I had the exactly same issue a while ago and finally got rid of it after many tries. It even appeared after a fresh git clone. I had to unset all (not just core.autocrlf) global Git configurations except username and password, even push.default (which I do on all other Git repositories locally, now).
BTW: I suspect that the issue with the broken GitLab pipelines for Trac tickets <https://gitlab.com/sagemath/dev/trac/-/pipelines> that I mentioned in this sage-devel thread <https://groups.google.com/g/sage-devel/c/Dbd4nFi8R7I/m/SEI0xx1vBwAJ> is related. dim...@gmail.com schrieb am Donnerstag, 11. Februar 2021 um 12:24:27 UTC+1: > It could also be that core.autocrlf is true by default for you. > (this is the option that does automatic convesion to LF from CRLF) > Try > > git config core.autocrlf false > > and see if it helps > > On Thu, Feb 11, 2021 at 9:21 AM John Cremona <john.c...@gmail.com> wrote: > > > > On Wed, 10 Feb 2021 at 18:07, Dima Pasechnik <dim...@gmail.com> wrote: > > > > > > On Wed, Feb 10, 2021 at 5:49 PM John Cremona <john.c...@gmail.com> > wrote: > > > > > > > > On Wed, 10 Feb 2021 at 16:56, Dima Pasechnik <dim...@gmail.com> > wrote: > > > > > > > > > > On Wed, Feb 10, 2021 at 2:58 PM John Cremona <john.c...@gmail.com> > wrote: > > > > > > > > > > > > Has anyone else been seeing the following problem, which has been > > > > > > plaguing me for a week or two. Here's a simple example. All > > > > > > computers mentioned here are running ubuntu. On a machine I had > not > > > > > > used for a while I had a sage build of the develop branch at an > old > > > > > > version (pre 9.0). There were no modified files (git status > showed > > > > > > nothing). Then I did "git pull trac develop", after which one > file is > > > > > > marked as having changed: > > > > > > > > > > > > modified: src/sage/misc/element_with_label.py > > > > > > > > > > > > It is always this file, on several machines where I have gone > through > > > > > > similar steps. file shows this: > > > > > > src/sage/misc/element_with_label.py: Python script, ASCII text > > > > > > executable, with CRLF line terminators > > > > > > > > > > does > > > > > > > > > > git config core.autolf > > > > > > > > > > show 'true'? > > > > > > > > No. > > > > > > > > > > > > > > Set it to false, IMHO this should fix this issue. > > > > > > > > It is not set at all. I saw that option after googling for help, but > > > > the issue it exists to solve is not one which has ever hit me (in > sage > > > > anyway, of course I know about using dos2unix sometimes when a > windows > > > > users sends a file). > > > > > > did you try setting this option to false, and see if it helps? > > > > Not yet systematically. After setting it to false and doing nothing > > else, git status still shows > > src/sage/misc/element_with_label.py as modified, even after 'git > > checkout --'. Trying to get back to sanity a different way, 'git > > stash' now outputs > > > > > > warning: CRLF will be replaced by LF in > src/sage/misc/element_with_label.py. > > The file will have its original line endings in your working directory. > > warning: CRLF will be replaced by LF in > src/sage/misc/element_with_label.py. > > The file will have its original line endings in your working directory. > > Saved working directory and index state WIP on develop: 8453ffb > > Updated SageMath version to 9.3.beta7 > > HEAD is now at 8453ffb Updated SageMath version to 9.3.beta7 > > > > but git status still shows that file as modified. Now I cannot just > > checkout a different existing branch (since there are apparently > > modified files), but I can checkout a new branch (g = alias for git) > > > > $ g co -b dud > > M src/sage/misc/element_with_label.py > > Switched to a new branch 'dud' > > > > the delete the 'develop' branch (which was the same as upstream anyway): > > > > $ g branch -d develop > > Deleted branch develop (was 8453ffb). > > > > and now recreate the develop branch > > > > $ g remote update trac > > (...) > > $ g co -b