[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Bug#60399: crashes on installation



[can anybody with knowledgs of the internals of dpkg/apt take a look at
this?]

On Thu, Mar 16, 2000 at 10:13:52AM +1100, Brian May wrote:
> 
>     Fabrizio> If you downgrade man-db (I'm reuploading 2.3.13 so
>     Fabrizio> you'll find it in Incoming or Incoming/REJECT) and then
>     Fabrizio> re-apt it, will it fail?
> 
> Anyway, have a look at the following:
> 
> lyell:~# apt-get upgrade
> ..............
> Unpacking replacement man-db ...
> dpkg-deb: subprocess paste killed by signal (Broken pipe)
> dpkg: error processing /var/cache/apt/archives/man-db_2.3.14_i386.deb (--unpack):
>  subprocess dpkg-deb --fsys-tarfile returned error exit status 2
>   Building manual page index in background.
> Errors were encountered while processing:
>  /var/cache/apt/archives/man-db_2.3.14_i386.deb
> E: Sub-process /usr/bin/dpkg returned an error code (1)
> 
> 
> We have success! Problem duplicated!

Great.
As you already have the old deb, can you please repete the test, but
using dpkg -i instead of apt-get (to re-upgrade I mean ... I don't think
apt-get can downgrade).

That is just to narrow the search of the package to reassign the 4 bugs
:-)

But seriously, it's quite clear that something in this man-db package
it's triggering this ugly behaviour.
So, what's changed between 2.3.13 and 2.3.14 ?

< -rwsr-xr-x root/root 82752 ./usr/lib/man-db/man
> -rwxr-xr-x root/root 82752 ./usr/lib/man-db/man
---
< -rwsr-xr-x root/root 66012 ./usr/lib/man-db/mandb
> -rwxr-xr-x root/root 66012 ./usr/lib/man-db/mandb
---
< -rw-r--r-- root/root 9008 ./usr/share/man/man1/man.1.gz
> -rw-r--r-- root/root 9006 ./usr/share/man/man1/man.1.gz
---
< -rw-r--r-- root/root 9986 ./usr/share/man/it/man1/man.1.gz
> -rw-r--r-- root/root 9984 ./usr/share/man/it/man1/man.1.gz
---
< -rw-r--r-- root/root 14485 ./usr/share/doc/man-db/changelog.Debian.gz
> -rw-r--r-- root/root 14617 ./usr/share/doc/man-db/changelog.Debian.gz


<  size 332058 bytes: control archive= 4325 bytes.
>  size 332200 bytes: control archive= 4320 bytes.
---
<      792 bytes,    20 lines      control              
>      816 bytes,    20 lines      control              
---
<  Version: 2.3.13
>  Version: 2.3.14
---
<  Depends: groff (>> 1.15-2) | jgroff (>> 1.15), libc6 (>= 2.1.2)
>  Depends: groff (>> 1.15-2) | jgroff (>> 1.15), libc6 (>= 2.1.2), libdb2 (>= 1:2.4.14-7)
---

Another thing that is changed is inside the md5sum file:
< a79d0fb7542beef4281c7d754707bac0  usr/bin/wrapper
< bc5ae3db0572b49aa938538f37c4147b  usr/bin/man
< bc5ae3db0572b49aa938538f37c4147b  usr/bin/mandb
> f9b7b6014efe84c9cea791c6e2b2f568  usr/bin/wrapper
> f9b7b6014efe84c9cea791c6e2b2f568  usr/bin/man
> f9b7b6014efe84c9cea791c6e2b2f568  usr/bin/mandb

In version 2.3.13 the md5sum of /usr/bin/man and /usr/bin/mandb is the
md5sum of the shell scripts files that are in the .deb and are removed
inside postinst, while in the 2.3.14 md5sum its listed the digest of the
files that are installed after the postinst (hardlinks to
/usr/bin/wrapper), and not the digests of the files shipped inside the
.deb.

Does anybody know if this could be the trigger for this ugly stuff?


fab
-- 
| fab@pukki.ntc.nokia.com                     fpolacco@debian.org
| pgp: 6F7267F5   57 16 C4 ED C9 86 40 7B 1A 69 A1 66 EC FB D2 5E
| fabrizio.polacco@nokia.com             gsm: +358 (0)40 707 2468


Reply to: