The debian-private mailing list leak, part 1. Volunteers have complained about Blackmail. Lynchings. Character assassination. Defamation. Cyberbullying. Volunteers who gave many years of their lives are picked out at random for cruel social experiments. The former DPL's girlfriend Molly de Blanc is given volunteers to experiment on for her crazy talks. These volunteers never consented to be used like lab rats. We don't either. debian-private can no longer be a safe space for the cabal. Let these monsters have nowhere to hide. Volunteers are not disposable. We stand with the victims.

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

Re: Unstripped executables



Ian Jackson <ian@chiark.chu.cam.ac.uk> said:

> As you can see below this turns dpkg from a 118K executable into a
> file which is well over *half a megabyte* !  The debugging symbols do
> clearly compress well, but in the simple test below the stripped file
> still compresses to well under half the size of the unstripped one.
(and, later)
> Quite why anyone would want to build a whole distribution unstripped
> is beyond me; a large proportion of the installed files are
> executables, and doing this will easily double the disk space
> requirements.

I'm a bit uncomfortable that I've cast myself as the proponent of
distributing unstripped executables, as I'm personally not convinced
that distributing unstripped executables would be all that useful.
Still, I think it ought to be argued fairly on its merits, or lack
thereof, and I think the above might present a somewhat distorted
picture by focusing on executable size differentials instead of
package size differentials.

Taking Ian's second point first, the suggestion was to distribute
unstripped executables in the binary package, but to strip them on
installation.  Therefore, in the normal case, the disk space requirements
on the user's system are not doubled.  In fact the disk space requirements
on the user's system are not increased at all.  (The suggestion was
also that installing stripped vs. unstripped executables could be a dpkg
config option and invocation option, with the idea that individual
executables could be installed unstripped for testing, and then
reinstalled stripped for space saving.)

Taking Ian's first point second, the effect of distributing unstripped
executables would be seen as an increase in the size of binary packages.
Speaking of it in terms of percentage increase in individual
executable size makes for more dramatic numbers, however the overall
differential is seen in package size, not in the size of individual
executables inside packages.  Someone recently posted comparative
results from a test package, giving a size penalty of about 30%
increase in package size, vs. a 300% increase in executable size.
I haven't done any systematic testing myself, and don't have any
good idea how representative that 30% figure might be as an average
over all the packages in the entire distribution.

Since this seems to be a prety emotionally charged issue, I think I'll
get off it at this point without exploring it further.

mitchell@mdd.comm.mot.com (Bill Mitchell)