Comment # 5
on bug 12228
from Guy Harris
Why on *earth* are we creating all those data sources? I *very sincerely* hope
that we're not creating registered data sources for *subset* tvbuffs, as that's
completely the wrong thing to do - create new data sources for data that
*isn't* just a subset of data in the parent tvbuff, such as reassembled data
and decrypted data and decompressed data, but if the data is in the parent
tvbuff, let the parent tvbuff handle it. That way, you don't waste CPU time,
memory, and screen real estate in the tab bar at the bottom with a ton of data
sources.
It might expand faster with fewer data sources. I doubt it's running out of
virtual memory, though - even though the version of Wireshark being run here is
32-bit, even though it's running on a 64-bit machine with a 64-bit OS, running
it (64-bit) on (64-bit) OS X and running vmmap on it shows
ReadOnly portion of Libraries: Total=319.1M resident=0K(0%)
swapped_out_or_unallocated=319.1M(100%)
Writable regions: Total=341.3M written=0K(0%) resident=0K(0%)
swapped_out=0K(0%) unallocated=341.3M(100%)
...
VIRTUAL REGION
REGION TYPE SIZE COUNT (non-coalesced)
=========== ======= =======
...
=========== ======= =======
TOTAL 818.8M 819
so, unless I'm misreading the output or vmmap is miscalculating the VM size,
it's less than 1GB of virtual address space total.
You are receiving this mail because:
- You are watching all bug changes.