01:39:51  * plutoniixquit (Quit: Leaving)
02:23:57  * plutoniixjoined
02:36:12  * cloudshuquit (Quit: Connection closed for inactivity)
03:34:34  <trungl-bot>Tree closed by buildbot@chromium.org: closed (https://ci.chromium.org/p/v8/builders/ci/V8%20Mac64%20ASAN/22194 from 83fa17178168dcb767527417fd91e49d3e908bb8)
08:08:18  <trungl-bot>Tree opened by yangguo@google.com: open
08:08:33  * Net147quit (Quit: Quit)
08:09:57  * Net147joined
09:59:48  * plutoniixquit (Quit: Leaving)
11:06:19  <trungl-bot>Tree closed by buildbot@chromium.org: closed (https://ci.chromium.org/p/v8/builders/ci/V8%20Mac64%20ASAN/22200 from 52dc1d799f49f57846f08a4b5562e52c638dd064)
11:46:35  <trungl-bot>Tree opened by mslekova@google.com: open (DEPS reverted)
12:14:46  * plutoniixjoined
12:37:51  * AtumTjoined
13:46:13  <jnes>caitp: so what happens is that i hit the bp, rcx is 0x1fafde147e00, i step once to to run `lea -0x7(%rip),%rbx # 0x1fafde147e00`, rbx is 0x1fafde147e07 !?!?!? dafuck...
13:53:45  <jnes>oh hold on.. i might have mixed up the registers.
13:57:38  <caitp>when you first break, can you evaluate `_v8_internal_Print_Code((void*)$pc)` (or $rip or whichever gdb lets you do?)
13:58:22  <caitp>and stick the output in a gist or something
14:02:39  <jnes>okey https://gist.github.com/sktt/be6ca7034c7ee9a2009e471db088d230
14:03:11  <jnes>Print_Code includes 0xcc instructions though... gdb does show those.. nor does xxd
14:06:11  <jnes>like so: https://gist.github.com/sktt/cee0eb0a15f528386e2c3ef8b9a47096
14:07:11  <jnes>so it looks like gdb inserts int3 for my bps at `...e00` and `...e16`. i guess that somehow breaks the lea relative to rip
14:09:33  <jnes>i would however assume that xxd is that ground truth.. which does not include those 0xcc bytes.. https://gist.github.com/sktt/cb694979debdec7d0f7bc7699bf44195
14:12:05  <jnes>s/gdb\ does/gdb\ doesnt/
14:34:36  <jnes>second try it works without any problem. am i drunk?
15:03:37  <caitp>those are your debug breaks
15:03:53  <caitp>they are actually there and you want them to show up so you can tell where you're breaking
15:04:38  <caitp>can you show the other information provided by print_code?
15:04:44  <caitp>like, the code header
15:05:44  <caitp>(I assume xxd omits them because it knows it's inserting them?)
15:15:42  <jnes>that's so crazy, how can xxd omit them?
15:18:45  <caitp>I just mean from the code listing
15:18:54  <caitp>beats me
15:29:46  * plutoniixquit (Quit: Leaving)
16:00:59  <jnes>but yea,, using hbreak seems to fix the problem :)
16:08:25  <caitp>I guess inserting int3s messes with RIP offsets
16:08:31  <caitp>makes sense when you think about it
16:09:56  <caitp>whereas hbreak wouldn't use int3 I guess :p
16:10:48  <caitp>I didn't know that was a thing in gdb, my gdb-fu is weak
16:31:34  * RT|Chatzillaquit (Quit: ChatZilla [Firefox])
16:32:30  <jnes>caitp: didnt know either haha
17:15:42  <trungl-bot>Tree opened by clemensh@google.com: open - Deepti looking at flakes
17:23:45  <trungl-bot>Tree opened by clemensh@google.com: open - fix for flakes incoming
17:46:54  <trungl-bot>Tree closed by buildbot@chromium.org: closed (https://ci.chromium.org/p/v8/builders/ci/V8%20Linux64%20-%20verify%20csa/12144 from b96c127f23587294b1be94bcc5502a451b8eca33)
17:49:55  <trungl-bot>Tree opened by clemensh@chromium.org: open - fix landed
17:50:55  <trungl-bot>Tree closed by buildbot@chromium.org: closed (https://ci.chromium.org/p/v8/builders/ci/V8%20Linux64/30577 from 3a16ee8708894ccc684e31341d41e4f9fcede835)
17:59:59  <trungl-bot>Tree opened by bbudge@google.com: open
22:26:45  * RT|Chatzillajoined