|
04:06
Geth joined,
lizmat joined
05:43
ShimmerFairy left
05:44
ShimmerFairy joined
07:24
[Tux] left
07:28
[Tux] joined
|
|||
| patrickb | In the debugger, I sometimes encounter a situation where the file being executed has the gen-cat SETTING as $?SOURCE but the file is not part of the gen-cat source. This seems to be the case for files of the compiler. E.g. src/Perl6/Metamodel/BaseType.nqp | 08:22 | |
| Any theories as to why this happens? | 08:23 | ||
| And ideas of how to best deal with this? | |||
|
10:15
ShimmerFairy left,
ShimmerFairy joined
10:22
librasteve_ joined
13:56
kurahaupo left,
kurahaupo joined
|
|||
| timo | could the $?SOURCE be wrongly picked up from the setting because the setting is "outside everything"? do nqp also get their $?SOURCE set up correctly already? | 14:37 | |
|
15:20
apogee_ntv left
15:21
apogee_ntv joined
|
|||
| lizmat | NQP doesn't know about $?SOURCE | 17:45 | |
| (yet, anyway) | |||
| nqp -e 'nqp::say($?SOURCE)' | 17:46 | ||
| doesn't show anything, but neither does: | |||
| nqp -e 'nqp::say($?FLABBERGASTED' | |||
| nqp -e 'nqp::say($?FLABBERGASTED)' # :-) | |||
| timo | right, the very common issue of "anything not actually declared gives you NQPMu" which can be very fun in situations of a type | 17:47 | |
| patrickb | Does the P6 Compiler has the setting as outer? If so that would explain why I see it as SOURCE. I would then just need to be a bit smarter in determining whether the source applies or is too far out. | 20:45 | |
| timo | I'm not exactly sure | 20:46 | |
| Geth | nqp/profile_sql_output_split_more_stuff: 6b443a0849 | (Timo Paulssen)++ | src/vm/moar/HLL/Backend.nqp Fix remaining issues with sql output after refactor |
21:33 | |
| timo | oops it ate the commit message huh | ||
| Geth | nqp/profile_sql_output_split_more_stuff: 9171877d8e | (Timo Paulssen)++ | src/vm/moar/HLL/Backend.nqp Refactor sql output, split more of the statements we can output nine++ tested sql profiler output with core setting compilation using RakuAST, and the resulting profile had an "allocations" insert statement that was 221 million characters long, which sqlite3 didn't like parsing as a single statement. This patch also has some attempts for re-using lists more. |
21:34 | |
| timo | that's better. | ||
|
22:04
apogee_ntv left
22:05
apogee_ntv joined
22:27
djinni` left,
djinni` joined
|
|||
| Geth | nqp/profile_sql_output_split_more_stuff: 42659f7cb0 | (Timo Paulssen)++ | src/vm/moar/HLL/Backend.nqp profiler sql: also split routines insert statements |
22:33 | |
| nqp: timo++ created pull request #845: Split all "insert" statements in profiler sql output |
22:34 | ||