[tor-bugs] #32230 [Core Tor/Tor]: configure summary is confusing or incorrect

[Tor Bug Tracker & Wiki]

#32230: configure summary is confusing or incorrect
--------------------------+------------------------------------
 Reporter:  teor          |          Owner:  dgoulet
     Type:  defect        |         Status:  needs_review
 Priority:  Medium        |      Milestone:  Tor: 0.4.3.x-final
Component:  Core Tor/Tor  |        Version:  Tor: unspecified
 Severity:  Normal        |     Resolution:
 Keywords:  043-must      |  Actual Points:
Parent ID:                |         Points:  0.5
 Reviewer:  teor          |        Sponsor:
--------------------------+------------------------------------
Changes (by dgoulet):

 * status:  needs_revision => needs_review


Comment:

 Replying to [comment:2 teor]:
 > Replying to [comment:1 dgoulet]:
 > > Replying to [ticket:32230 teor]:
 > > > Inverted, should be --enable-* :
 > > > * --disable-seccomp
 > > > * --disable-libscrypt
 > > > * --disable-gcc-hardening ?
 > > > * --disable-linker-hardening ?
 > > > * --disable-module-dirauth
 > > > * --disable-unittests
 > >
 > > All these are enabled by default and thus why we put in the
 `--disable`. If one does autocompletion on `./configure --enable-`, none
 of the above shows up.
 >
 > Ok, but the output is still wrong:
 >
 > {{{
 > test "x$enable_seccomp" != "xno" && value=1 || value=0
 > PPRINT_PROP_BOOL([libseccomp (--disable-seccomp)], $value)
 > }}}
 >
 > If enable_seccomp is yes, then value is 1, and we print --disable-
 seccomp yes.
 > But we should actually print one of these:
 > * --enable-seccomp yes
 > * --disable-seccomp no
 >
 > I prefer the output without the double-negative.
 > But I can understand if you want to print the non-default option as
 well.

 Right, the string is hardcoded regardless of the value of the option.
 Basically, the `--disable-seccomp` is the configure option that is
 available, regardless if you enabled it or not.

 The way to look at it is `libseccomp = yes/no`. The configure options
 there were added so the user knows how to switch the option on or off.

 I agree it is confusing. We would need to handle different text output for
 _each_ option depending on their values. Maybe another ticket?

 Another way to look at it is that `--enable-seccomp` virtually doesn't
 exists, you either have no option which enables seccomp by default or you
 disable it with `--disable`.

--
Ticket URL: <https://trac.torproject.org/projects/tor/ticket/32230#comment:4>
Tor Bug Tracker & Wiki <https://trac.torproject.org/>
The Tor Project: anonymity online