vrurg |
Any potential difference in how well code would be inlined via `boot-code-constant` and `raku-meth-call` dispatchers? Considering that actual code bodies are gonna be basically the same up to use of arguments vs. attributes. |
22:09 |
|
3 Jul 2022 |
nine |
Now 422 passing without doing a non-RakuAST spectest run first! Narrowing the gap to just 81 |
11:18 |
|
timo |
\o/ |
|
|
nine |
2 more fixes and we are at 506/487 respectively, closing the gap to just 19! That's now more spec test files passing without requiring precompilation than we did pass with precompilation just 3 days ago :) |
14:06 |
|
|
And 512 in both cases! |
16:20 |
|
|
jnthn: there's no more need to run a normal spectest before doing a RakuAST count :) |
|
|
japhb |
nine++ |
16:55 |
|
timo |
you mean the gap is now at 0? |
17:07 |
|
nine |
timo: yes |
17:09 |
|
timo |
fantastic news |
|
|
nine |
In other words RakuAST compiles fully functional Test, Test::Utils and friends |
17:10 |
|