Where communities thrive


  • Join over 1.5M+ people
  • Join over 100K+ communities
  • Free without limits
  • Create your own community
People
Repo info
Activity
  • Mar 07 17:29
    @directhex banned @CharlieIsHere
  • Oct 11 2018 17:23
    @directhex banned @deleteaccount1234567890
Calvin Buckley
@NattyNarwhal
i wanna see consistent graphs again, i should just allocate those cores back ;_;
Jay Krell
@jaykrell

mono-codeman.c:

        /* Try to allocate code chunks next to each other to help the VM */
        ptr = NULL;
        if (last)
            ptr = codechunk_valloc ((guint8*)last->data + last->size, chunk_size);
        if (!ptr)
            ptr = codechunk_valloc (NULL, chunk_size);
        if (!ptr)
            return NULL;

Which VM?
The kernel virtual memory manager or the Mono virtual machine?
The kernel virtual memory manager surely will help itself most with NULL there.
I imagine it is the map32bit stuff for amd64, so the Mono virtual machine.
Comments should/could be clearer imho.
As well, only systems that use the map32bit stuff should have this logic, I think.

Alexander Köplinger
@akoeplinger
@NattyNarwhal I bumped the timeout to 360 minutes
Jay Krell
@jaykrell

arm/mips/powerpc, why the 32bit here?

        ss_trigger_page = mono_valloc (NULL, mono_pagesize (), MONO_MMAP_READ|MONO_MMAP_32BIT, MONO_MEM_ACCOUNT_OTHER);
        bp_trigger_page = mono_valloc (NULL, mono_pagesize (), MONO_MMAP_READ|MONO_MMAP_32BIT, MONO_MEM_ACCOUNT_OTHER);

I assume mips implies 32bit.
I know arm does (arm64 is separate).
PowerPC I guess was dual purpose, but is now 64bit only?, so it seems the only user like this (I know amd64 has a different use).

On Windows there could be some hypothetical creation of a truncatable address,
for a 32bit debugger to poke at a 64bit target. Or to push the address
over a transport that truncates to 32bits even if to a 64bit target.
But mono does not target Windows with these architectures (it did/does run on them!).

But still, it is only for powerpc64 seemingly.

mono/mono#16469 ?

