Can you try "beagle-extract-content --mimetype=message/rfc822 /path/to/file"? It worked for me here.

I tried the following:

tamas@milleniumfalcon:~/beagle> beagle-extract-content --mimetype=message/rfc822 1191841883.13274.RPjw4

The output was:

Filename: file:///home/tamas/beagle/1191841883.13274.RPjw4
Debug: Loaded 53 filters from /usr/lib/beagle/Filters/Filters.dll

And it hang forever with 100% cpu utilization. By the way I use openSUSE 10.3.

Ok, I tried using the branch for 0.2.18 and still it runs fine. This could be some problem with the OpenSUSE gmime-sharp or some other local patch in the OpenSUSE 10.3.

Joe, do you know anything offhand ?


----output of extract-content (0.2.18 is made out of 0.2.16 svn branch)-----

Filename: file:///home/debajyoti/1191841883.13274.RPjw4
Debug: Loaded 53 filters from /usr/share/devel/beagle-gnome/branches/beagle-0.2.16/Filters/Filters.dll
Filter: Beagle.Filters.FilterMail (determined in .60s)
MimeType: message/rfc822

Properties:
  Timestamp = 2007-10-09 11:26:07 (Utc)
  dc:title = Job No: 1 Job begin
  fixme:date = 20060704132711
  fixme:from = Webmaster hrlevl teszt <webmaster@nyme.hu>
  fixme:from_address = webmaster@nyme.hu
  fixme:from_name = Webmaster hrlevl teszt
  fixme:msgid = 20060704132711.084A0CB3@bismarck.nyme.hu
  fixme:to = webmaster@nyme.hu
  fixme:to_address = webmaster@nyme.hu

Content:
Job begin: 04-07-06 03:27:11

HotContent:
(no hot content)

Text extracted in .05s

*** Bug 490074 has been marked as a duplicate of this bug. ***

I figured out that if I change the subject line from:
Subject: =??Q?Job_No:_1_Job_begin?=
to something like:
Subject: Job_No:_1_Job_begin
then beagle is able to index the file!

I forgot to mention that i run openSUSE 10.3 on i386 (32 bit)

Markus, Nemeth,
Do you use non-english language or encoding ?
Nemeth, the "?" in the email is really english "?" or the base64 changed it to english "?". If there is no funky encoding involved, then I dont see any way beagle could be using gmime incorrectly. Then this bug should really be investigated by the gmime people.

gmime contains several test programs, e.g. there is gmime/tests/test-parser.c. If you want, you can download that test program and run it with the path to the bad file as argument. Maybe that will help in figuring out why gmime is taking forever on these emails.

Created attachment 98660 [details]
problematic file (mail)

I attached the file tu this bug report, but you also can extract it from byte to byte from the base64 encoded version above. As you can see this way, the question marks in the Subject line are really question marks (0x3f). The From line itself is actually coded 8 bit iso8859-2, because it is in hungarian. It does not show the encoding however, and I guess it may be incorrect this way, isn't it? I'm not an RFC822 expert at all, but I feel that this mail is malformed at many points. Maybe the program, which generated the mail was not capable of handling the iso8859-2 encoding.

Created attachment 98675 [details] [review]
possible fix

I wasn't able to reproduce the problem on my OpenSuSE 1.3 machine w/ gmime-2.2.10, so I'm not sure that this patch actually fixes the problem... but it's the only thing I can see in the code which /may/ be related.

that hould read OpenSuSE 10.3

Not sure but that could be the problem. If any of the reporters could try beagle with gmime+the above patch (gmime is extremely easy to build), we can try to fix this forever.

Dear Jeff!

 I did the following.

I downloaded and installed http://download.opensuse.org/distribution/10.3/repo/src-oss/suse/src/gmime-2.2.10-35.src.rpm in order to be able to reproduce the openSUSE gmime rpms. I did so, but I did't use the build script by SuSE, just the raw rpmbuild program like this.
cd /usr/src/packages/SPECS
rpmbuild -ba gmime.spec

After building the rpms I replaced the already installed original gmime-sharp, libgmime-2_0-2 and gmime packages with the respective (and hopefully totally identical) ones generated by me just now.

I tested beagle by copying my mail file and Markus Ehrnsperger's file from Bug 490074 to my home directory and the symptoms were the same as with the original openSUSE 10.3 version of gmime.

After doing so, I modified /usr/src/packages/SPECS/gmime.spec the following way:

-I increased the release number to 36.
-I downloaded your patch as
   /usr/src/packages/SOURCES/gnome-bugzilla-id-485005.patch
   and added the following line ti gmime.spec.
   Patch2:         gnome-bugzilla-id-485005.patch
-At the %prep section I added: %patch2 -p0

After doing so I build the version 36 counterparts of my gmime rpm packages and replaced the old ones with the brand new ones (hopefully) including your patch.

And now the symptoms are the same for both problematic file:((((

PS: To tell the truth, I haven't even heard about rpmbuild so far, so it's possible that I made a mistake. Whether or not is this the case, the building of the version 36 rpms went without problem and I attach the src rpm to this bug report for anyone to be able to reproduce the rpms with your patch.

Created attachment 98731 [details] [review]
Patched rpm, but no positive results for me

Created attachment 98732 [details]
strace beagle-extract-content /beagle/1191841883.13274.RPjw4

Created attachment 98733 [details]
strace beagle-extract-content /beagle/1108158121.18206.VP5fq\:2\,S

Dear Jeff!

 I attached the outputs of stracing beagle-extract-content on both problematic files. By the way, how can it be possible that you can not reproduce the symptom on your openSUSE 10.3 machine??? Don't you run the 32 bit intel version? Or you have some piece of code compiled by hand, replacing some binaries of the original rpms? Or you have installed one of the RC versions instead of the final openSUSE 10.3?

aha, found the problem... it was a bug with the --enable-rfc2047-workarounds code path.

Created attachment 98977 [details] [review]
485005.2.patch

this really fixes it

OK, I added both patches to the openSUSE src.rpm, built the actual rpms and installed them. Now beagled is able to index both problematic files, I can find them via the SuSE style KDE menu and via kerry beagle too. beagle-extract-content also easily parses the files.

Great job! I say thank you in the name of the community.

You may possibly change this bug's status to fixed.

When will these patches find their way to openSUSE (and official gmime) packages?

Changing product and summary. Jeffrey, mark it as fixed if the patch has been checked in.

the fix is in the 2.2.11 release of GMime