Results 1 to 6 of 6

Thread: PET CTDs; selecting expensive parent particle as child particle/instability crash

  1. #1
    Basic Member Toyoka's Avatar
    Join Date
    Oct 2013
    Location
    Current position co-ordinates non-tracable. Unknown space.
    Posts
    833

    PET CTDs; selecting expensive parent particle as child particle/instability crash

    Details: Encountered several possibly-related CTDs concerning the PET (Particle Editor), described below. (Greyed-out text refers to previous assumptions)

    1 PET can be unstable and may cause a delayed CTD (after system hangs for some seconds) when browsing and selecting random particles within particle asset browser, to be assigned as a child particle of another particle (possibly cache-related, as this does not happen more than once within the same system boot session; have yet to test during multiple system boot sessions). Edit: After some testing it seems this (and subsequent) CTD(s) actually happens as a result of browsing for and selecting a particle to be a child particle of itself (ie. selecting particle "X" [or particles containing particle "X"] as a child particle of particle "X"). The delay experienced during first crash may have been due to external factors within the workshop tools, possibly vconsole-related (unfortunately, no vconsole log to verify this).

    2 PET immediately causes CTD when selecting an expensive (non-compiled version, ie. .vpcf) parent particle (ie. one with many child particles, such as an "*_ambient" particle where "*" is a hero).

    2b Same crash cause and result as above, but produces different Exception Code (0xC0000005 rather than 0X80000003)

    These 3 separate crashes may in fact be connected, possibly by a common error or instability within PET that increases with severity after consecutive CTDs.
    Edit: These crashes do indeed seem to be all related to browsing and selecting a particle to be assigned as a child particle of itself.

    Status
    : 1 is reproducible; 2 happens once, leads to 2b with consecutive CTDs (within same system boot session).

    Crash reports
    :
    1 crash_dota2_20150114163619_1.zip
    2
    crash_dota2_20150114164619_1.zip
    2b crash_dota2_20150114165815_1.zip
    Last edited by Toyoka; 01-14-2015 at 04:02 PM. Reason: Further clarified cause of crashes; greyed text refers to previous assumptions

  2. #2
    Basic Member Noya's Avatar
    Join Date
    Dec 2011
    Location
    Uruguay
    Posts
    12,904
    Hmm I should probably make a compilation of all my particle editor crashes, but I don't know what causes most or how to reproduce them.

  3. #3
    Basic Member Toyoka's Avatar
    Join Date
    Oct 2013
    Location
    Current position co-ordinates non-tracable. Unknown space.
    Posts
    833
    Please do. Also, keep vconsole open to more accurately debug and/or reproduce the steps in case you run into a crash

  4. #4
    Basic Member Toyoka's Avatar
    Join Date
    Oct 2013
    Location
    Current position co-ordinates non-tracable. Unknown space.
    Posts
    833
    For convenience (and because PET seems to be prone to crashes quite frequently), I'll try and post all my particle editor tool-related crashes in one place; in this thread. If for only my own reference, should these reports go unnoticed. Will bump the thread with a CTD report when I encounter additional CTDs.

    Encountered one just recently, logged below.

    Details: Remapping control point to scalar (or vector) with an output field of Scratch Float 1 or Scratch Float 2 causes an immediate CTD. Scratch Float 1 CTD only seems to happen on a particle system with an assortment of functions and may be affected directly or indirectly by these functions (it does not cause a crash on a simple particle system as far as I have tested). While Scratch Float 2 on the other hand will cause an immediate CTD even on the simplest particle system.

    Status: Scratch Float 1 CTD reproducible. Take any system that utilizes remapping of a control point to scalar or vector and change the output field to Scratch Float 1, it will likely crash if it is functionally complex enough (more likely a limitation than a bug, in this case). The Scratch Float 2 crash is immediately reproducible by creating a simple particle system with a Render Sprites function, a continuous emitter and the remap control point to scalar/vector operator function. Not really sure what the Scratch Float output fields actually do (if anything), but I thought I'd report it just in case

    Crash report: crash_dota2_2015-01-21__01-00-19am_1.zip

  5. #5
    Basic Member Toyoka's Avatar
    Join Date
    Oct 2013
    Location
    Current position co-ordinates non-tracable. Unknown space.
    Posts
    833
    Just encountered another PET CTD. This one seems to be related to instability.

    Details: Was working on a particle system for Io, changing some values while observing the results in-game. Was doing this for about 2 hours, when suddenly after changing a function of a particle (radius scale, not that it matters.. probably), the tool hung up for about half a minute and then a CTD occurred.

    Status: Not sure if it is exactly an instability crash. Not sure if reproducible either, since it happened seemingly at random, although there might be a hard-limit to particle rendering in-game that I am probably not aware of. Also likely the result of prolonged editing of functions, but perhaps this is not the case. No memory leaks or post-crash system lag is present, so probably tools-related and not memory-related.

    Crash report: crash_dota2_2015-02-09-211312_1.zip



    I must also mention that I have encountered a minor visual bug concerning particles being worked on in real-time (observed in-game while simultaneously being edited in the PET); the particle being worked on can become glitched if you alt-tab out of the game window while the particle is being fired off, and subsequently edited within the PET. It does not properly fade/decay as it should, and stays rendered within the map beyond its set lifetime. It is also not seemingly affected by certain parameters that would otherwise kill the particle (ie. Io's tether can be glitched to stay active indefinitely, even beyond its range limit at which it would otherwise snap). Any changes made to the particle will be reflected within the map.

  6. #6
    Basic Member Toyoka's Avatar
    Join Date
    Oct 2013
    Location
    Current position co-ordinates non-tracable. Unknown space.
    Posts
    833
    The reports in this thread are confirmed to be fixed in the beta tools update.

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •