Memory leak in nemo 1.8.3

[laecdtu]

I like Mint 15 but wonder why nemo is using more than 3GB of memory and %CPU at 30+! Which causes my i7 to appear slow in some programs, e.g. 5 seconds delays multiple times while writing this email as I have changed windows several times.

I tried to file a bug using:
ubuntu-bug nemo

but got:

ERROR: hook /usr/share/apport/package-hooks/source_nemo.py crashed:
Traceback (most recent call last):
File "/usr/lib/python3/dist-packages/apport/report.py", line 197, in _run_hook
symb['add_info'](report, ui)
TypeError: add_info() takes 1 positional argument but 2 were given

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
File "/usr/lib/python3/dist-packages/apport/report.py", line 202, in _run_hook
symb'add_info'
File "/usr/share/apport/package-hooks/source_nemo.py", line 6, in add_info
if report.has_key("Stacktrace") and "/usr/lib/nemo" in report["Stacktrace"]:

AttributeError: 'Report' object has no attribute 'has_key'

Then I ran valgrind:
valgrind --leak-check=full --show-reachable=yes nemo
I dumped what was sent to stderr - a total of 4.7Mb.

The last part is:
==18381== LEAK SUMMARY:
==18381== definitely lost: 0 bytes in 0 blocks
==18381== indirectly lost: 0 bytes in 0 blocks
==18381== possibly lost: 393,364 bytes in 2,939 blocks
==18381== still reachable: 495,048 bytes in 4,598 blocks
==18381== suppressed: 0 bytes in 0 blocks
==18381==
==18381== For counts of detected and suppressed errors, rerun with: -v
==18381== ERROR SUMMARY: 839 errors from 839 contexts (suppressed: 2 from 2)

After a reboot it doesn't jump directly to this buggy state and I don't know what causes it. I've experienced it multiple times.

I hope that someone will look into this issue and hope for an update ;-)

Best regards
Lasse

PS: Don't hesitate to contact me if you need more info.
$ uname -a
Linux comp-nblaec 3.8.0-19-generic #29-Ubuntu SMP Wed Apr 17 18:16:28 UTC 2013 x86_64 x86_64 x86_64 GNU/Linux

[mtwebster]

See #346 and #349

There's a patched version 1.8.4 in Romeo repos now, that will be rolled out shortly to Main that should address this. It's possible this is unrelated, however I've not seen any issues apart from the thumbnailing one.

[laecdtu]

I don't know if it is related to thumbnails. But I hope so. I'll wait for 1.8.4 and get back back if it doesn't solve my issue.

Thanks for a quick comment!

[eljaydub]

I see this is closed but I am having the same issue on Mint 15 and just upgraded from Nemo 1.8.3 to 1.8.4 through the Romeo repos. The issue is apparently still there for me with 1.8.4: I killed and restarted nemo after the upgrade and later performed a full system reboot but the problem persists. Nemo 1.8.4 is currently using 21% of CPU on a dual core pentium 2.3Ghz (ie 42% of one CPU) and eating memory at a rate of about 0.5Mb/s.

In 1.8.3 I saw it get to about 2Gb before my system slowed down to unusable.

Nemo was stable at 36Mb of ram until I navigated to a folder with several .svg files I am currently working on. The SVG's are generated with java code and every time the code is run the thumbnail preview in the window is refreshed - this is when Nemo starts eating memory. The thumbnails always show the clock icon (ie refreshing) but never finish. Hope that helps identify the issue!

In the mean time to avoid the issue I just turned off thumbnail previews in Nemo's Edit->Prefences->Preview menu.

[cipri-tom]

Will 1.8.4 be available in Linux Min 15? I am also facing this issue