develop trac/develop > > M src/sage/misc/element_with_label.py > > Branch develop set up to track remote branch develop from trac. > > Switched to a new branch 'develop' > > > > and we are back to where we started. All of the above was with > > core.autolf set to false. > > > > On the same machine, in a completely new clone made and built > > yesterday, git status shows > > > > $ g st > > On branch develop > > Your branch is up-to-date with 'origin/develop'. > > Changes not staged for commit: > > (use "git add <file>..." to update what will be committed) > > (use "git checkout -- <file>..." to discard changes in working directory) > > > > modified: src/sage/rings/invariants/__init__.py > > > > no changes added to commit (use "git add" and/or "git commit -a") > > > > So the same problem with a new file -- which only has 2 bytes in it! > > AFter dos2unix-ing it (so it now has 1 byte) it is still showing up in > > git status. AFter git checkout -- it goes back to having 2 bytes and > > still shows up. > > > > It seems to be impossible to do any sage development here. I am now > > trying yet another git clone -- so the global config of core.autolf is > > in force as the clone is made -- but then after the clone from > > git://github.com/sagemath/sage.git that same file is shown as > > modified. For some reason I have been making the cones from github > > and then adding trac as a remote. As a variant I made a fresh clone > > from trac -- but see exactly the same. > > > > A similar problem was reported on stackexchange *over 9 years ago* by > > a Mac users whose collaborators were all ubuntu users. There, it is > > suggest to look at the .gitattributes file, and *in this fresh clone* > > that file contains > > > > # Force LF normalization > > * text=auto eol=lf > > # except for Windows batch files > > *.{cmd,[cC][mM][dD]} text eol=crlf > > *.{bat,[bB][aA][tT]} text eol=crlf > > *.diff_bin binary > > > > This is commit 5328570 which is 9.2, master branch, but the develop > > branch has the same. > > > > To me it seems (from reading all that stackexchange page) that this is > > not just a problem caused by something in my local configuration, but > > something which has happened to the repository itself, which needs to > > be fixed. But I am no expert. > > > > Just to add to the fun, on a second machine with the same ubuntu > > version as the one above (16.04.7 LTS) and same git version 2.7.4, a > > clone I made yesterday shows no problems at all, while and older clone > > on which I had done development for a long time has some CRLF file > > issues. The only differences in the global .gitconfig between the two > > machines is the core.autolf which I just added to the first. > > > > I do have some clones without these problems so can keep on working, > > but I also have branches I wanted to do more work on in clones which > > are unusable. > > > > John > > > > > > > > > > > > > > > So, I did not change anything (as far as I can tell -- except get > > > > older) but my git state changed. These files stay changed even after > > > > trying to reset them (though *I* did not change them) using git > > > > checkout --. > > > > > > It's a new version of git that got you, I suppose. > > > > Yes, that is likely, since I did do package updates across the board > > before this started to happen. I am seeing it on machines running: > > ubuntu 16.04.7 LTS and git 2.7.4 > > ubuntu 18.04.5 LTS and git 2.17.1 > > ubuntu 20.04.2 LTS and git 2.25.1 > > > > In no cases was the whole OS updated, just packages so possibly git, > > but I don't know what git version any had before this. > > > > > > We've had some "fun" with this lately, as there are Sage spkgs which > > > need patches for files > > > with CRLF endings, and so these patches had to be dealt with somehow. > > > It was not pleasant. > > > See https://trac.sagemath.org/ticket/30403 > > > > > > > > > > > > > > John > > > > > > > > > > > > > > > > > > > > > > > > > > git diff shows this: > > > > > > > > > > > > $ git diff -w src/sage/misc/element_with_label.py > > > > > > warning: CRLF will be replaced by LF in > src/sage/misc/element_with_label.py. > > > > > > The file will have its original line endings in your working > directory. > > > > > > > > > > > > I cannot get this to go away. git stash does nothing, and I > cannot > > > > > > change branches since git thinks there is an uncommitted change. > I > > > > > > have been reduced to deleting the entire build and starting > again from > > > > > > scratch. > > > > > > > > > > > > No, I have not knowingly changed any configuration on these > computers, > > > > > > though I have done some normal package upgrades using apt. > > > > > > > > > > > > I just went through the same steps in another directory on the > same > > > > > > machine where I had an old build (8.7) and after "git pull trac > > > > > > develop" there are now 5 modified files: > > > > > > > > > > > > modified: > > > > > > > build/pkgs/sage_sws2rst/src/test/Adding_Pictures_and_screenshots.sws > > > > > > modified: src/doc/en/reference/combinat/media/k-rim.JPG > > > > > > modified: src/sage/ext_data/doctest/rich_output/example.png > > > > > > modified: src/sage/ext_data/notebook-ipython/logo-64x64.png > > > > > > modified: src/sage/misc/element_with_label.py > > > > > > > > > > > > -- the same one as before and 3 others which are not even normal > text > > > > > > files. I don't want to trash all these since I am trying to > resurrect > > > > > > and complete work done on some old trac tickets. > > > > > > > > > > > > John > > > > > > > > > > > > -- > > > > > > You received this message because you are subscribed to the > Google Groups "sage-devel" group. > > > > > > To unsubscribe from this group and stop receiving emails from > it, send an email to sage-devel+...@googlegroups.com. > > > > > > To view this discussion on the web visit > https://groups.google.com/d/msgid/sage-devel/CAD0p0K7KrVUvvkmqt6Nwhe1Tt9ef_w3ZWAqU5GK6Y6%3Dysqq3Uw%40mail.gmail.com > . > > > > > > > > > > -- > > > > > You received this message because you are subscribed to the Google > Groups "sage-devel" group. > > > > > To unsubscribe from this group and stop receiving emails from it, > send an email to sage-devel+...@googlegroups.com. > > > > > To view this discussion on the web visit > https://groups.google.com/d/msgid/sage-devel/CAAWYfq10WrFPb-2wQfLVyhSSRRGdWr_eEQ7gyanQ9Pbgm6o6qQ%40mail.gmail.com > . > > > > > > > > -- > > > > You received this message because you are subscribed to the Google > Groups "sage-devel" group. > > > > To unsubscribe from this group and stop receiving emails from it, > send an email to sage-devel+...@googlegroups.com. > > > > To view this discussion on the web visit > https://groups.google.com/d/msgid/sage-devel/CAD0p0K41axYDsedDj4-Dc7TkANM7T3CmVyMmS%3DDHhEXQQZqn4A%40mail.gmail.com > . > > > > > > -- > > > You received this message because you are subscribed to the Google > Groups "sage-devel" group. > > > To unsubscribe from this group and stop receiving emails from it, send > an email to sage-devel+...@googlegroups.com. > > > To view this discussion on the web visit > https://groups.google.com/d/msgid/sage-devel/CAAWYfq19RqXJ1oNgiTo3hX0yV06dXtMEenu-yuXjoM5S9XLhfg%40mail.gmail.com > . > > > > -- > > You received this message because you are subscribed to the Google > Groups "sage-devel" group. > > To unsubscribe from this group and stop receiving emails from it, send > an email to sage-devel+...@googlegroups.com. > > To view this discussion on the web visit > https://groups.google.com/d/msgid/sage-devel/CAD0p0K4PgJG0PsV1cL8agta4262O5PU%2B7QqeESdSot%3D7-x1dpw%40mail.gmail.com > . > -- You received this message because you are subscribed to the Google Groups "sage-devel" group. To unsubscribe from this group and stop receiving emails from it, send an email to sage-devel+unsubscr...@googlegroups.com. To view this discussion on the web visit https://groups.google.com/d/msgid/sage-devel/286d96a6-827e-4ffa-89a2-eda4c29ab8abn%40googlegroups.com.
