Skip to contentSkip to conversation

Git commit behaving strangely

PATCH Zeeshan Qureshi 2 participants 5 messages Jun 15, 2016 2 messages not in the archive

lore
Share & tools
lore
#0 lore raw

Hello,

I am new to git and was going through gittutorial-2 when i noticed something weird.

1. I first create a file file.txt with some text and committed it.
2. I create a file called closing.txt with some text and committed it.
3. then i added 1 new line to both closing.txt and file.txt
4. i added the file closing.txt to the staging are by running 'git add closing.txt'
5. now i ran 'git checkout' (by mistake)
6. Now i run 'git diff --cached' (which should show the changes i have in the index), the output is this

closing.txt +1 -0
index da55920..953ec3a 100644
@@ -1 +1,2 @@
 closing
+again?

7. Now when i run 'git commit -m "closing again"', the change is not committed even though i get a new commit id.

The output of 'git log -p' is this (as you can clearly see, there are new commits but no change has been recorded)

commit 7dbe9a272c58decc08f9fd5ce07efd5bf28d89f8

Author: Zeeshan Qureshi <{email}>
Date:   Tue Jul 20 15:36:05 2010 -0400

closing again

commit 1bfda310eb82e62925b28d62e98097d6fbd61620

Author: Zeeshan Qureshi <{email}>
Date:   Tue Jul 20 15:34:46 2010 -0400

closing again

commit 3d0e35717bea56adbc7c578b3b17db3657e30040

Author: Zeeshan Qureshi <{email}>
Date:   Tue Jul 20 15:33:40 2010 -0400

closing

closing.txt new file +1 -0
new file mode 100644
index 0000000..da55920
@@ -0,0 +1 @@
+closing

commit 2a0a0e4edb1810b6e31b0739bd7736d1360e6d13

Author: Zeeshan Qureshi <{email}>
Date:   Tue Jul 20 15:32:13 2010 -0400

initial commit

file.txt new file +1 -0
new file mode 100644
index 0000000..3b18e51
@@ -0,0 +1 @@
+hello world
Signature · 2 lines
Zeeshan Qureshi
[email protected]
#1 lore raw

Hi Zeeshan,

Quoted 4 lines · Zeeshan Qureshi

3. then i added 1 new line to both closing.txt and file.txt
4. i added the file closing.txt to the staging are by running 'git add closing.txt'
5. now i ran 'git checkout' (by mistake)

[...]

Quoted 5 lines

7. Now when i run 'git commit -m "closing again"', the change is not committed even though i get a new commit id.

The output of 'git log -p' is this (as you can clearly see, there are new commits but no change has been recorded)

Strange. I tried to reproduce this but I got different results (the new commit includes in a change to closing.txt). What am I doing wrong?

git init repo
cd repo

echo 'hello world' >file.txt
git add .
git commit -m 'Initial version of repo, a test repository'

echo closing >closing.txt
git add closing.txt
git commit -m 'Goodbye, all'

echo 'hi!' >>file.txt
echo 'again?' >>closing.txt
git add closing.txt

git checkout
git diff --cached

git commit -m 'closing again'
git log -p

#4 lore raw
In reply to a message not in the archive

Hello Jonathan,

Quoted 9 lines · Jonathan Nieder
Quoted 2 lines · Zeeshan Qureshi

I will do it again when I get to work tomorrow (i don't have remote access to the workstation) and post the results.

Sounds good.  Thanks for looking into it.

Jonathan

I did the test again on the Solaris machine, and encountered the same problem.

I am on a Sun Blade 1500 machine running Solaris 10. Zeeshan

#5 lore raw
Quoted 6 lines · Zeeshan Qureshi
Quoted 6 lines
Quoted 6 lines
Quoted 6 lines
Quoted 4 lines · Zeeshan Qureshi

3. then i added 1 new line to both closing.txt and file.txt
4. i added the file closing.txt to the staging are by running 'git add closing.txt'
5. now i ran 'git checkout' (by mistake)

[...]

Quoted 5 lines
Quoted 5 lines
Quoted 5 lines
Quoted 5 lines
Quoted 5 lines

7. Now when i run 'git commit -m "closing again"', the change is not committed even though i get a new commit id.

The output of 'git log -p' is this (as you can clearly see, there are new commits but no change has been recorded)

[...]

Quoted 3 lines

I did the test again on the Solaris machine, and encountered the same problem.

I am on a Sun Blade 1500 machine running Solaris 10.

Since this seems to be hard to replicate without blastwave, I would suggest reporting it through the blastwave forum.

Thanks, and good luck.

#6 lore raw
In reply to Jonathan Nieder

Hello Jonathan,

The problem seems to be with the Blastwave git package (it 1.6.x) on Solaris.

I tried reproducing it on 2 different linux machines (both running git 1.7) and I couldn't reproduce it.

I will do it again when I get to work tomorrow (i don't have remote access to the workstation) and post the results.

Thanks,
Zeeshan