git bug?
Just upgraded to the latest git, pulled the latest Linux kernel tree, and made a local modification. Here is the strange git-diff-cache output:
Quoted 6 lines
[jgarzik@pretzel libata-dev]$ git-diff-cache -p HEAD diff --git
a/arch/arm/mm/minicache.cb/arch/arm/mm/minicache.cdeleted file mode 100644
drivers/scsi/libata-core.c static void __ata_qc_complete(struct ata · line 3064
Quoted 8 lines
drivers/scsi/libata-core.c +0
-2@@ -3059,8 +3059,6 @@ static void __ata_qc_complete(struct atastruct ata_port *ap = qc->ap;unsigned int tag, do_clear =0 ; -WARN_ON (!assert_spin_locked (&ap->host_set->lock)); -if (likely (qc->flags &ATA_QCFLAG_ACTIVE )) {assert (ap->queue_depth); ap->queue_depth--;
The libata-core part is correct, the arch/arm/mm part is not.
I pulled using unmodified, upstream git scripts. No sugar added. Here's what the repo looks like:
Quoted 9 lines
[jgarzik@pretzel libata-dev]$ git-rev-list --pretty HEAD ^master | git-shortlog Jeff Garzik:
Merge /spare/repo/linux-2.6/
Automatic merge of /spare/repo/linux-2.6/.git branch HEADJens Axboe:
libata: fix spinlock bug introduced by NCQ code
libata: ncq support update
libata-scsi: better placement of cmd completion on err SATA NCQ support
It's mirroring to kernel.org right now, at rsync.kernel.org://...jgarzik/libata-dev.git if somebody wants to poke at it. Top of tree is d032ec9048ff82a704b96b93cfd6f2e8e3a06b19 (when fully uploaded and mirrored).
Jeff
Quoted 27 lines · Jeff Garzik
Just upgraded to the latest git, pulled the latest Linux kernel tree, and made a local modification. Here is the strange git-diff-cache output:
Quoted 6 lines
[jgarzik@pretzel libata-dev]$ git-diff-cache -p HEAD diff --git
a/arch/arm/mm/minicache.cb/arch/arm/mm/minicache.cdeleted file mode 100644drivers/scsi/libata-core.c static void __ata_qc_complete(struct ata · line 3064
Quoted 8 lines
drivers/scsi/libata-core.c +0
-2@@ -3059,8 +3059,6 @@ static void __ata_qc_complete(struct atastruct ata_port *ap = qc->ap;unsigned int tag, do_clear =0 ; -WARN_ON (!assert_spin_locked (&ap->host_set->lock)); -if (likely (qc->flags &ATA_QCFLAG_ACTIVE )) {assert (ap->queue_depth); ap->queue_depth--;The libata-core part is correct, the
arch/arm/mmpart is not.I pulled using unmodified, upstream git scripts. No sugar added. Here's what the repo looks like:
Quoted 4 lines
[jgarzik@pretzel libata-dev]$ git-rev-list --pretty HEAD ^master | git-shortlog Jeff Garzik:
Merge /spare/repo/linux-2.6/
Automatic merge of /spare/repo/linux-2.6/.git branch HEAD
Quoted 8 lines
Quoted 5 lines
Jens Axboe:
libata: fix spinlock bug introduced by NCQ code
libata: ncq support update
libata-scsi: better placement of cmd completion on err SATA NCQ supportIt's mirroring to kernel.org right now, at rsync.kernel.org://...jgarzik/libata-dev.git if somebody wants to poke at it. Top of tree is
d032ec9048ff82a704b96b93cfd6f2e8e3a06b19(when fully uploaded and mirrored).
Well, after switching branches and switching back, the behavior is gone.
<shrug>
Jeff
Jeff, I assume you are talking about this part:
Quoted 3 lines
drivers/scsi/libata-core.c
Quoted 3 lines
arch/arm/mm/minicache.c deleted +0
-0drivers/scsi/libata-core.c +0
-0
that does not have anything between "deleted file..." and the next diff about libata-core.c.
In short, this is not really a "bug", but I am open to suggestions for improvements (but you need to talk Linus into accepting the changes, because changing this would also affect his "git-apply" program as well).
The current "diff-patch" format is generated this way:
(0) if a modification is between different kind of objects (e.g. a regular file changed to a symlink), then the following steps are done twice; one "deletion" of the original, and another "creation" of the result.
(1) "diff --git a/oldname b/newname" line is shown first and it is always shown. Unless it is a rename/copy diff, oldname and newname always match. We never say /dev/null on this line to represent creation/deletion.
