Mercurial slow, impoosible to clone/pull/push

[junk2 at skjaerris.com]

The xml files contains a huge number of duplicate lines, E.g. The xml 
file consists of info about 5000+ parameters in the system.
E.g one entry is:
   <parameter>
     <name>Par100001</name>
     <num>100001</num>
     <type>int</type>
     <text>This is parameter xxx</text>
     <reset>yes</reset>
     <readlevel>2</readlevel>
     <writelevel>2</writelevel>
     <clearanceLevel>1</clearanceLevel>
     <owner>Brian</owner>
     <wtc3num>99999</wtc3num>
     <wtc3comment>STC-1_specific</wtc3comment>
     <value>
       <min>0</min>
       <max>0</max>
       <def>0</def>
     </value>
   </parameter>

And e.g. the tags readlevel, writelevel and clearancelevel, can only be 
to different values, so I guess for 90% of the 5000+ parameters these 
lines are identical.

And we have such an entry for each parameter.

Matt Mackall skrev den 20.01.2015 22:19:
> On Tue, 2015-01-20 at 15:36 -0500, Augie Fackler wrote:
>> On Mon, Jan 19, 2015 at 5:20 AM,  <junk2 at skjaerris.com> wrote:
>> > Hi All
>> >
>> > I have a mercurial repository, which includes 2 xml files..
>> > Par1.xml (size approx 1.5 MB) (600 revisions)
>> > Par2.xml (size approx 5 MB) (1200 revisions)
>> >
>> > It has been like that for some time, but lately I am struggling by making
>> > clone/pull/push from the mercurial repository. When I do a clone it's
>> > hanging when it reaches these 2 xml files, sometimes it goes though after
>> > 3-5 minutes other times it fails.
>> 
>> 
>> Are these xml files remarkable? Do they contain any newlines at all?
> 
> Alternately, do they have a huge number of duplicate lines? This can
> trigger some worst-case behavior in our delta generation algorithm.