[PATCH v6 0/5] Fix and extend encoding handling in fast export/import

While stress testing `git filter-repo`, I noticed an issue with encoding; further digging led to the fixes and features in this series. See the individual commit messages for details.

Changes since v5 (full range-diff below):
  * s/utf-8/UTF-8/, as pointed out by Torsten (in commit messages and comments)
  * small code cleanup pointed out by Eric
  * rewrap the first commit message

Elijah Newren (5):
  t9350: fix encoding test to actually test reencoding
  fast-import: support 'encoding' commit header
  fast-export: avoid stripping encoding header if we cannot reencode
  fast-export: differentiate between explicitly UTF-8 and implicitly
    UTF-8
  fast-export: do automatic reencoding of commit messages only if
    requested

 Documentation/git-fast-export.txt            |  7 ++
 Documentation/git-fast-import.txt            |  7 ++
 builtin/fast-export.c                        | 55 ++++++++++++--
 fast-import.c                                | 11 ++-
 t/t9300-fast-import.sh                       | 20 +++++
 t/t9350-fast-export.sh                       | 78 +++++++++++++++++---
 t/t9350/broken-iso-8859-7-commit-message.txt |  1 +
 t/t9350/simple-iso-8859-7-commit-message.txt |  1 +
 8 files changed, 163 insertions(+), 17 deletions(-)
 create mode 100644 t/t9350/broken-iso-8859-7-commit-message.txt
 create mode 100644 t/t9350/simple-iso-8859-7-commit-message.txt

Range-diff:
1:  37a68a0ffd ! 1:  b5dcdab662 t9350: fix encoding test to actually test reencoding
    @@ -2,13 +2,13 @@
     
         t9350: fix encoding test to actually test reencoding
     
    -    This test used an author with non-ascii characters in the name, but
    -    no special commit message.  It then grep'ed for those non-ascii
    -    characters, but those are guaranteed to exist regardless of the
    -    reencoding process since the reencoding only affects the commit message,
    -    not the author or committer names.  As such, the test would work even if
    -    the re-encoding process simply stripped the commit message entirely.
    -    Modify the test to actually check that the reencoding in utf-8 worked.
    +    This test used an author with non-ascii characters in the name, but no
    +    special commit message.  It then grep'ed for those non-ascii characters,
    +    but those are guaranteed to exist regardless of the reencoding process
    +    since the reencoding only affects the commit message, not the author or
    +    committer names.  As such, the test would work even if the re-encoding
    +    process simply stripped the commit message entirely.  Modify the test to
    +    actually check that the reencoding into UTF-8 worked.
     
         Signed-off-by: Elijah Newren <newren@gmail.com>
     
    @@ -40,7 +40,7 @@
     +		 # The commit object, if not re-encoded, would be 240 bytes.
     +		 # Removing the "encoding iso-8859-7\n" header drops 20 bytes.
     +		 # Re-encoding the Pi character from \xF0 (\360) in iso-8859-7
    -+		 # to \xCF\x80 (\317\200) in utf-8 adds a byte.  Check for
    ++		 # to \xCF\x80 (\317\200) in UTF-8 adds a byte.  Check for
     +		 # the expected size.
     +		 test 221 -eq "$(git cat-file -s i18n)" &&
     +		 # ...and for the expected translation of bytes.
2:  3d84f4613d ! 2:  af7d4e18fa fast-import: support 'encoding' commit header
    @@ -2,7 +2,7 @@
     
         fast-import: support 'encoding' commit header
     
    -    Since git supports commit messages with an encoding other than utf-8,
    +    Since git supports commit messages with an encoding other than UTF-8,
         allow fast-import to import such commits.  This may be useful for folks
         who do not want to reencode commit messages from an external system, and
         may also be useful to achieve reversible history rewrites (e.g. sha1sum
3:  baa8394a3a ! 3:  d5b300692a fast-export: avoid stripping encoding header if we cannot reencode
    @@ -3,8 +3,8 @@
         fast-export: avoid stripping encoding header if we cannot reencode
     
         When fast-export encounters a commit with an 'encoding' header, it tries
    -    to reencode in utf-8 and then drops the encoding header.  However, if it
    -    fails to reencode in utf-8 because e.g. one of the characters in the
    +    to reencode in UTF-8 and then drops the encoding header.  However, if it
    +    fails to reencode in UTF-8 because e.g. one of the characters in the
         commit message was invalid in the old encoding, then we need to retain
         the original encoding or otherwise we lose information needed to
         understand all the other (valid) characters in the original commit
4:  49960164c6 ! 4:  2cef40c613 fast-export: differentiate between explicitly utf-8 and implicitly utf-8
    @@ -1,11 +1,11 @@
     Author: Elijah Newren <newren@gmail.com>
     
    -    fast-export: differentiate between explicitly utf-8 and implicitly utf-8
    +    fast-export: differentiate between explicitly UTF-8 and implicitly UTF-8
     
         The find_encoding() function returned the encoding used by a commit
    -    message, returning a default of git_commit_encoding (usually utf-8).
    +    message, returning a default of git_commit_encoding (usually UTF-8).
         Although the current code does not differentiate between a commit which
    -    explicitly requested utf-8 and one where we just assume utf-8 because no
    +    explicitly requested UTF-8 and one where we just assume UTF-8 because no
         encoding is set, it will become important when we try to preserve the
         encoding header.  Since is_encoding_utf8() returns true when passed
         NULL, we can just return NULL from find_encoding() instead of returning
5:  d8be4ee826 ! 5:  d18f03d1bf fast-export: do automatic reencoding of commit messages only if requested
    @@ -62,7 +62,7 @@
     +		reencode_mode = REENCODE_YES;
     +		break;
     +	default:
    -+		if (arg && !strcasecmp(arg, "abort"))
    ++		if (!strcasecmp(arg, "abort"))
     +			reencode_mode = REENCODE_ABORT;
     +		else
     +			return error("Unknown reencoding mode: %s", arg);
    @@ -156,7 +156,7 @@
     +		 # The commit object, if not re-encoded, is 240 bytes.
     +		 # Removing the "encoding iso-8859-7\n" header would drops 20
     +		 # bytes.  Re-encoding the Pi character from \xF0 (\360) in
    -+		 # iso-8859-7 to \xCF\x80 (\317\200) in utf-8 adds a byte.
    ++		 # iso-8859-7 to \xCF\x80 (\317\200) in UTF-8 adds a byte.
     +		 # Check for the expected size...
     +		 test 240 -eq "$(git cat-file -s i18n-no-recoding)" &&
     +		 # ...as well as the expected byte.