(2) Next, optional "diff --git extended headers" appear. See apply.c for the list of things to expect.
(3) Next, the output from "diff -u -L<oldname> -L<newname>" between the two file contents follows. Here, oldname and newname are usually of the form "a/oldname" and "b/oldname", but we _do_ follow /dev/null convention for file creation and deletion here. That is, you would see
--- /dev/null
+++ frotz.c
@@ ...
for a file creation, and the opposite for a file deletion.
Unfortunately, if "diff" finds nothing to report, you would not even see these ---/+++ lines. I think that is what puzzled you.
By the way, this is all documented (?) behaviour, reported first by Linus in this message.
Message-ID: <[email protected]>
From: Linus Torvalds <[email protected]>
Subject: git full diff output issues..
Date: Thu, 26 May 2005 12:19:21 -0700 (PDT)
Cc: Git Mailing List <[email protected]>
While testing my "git-apply" thing (coming along quite nicely, thanks for asking), I've hit a case that is nasty to parse.
This is from the 2.6.12-rc4 -> 2.6.12-rc5 patch:
diff --git a/arch/um/kernel/checksum.c b/arch/um/kernel/checksum.c
deleted file mode 100644
diff --git a/arch/um/kernel/initrd.c b/arch/um/kernel/initrd.c
new file mode 100644
--- /dev/null
+++ b/arch/um/kernel/initrd.c
@@ -0,0 +1,78 @@and the magic here is that deleted file that was empty to begin with, so it didn't have a patch, just a note on deletion.
Why is that nasty? Because we don't have the file _name_ in any good format. The filename only exists int he "diff --git" header, and that one has the space-parsing issue, which makes it less than optimal.
(the rest omitted)
We had a handful of back-and-forth back then and the resolution was this:
Message-ID: <[email protected]>
From: Linus Torvalds <[email protected]>
Subject: Re: git full diff output issues..
Date: Thu, 26 May 2005 13:33:26 -0700 (PDT)
Cc: Junio C Hamano <[email protected]>, Git Mailing List <[email protected]>
(some omitted)
However, I ended up just validating the name parsing by making sure that when I parse the "git --diff" line, I only take the name if I can see it being the same for both the old and the new. ...
(the rest omitted)
By the way, this is not just "deleting empty files". Any change that does not involve content changes you would see something similar to what you found.
Message-ID: <[email protected]>
From: Linus Torvalds <[email protected]>
Subject: Re: git full diff output issues..
Date: Sun, 5 Jun 2005 08:11:02 -0700 (PDT)
Cc: [email protected]
(some omitted)
The only case that was special was literally the "same name, no content changes, new mode" case, which looked like
diff --git a/oldname.c b/oldname.c
new mode 100755
old mode 100644and thus _only_ had the name in the (normally ambiguous wrt whitepsace) header line.
Quoted 11 lines · Junio C Hamano
Jeff, I assume you are talking about this part:
Quoted 3 lines
drivers/scsi/libata-core.c
Quoted 3 lines
arch/arm/mm/minicache.c deleted +0
-0drivers/scsi/libata-core.c +0
-0that does not have anything between "deleted file..." and the next diff about
libata-core.c.
Quoted 4 lines
In short, this is not really a "bug", but I am open to suggestions for improvements (but you need to talk Linus into accepting the changes, because changing this would also affect his "git-apply" program as well).
It's a bug and a new behavior.
'git-diff-cache -p HEAD' should not show files that I have not modified.
This screws up
git-diff-cache -p HEAD | diffstat -p1 | awk '{print $1}'
for example.
Jeff
Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line
"JG" == Jeff Garzik <[email protected]> writes:
JG> 'git-diff-cache -p HEAD' should not show files that I have not modified. JG> This screws up
JG> git-diff-cache -p HEAD | diffstat -p1 | awk '{print $1}'
JG> for example.
Agreed 100%. If there is a file exists (rather, if the git thinks the file is there), then diff-cache should not say it is deleted. That would be wrong.
Sorry for not having a crystal ball to poke into your terminal, so let me get something straight. I am still unsure the problem you are reporting is coming from diff-cache bug, or somehow yesterday's (well it is 01:13AM where I sit ;-)) merge updates trashed your work tree or index file, so these questions would help me tell which. When "diff-cache" reports the file "deleted":
- Do you have that arch/arm/mm/minicache.c file in your working directory?
- Does git-ls-files know about the file? Does the path show in:
git-ls-files --stage | grep arch/arm/mm/minicache.c
- Does the HEAD tree have the path as well? Does it show in:
git-ls-tree -r HEAD | grep arch/arm/mm/minicache.c