New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Change $RM in sage-env #3537
Comments
Attachment: trac_3537_scripts.patch.gz |
comment:2
The intent of -f is to avoid errors for the conditions cited as a problem. |
comment:3
I am not convinced yet that this is the right thing to do and it is rather likely that this will break something else in the tree. The solution to M2's problem is to unset RM in spkg-install and then to set RM to some value you desire. And this is definitely not a blocker for 3.0.4. Cheers, Michael |
comment:4
If M2 depends on the default behavior of RM in its Makefiles it should set RM in its top level makefile. I see no reason to change Sage's behavior and as is if someone sets RM outside of sage-env it is propagated anyway. So this is wontfix. Cheers, Michael |
comment:5
I'm reopening this bug since people keep tripping over this issue. We need to fix this or we'll end up with every spkg working around the For the record, gfurnish's patch applies cleanly on Sage-4.6.1.alpha3 (apply to the $SAGE_LOCAL/bin repo) I'd be happy to give this a positive review. Maybe mabshoff can reconsider his objections? |
Author: gfurnish |
This comment has been minimized.
This comment has been minimized.
comment:7
Replying to @vbraun:
Well, is this a bug?
That's IMHO a problem of autotools.
Michael has quit a while ago, though he perhaps still reads trac notifications... ;-) Note that redefining Also, as Michael said, changing the default value in I think we should prominently document this issue, and close this ticket again. |
comment:18
My opinion still is that |
comment:19
Replying to @jdemeyer:
I agree 100%. I can understand the logic of variables like $CC, $CXX, $MAKE, but not $RM. I don't know of any system where one might want to specify what version of BTW, take a look at http://boxen.math.washington.edu/home/kirkby/bad-code/sympow-1.018.1.p7/src/Configure where you will find some interesting use of variable names. A script which starts
|
comment:20
See #9497 for a patch to change "$RM" to "rm" in Singular's spkg-install script. |
This comment has been minimized.
This comment has been minimized.
Reviewer: Mariah Lenox |
comment:21
When 50 percent or more of spkgs require $RM to be "rm -f" then Positive review. |
This comment has been minimized.
This comment has been minimized.
comment:22
Singular's spkg-install no longer uses $RM, so the documentation is not up to date. |
Do not set $RM in sage-env |
comment:23
Attachment: trac_3537-unset-RM.patch.gz attachment: trac_3537-unset-RM.patch applied to sage-4.7.rc2 tested on skynet/taurus with make testlong. All tests passed. Positive review. |
comment:24
Replying to @sagetrac-mariah:
Did you try building Sage from source with this patch applied? Because this patch affects the Sage build, not the Sage library. |
comment:25
Apologies for not be clear. I first applied the patch to the sage-4.7.rc2 source, then built the source, then did make testlong. |
Changed author from gfurnish to Jeroen Demeyer |
comment:27
I just checked all spkg-install scripts and |
Merged: sage-4.7.1.alpha1 |
The env variable RM is set to
rm
instead ofrm -f
. This breaks newer libtools, for example anything in Fedora 12 or later (libtool 2.2.6, autoconf 2.63, automake 1.11.1). They assume that$RM
is either unset orRM="rm -f"
, that is, deleting non-existing files must not cause an error.One of the symptoms of this breakage is that configure ends with
Compile will break later on...
There are three approaches presented here:
Apply attachment: trac_3537-unset-RM.patch
CC: @sagetrac-mabshoff @SnarkBoojum @sagetrac-drkirkby @jdemeyer
Component: scripts
Author: Jeroen Demeyer
Reviewer: Mariah Lenox
Merged: sage-4.7.1.alpha1
Issue created by migration from https://trac.sagemath.org/ticket/3537
The text was updated successfully, but these errors were encountered: