Subject: debdelta: Improve performance info
Package: debdelta
Version: 0.67
Severity: minor
Dear Maintainer,
I'm very happily using `debdelta-upgrade` now that I finally heard about
it (I've been using Debian for 20 years and wishing for something like
DebDelta for a big part of it), so first: thanks!
One think that annoys me is that I find the output unclear. E.g. I see
[note the output was actually in French, so I translated it by hand, it
probably doesn't match 100% what debdelta would say]:
Created, time 0.42sec, speed 202kB/sec, avahi-daemon_0.8-12_i386.deb
[...]
The delta est too large: libmbim-glib4_1.28.4-2_1.30.0-1_i386.debdelta
Delta was not created since new package is too small:
libmbim-proxy_1.28.4-2_1.30.0-1_i386.debdelta
[...]
Created, time 1.07sec, speed 239kB/sec,
libedata-book-1.2-27_3.50.1-1_i386.deb
[...]
Downloaded, time 0.11sec, speed 8kB/sec,
libudisks2-0_2.10.1-1_2.10.1-2_i386.debdelta
Downloaded, time 0.24sec, speed 13MB/sec,
libreoffice-calc_4%3a7.5.6-1_4%3a7.5.8~rc1-2_i386.debdelta
[...]
Downloaded, time 0.02sec, speed 488kB/sec, libmbim-proxy_1.30.0-1_i386.deb
Downloaded, time 0.03sec, speed 7658kB/sec, libmbim-glib4_1.30.0-1_i386.deb
[...]
Created, time 27.30sec, speed 296kB/sec,
libreoffice-calc_4%3a7.5.8~rc1-2_i386.deb
[...]
Statistics of delta-upgrade:
debs result total, size 115MB time 257sec virtual speed 458kB/sec
Questions that I can't answer based on the above output:
- I don't see any other mention of "avahi-daemon" than the "created" line,
so how was it created?
[ Could it be due to some previous `debdelta-upgrade` run which
was maybe interrupted? I can't remember causing that recently
enough, tho. ]
- Why/how were `libmbim-proxy` and `libmbim-glib4` downloaded (since
there's a delta missing for them) and from where?
There's no subsequent matching "created" line, so IIUC it was
downloaded in non-delta form, which I'd expect `debdelta-upgrade`
never does (leaving it to `apt upgrade` instead).
- The "Downloaded" speed varies very widely (which is admittedly fair
game, it might just be the result of the state of the network and
server), and oddly enough I've seen it often higher than my DSL
connection speed (even for large packages, so it doesn't seem to be
some kind of rounding error), so it makes me wonder what is it
exactly measuring.
Could it be that it's reporting the speed of "virtual bytes"
(i.e. the number of bytes of the resulting `.deb` after patching,
rather than the number of bytes of the actual xdelta file)?
- I'd appreciate seeing the actual size of the downloaded data on each
line (maybe instead of the time?).
- In the final statistics, I'd be interested to see a report comparing
the amount of bytes that passed over the network compared to the
number of bytes in the resulting `.deb` files (so as to see how much
we gained in this). I'd also be interested to see the average network
download speed (not virtual) and the average "creation speed" (number
of `.deb` bytes generated via patching divided by the time it took to
do it).
Stefan
-- System Information:
Debian Release: trixie/sid
APT prefers testing
APT policy: (990, 'testing'), (500, 'stable-security'), (100, 'stable')
Architecture: i386 (x86_64)
Foreign Architectures: amd64
Kernel: Linux 6.1.0-7-amd64 (SMP w/2 CPU threads; PREEMPT)
Kernel taint flags: TAINT_WARN, TAINT_OOT_MODULE, TAINT_UNSIGNED_MODULE
Locale: LANG=fr_CH.UTF-8, LC_CTYPE=fr_CH.UTF-8 (charmap=UTF-8), LANGUAGE not set
Shell: /bin/sh linked to /usr/bin/dash
Init: systemd (via /run/systemd/system)
LSM: AppArmor: enabled
Versions of packages debdelta depends on:
ii binutils 2.41-6
ii bzip2 1.0.8-5+b1
ii libbz2-1.0 1.0.8-5+b1
ii libc6 2.37-12
ii python3 3.11.4-5+b1
ii python3-requests 2.31.0+dfsg-1
ii zlib1g 1:1.2.13.dfsg-3
Versions of packages debdelta recommends:
ii bsdiff 4.3-23
ii gnupg2 2.2.40-1.1
ii gpg-agent [gnupg-agent] 2.2.40-1.1
ii lzma 9.22-2.2
ii python3-apt 2.6.0
ii python3-debian 0.1.49
ii xdelta 1.1.3-10.4
ii xdelta3 3.0.11-dfsg-1.2
ii xz-utils [lzma] 5.4.4-0.1
Versions of packages debdelta suggests:
pn debdelta-doc <none>
-- no debconf information
