<?xml version="1.0" encoding="UTF-8"?>
<issue>
  <id>58</id>
  <project name="Gummi" id="1"/>
  <tracker name="Bug" id="1"/>
  <status name="New" id="1"/>
  <priority name="High" id="5"/>
  <author name="Alexander van der Mey" id="3"/>
  <subject>Memory leak</subject>
  <description>&gt;&gt;copy paste:

Hi,

I believe that there is a memory leak in Gummi, since I've been editing only a single few kB file for a few hours now, and its memory usage has reached 500 MB.
That's way more than FireFox!

It also seems that it consumes about 0.4 MB every time the Preview Pane is refreshed - I assume that Gummi keeps all the previews in memory, in case undo is used so that they don't have to be redrawn. The number of undo's is limited, however, the number of previews that are stored in the memory is either not limited, or the memory is just not freed. Either way, after some time, Gummi will eat up lots more memory than it should ever need.

I'm using Gummi 0.4.2 on Ubuntu 9.10 with 2.6.31-14-generic x86_64 kernel.
</description>
  <start_date>2010-01-23</start_date>
  <due_date></due_date>
  <done_ratio>0</done_ratio>
  <estimated_hours></estimated_hours>
  <spent_hours>0.0</spent_hours>
  <created_on>Sat Jan 23 18:06:58 +0000 2010</created_on>
  <updated_on>Tue Jan 26 18:17:07 +0000 2010</updated_on>
  <journals>
    <journal id="109">
      <user name="Alexander van der Mey" id="3"/>
      <notes>I investigated and can confirm the issue. "This page":https://bugs.launchpad.net/poppler-python/+bug/509408 seems to describe the problem we are experiencing, and includes a fix. I'm somewhat hesistant to make calls to a c-library function, but will include this fix is no better solution presents itself. </notes>
      <details>
      </details>
    </journal>
    <journal id="110">
      <user name="Alexander van der Mey" id="3"/>
      <notes>Implemented the beforementioned patch.. let's see how it holds up. Feel free to reopen/comment if you are experiencing problems. </notes>
      <details>
        <detail old="1" property="attr" name="status_id" new="3"/>
      </details>
    </journal>
    <journal id="111">
      <user name="Alexander van der Mey" id="3"/>
      <notes></notes>
      <details>
        <detail old="3" property="attr" name="status_id" new="1"/>
      </details>
    </journal>
    <journal id="113">
      <user name="Alexander van der Mey" id="3"/>
      <notes>Implemented the before mentioned c-library trick, will stay in Gummi until a better solution presents itself. Bug remains open until then. </notes>
      <details>
      </details>
    </journal>
  </journals>
</issue>