Jay Krell
@jaykrell
Anyone consider an interpreter for the JIT IR? i.e. because it has some optimization? Or transform.c is similar enough (I haven't looked)? Or the IR is not a good target for an interpreter?
Mikkel Kruse Johnsen
@mikkeljohnsen
@lambdageek Did you look more on #16320 ?
I don't understand why this bug is not a big problem for other users. Not being able to execute another program from Mono, is a huge problem.
Aleksey Kliger (λgeek)
@lambdageek
@mikkeljohnsen it's some interaction with GSharpKit - Process.Start works for normal console apps in Mono. I don't remember if I already suggested a workaround - try setting MONO_THREADS_SUSPEND=preemptive before running your app.
Mikkel Kruse Johnsen
@mikkeljohnsen
@lambdageek Setting MONO_THREADS_SUSPEND=preemptive has no effect.
@lambdageek But what could GtkSharp do that would make mono unable to use "fork" ?
Aleksey Kliger (λgeek)
@lambdageek
I'm not sure yet.
Mikkel Kruse Johnsen
@mikkeljohnsen
I can't even use DllImport to launch a script that uses the "open" command. Somehow it seems that macOS is preventing the entire app from spawning a new process
Phil Jaenke
@rootwyrm
Oooooooooof.
I have an unhappy idea there. You have truss on the system, yes?
Sorry, that Q was for @mikkeljohnsen
Nikolay Sivov
@nsivov
@jaykrell regarding warnings, I see a lot of -Wsign-compare in "Windows x64 C++" logs, do you think it's worth cleaning up? Personally I'd like to keep everything warning-free for regularly used toolchains
Jay Krell
@jaykrell
I gave up and pragma disabled a lo of this, but missed some.
Nikolay Sivov
@nsivov
alright then
Jay Krell
@jaykrell
But it would be nice to have fewer warnings. Some I disabled is pointer type mismatches for example, Maybe whittle down the pragmas gradually?
@nsivov see here C:\s\mono2\msvc\msvc-disabled-warnings.h and maybe try whittling it more down than up?
Nikolay Sivov
@nsivov
ok, that's a lot
Phil Jaenke
@rootwyrm
Now I just have to figure out how to accept proposed changes on Github... which is not as easy as one would think.
Calvin Buckley
@NattyNarwhal
oh right, i need to go into SST.
also probably should de-WIP that PR because i think it works even though i have nfc how to activate the crash reporter
Phil Jaenke
@rootwyrm
Hey, on the plus side, not gonna break all your stuff (more than it's already broken.)
Calvin Buckley
@NattyNarwhal
oh, i can't give this LPAR more units. because this half-configured LPAR is in the way
moderate chuckle
5 second later: time to find more BUILD SERVERS
Phil Jaenke
@rootwyrm
Find me some while you're at it?
On the plus side...
root@teldrassil:/usr/ports/lang/mono6.0 # ldd /usr/local/bin/mono
/usr/local/bin/mono:
        libz.so.6 => /lib/libz.so.6 (0x8006b5000)
        libutil.so.9 => /lib/libutil.so.9 (0x8006cf000)
        libiconv.so.2 => /usr/local/lib/libiconv.so.2 (0x8006e6000)
        libm.so.5 => /lib/libm.so.5 (0x8007e4000)
        libinotify.so.0 => /usr/local/lib/libinotify.so.0 (0x800816000)
        libthr.so.3 => /lib/libthr.so.3 (0x800821000)
        libc.so.7 => /lib/libc.so.7 (0x80084c000)
Calvin Buckley
@NattyNarwhal
yeah if i can get my PR merged then libexecinfo will show up there too
if you wanna halfass it though, i think adding -lexecinfo is enough for now
Phil Jaenke
@rootwyrm
I'm gonna test that next... that was testing the fix in corefx for libinotify shim.
Now we can fail tests more spectacularly on FreeBSD because we're missing inotify functions! Yay!
Phil Jaenke
@rootwyrm
Darn... of course it couldn't be as simple as ac_cv_header_execinfo_h=yes :(
Calvin Buckley
@NattyNarwhal
more -I for the preprocessor god
Phil Jaenke
@rootwyrm
MOAR AUTOCONF MANGLING
Mikkel Kruse Johnsen
@mikkeljohnsen
@rootwyrm What do you mean by "truss" ?
Alexander Köplinger
@akoeplinger
@rootwyrm you're probably talking about the change I proposed on the TZ PR, I didn't use the github feature for proposed changes so it's just a comment :smile: you need to manually make the change and push it to the branch
Calvin Buckley
@NattyNarwhal
@mikkeljohnsen truss is a syscall tracer you can attach to programs
Phil Jaenke
@rootwyrm
@akoeplinger maybe Github should figure that out instead of telling me you proposed changes <sigh> But hey! I'm learning C#.
Alexander Köplinger
@akoeplinger
:wink:
Phil Jaenke
@rootwyrm
Fix committed, notes taken for how to improve the test at a later date. :)
Alexander Köplinger
@akoeplinger
@rootwyrm two more changes, looks good otherwise :wink:
Phil Jaenke
@rootwyrm
Augh, stop finding things I missed! ;)
Actually, let me noodle a little on that first one, because that's the only Central Standard Time test case.
Alexander Köplinger
@akoeplinger
I'm pretty sure that changing it to America/Chicago should be fine too
Phil Jaenke
@rootwyrm
Theoretically, unless mobile doesn't have that, which I honestly don't know.
Alexander Köplinger
@akoeplinger
doubtful. do you know when the TZ deprecation happened?
Phil Jaenke
@rootwyrm
Unfortunately, no... it's really poorly documented. But that only applies to the IANA (Olson) database.