Join GitHub today
GitHub is home to over 36 million developers working together to host and review code, manage projects, and build software together.Sign up
Configure script reject legitimate arguments #6628
Original bug ID: 6628
The configuration script says
This is broken because it matches legitimate arguments, like:
Fix: Do not try to pseudo-validate input in a broken way. We could report suspicious arguments in the
Steps to reproduce
./configure -cc "gcc48 -O2 -pipe -Wl,-rpath=/usr/local/lib/gcc48 -fno-strict-aliasing"
Comment author: michipili
Here it is!
I also propose another error message, as the previous is not very clear. Until I read the source, I understood it as if it referred to the actual arguments of the script. A more correct statement could have been “This script expects arguments looking like '-prefix /foo/bar', not '-prefix=/foo/bar' (note the '=').”.
Comment author: michi
Can we reopen this issue? The argument validation stroke two new FreeBSD users, and I asked them
Argument validation is expected to make people life's easier. Is there some of these people?
Comment author: @gasche
The reason I closed this issue is because of this commit by Damien:
This change is not part of any released OCaml version so far (it should be part of the next release), so when you say "stroke two new users" I'm not sure whether you mean that even this refined test is not correct, or just that they were hit with the same old bug with the released version.
I tested that the configure command you provide works fine under not-yet-released versions (for example the stable 4.02 branch). If you are not sure whether the above patch fixes the issues that those two new users have, yet don't want to setup an experimental version just for this, I would be glad to test proposed inputs and report whether they are still problematic.
Comment author: michi
Ooooh, there is a bit of a misunderstanding here! Thank you for pointing out that!
I erroneously assumed the change would be in 4.02.1 – maybe I should have read better the “fixed in” field. So we can keep this issue closed, add the patch in FreeBSD and remove it as 4.02.2 arrives.
Thank you for the clarification!