I have made https://trac.sagemath.org/ticket/31379 to get rid of the CRLF
in these 3 files.
Le jeudi 11 février 2021 à 10:43:27 UTC+1, Frédéric Chapoton a écrit :
> I have been hit by this kind of annoying issue some time ago. No idea what
> is happening.
> There are only 3 offending files containing CRLF:
>
> sage$ git grep -Il $'\r' src/
> src/sage/combinat/crystals/star_crystal.py
> src/sage/misc/element_with_label.py
> src/sage/rings/invariants/__init__.py
>
> We should probably change them to use the correct line break
> Le jeudi 11 février 2021 à 10:21:26 UTC+1, john.c...@gmail.com a écrit :
>
>> 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+unsubscr...@googlegroups.com.
To view this discussion on the web visit
https://groups.google.com/d/msgid/sage-devel/0e364799-33a8-4623-8faa-c0d8a17f5ac4n%40googlegroups.com.
