fast-import doesn't have a syntax to support tree-objects (and some other
object-types), so fast-export shouldn't handle them. However, aborting the
operation is a bit drastic. This patch turns the error into a warning instead.
Signed-off-by: Erik Faye-Lund <kusmabite@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
Ideally only errors should be output in this mode so fetch
can be run from cron and normally produce no output. Without
this change it would output a single line on each git commit,
e.g.
r1909 = 32ef87860662526d4a62f903949ed21e0341079e (u2_10_12_branch)
Signed-off-by: Simon Arlott <simon@fire.lp0.eu> Acked-by: Eric Wong <normalperson@yhbt.net>
git-svn: fix ls-tree usage with dash-prefixed paths
To find the blob object name given a tree and pathname, we were
incorrectly calling "git ls-tree" with a "--" argument followed
by the pathname of the file we wanted to get.
git ls-tree <TREE> -- --dashed/path/name.c
Unlike many command-line interfaces, the "--" alone does not
symbolize the end of non-option arguments on the command-line.
ls-tree interprets the "--" as a prefix to match against, thus
the entire contents of the --dashed/* hierarchy would be
returned because the "--" matches "--dashed" and every path
under it.
Thanks to Anton Gyllenberg for pointing me toward the
Twisted repository as a real-world example of this case.
set_shared_perm(): sometimes we know what the final mode bits should look like
adjust_shared_perm() first obtains the mode bits from lstat(2), expecting
to find what the result of applying user's umask is, and then tweaks it
as necessary. When the file to be adjusted is created with mkstemp(3),
however, the mode thusly obtained does not have anything to do with user's
umask, and we would need to start from 0444 in such a case and there is no
point running lstat(2) for such a path.
This introduces a new API set_shared_perm() to bypass the lstat(2) and
instead force setting the mode bits to the desired value directly.
adjust_shared_perm() becomes a thin wrapper to the function.
move_temp_to_file(): do not forget to chmod() in "Coda hack" codepath
Now move_temp_to_file() is responsible for doing everything that is
necessary to turn a tempfile in $GIT_DIR into its final form, it must make
sure "Coda hack" codepath correctly makes the file read-only.
format-patch supports the format.headers configuration for adding
arbitrary email headers to the patches it outputs. This patch adds
support for an --add-header argument which makes the same feature
available from the command line. This is useful when the content of
custom email headers must change from branch to branch.
This patch has been sponsored by Grant Street Group
Signed-off-by: Michael Hendricks <michael@ndrix.org> Signed-off-by: Junio C Hamano <gitster@pobox.com>
When writing out a loose object or a pack (index), move_temp_to_file() is
called to finalize the resulting file. These files (loose files and packs)
should all have permission mode 0444 (modulo adjust_shared_perm()).
Therefore, instead of doing chmod(foo, 0444) explicitly from each callsite
(or even forgetting to chmod() at all), do the chmod() call from within
move_temp_to_file().
Signed-off-by: Johan Herland <johan@herland.net> Signed-off-by: Junio C Hamano <gitster@pobox.com>
"core.sharedrepository = 0mode" should set, not loosen
This fixes the behaviour of octal notation to how it is defined in the
documentation, while keeping the traditional "loosen only" semantics
intact for "group" and "everybody".
Three main points of this patch are:
- For an explicit octal notation, the internal shared_repository variable
is set to a negative value, so that we can tell "group" (which is to
"OR" in 0660) and 0660 (which is to "SET" to 0660);
- git-init did not set shared_repository variable early enough to affect
the initial creation of many files, notably copied templates and the
configuration. We set it very early when a command-line option
specifies a custom value.
- Many codepaths create files inside $GIT_DIR by various ways that all
involve mkstemp(), and then call move_temp_to_file() to rename it to
its final destination. We can add adjust_shared_perm() call here; for
the traditional "loosen-only", this would be a no-op for many codepaths
because the mode is already loose enough, but with the new behaviour it
makes a difference.
Test for local branches being followed with --track
According to the documentation, it is perfectly okay to follow local
branches using the --track option. Introduce a test which checks whether
they behave the same. Currently one test fails.
Signed-off-by: Michael J Gruber <git@drmicha.warpmail.net> Signed-off-by: Junio C Hamano <gitster@pobox.com>
Users were confused about the meaning and use of the --root option.
Notably, since 68c2ec7 (format-patch: show patch text for the root
commit, 2009-01-10), --root has nothing to do with showing the patch
text for the root commit any more.
Shorten and clarify the corresponding paragraph in the DESCRIPTION
section, document --root under OPTIONS, and add an explicit note that
root commits are formatted regardless.
Signed-off-by: Thomas Rast <trast@student.ethz.ch> Signed-off-by: Junio C Hamano <gitster@pobox.com>
Documentation: option to render literal text as bold for manpages
This allows manpages viewed on a tty to render inline literal
text in a manner that is distinct from the surrounding text.
The initial implementation (pre-mailing-list) of this patch
included a conditional variant of the XSLT code in
manpage-base.xsl and use xmlto's --stringparam option to
optionally enable the functionality. It turns out that
--stringparam is broken in all versions of xmlto except for the
pre-release, SVN version. Since xmlto is a shell script the patch
to fix it is simple enough, but I instead opted to use xmlto's
"module" functionality.
Testing done with asciidoc 8.3.1 and docbook-xsl 1.74.0.
Signed-off-by: Chris Johnsen <chris_johnsen@pobox.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
Documentation: asciidoc.conf: fix verse block with block titles
No files use the variant of block-title with verse-block, but
such a case would have generated broken docbook XML (<simpara> is
not allowed inside <para>). This fixes the potential deviation from
valid docbook XML.
Testing done with asciidoc 8.3.1 and docbook-xsl 1.74.0.
Signed-off-by: Chris Johnsen <chris_johnsen@pobox.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
Documentation: asciidoc.conf: always use <literallayout> for [blocktext]
Make the docbook-xsl-no-raw-roff variant match the
no-docbook-xsl-no-raw-roff variant in terms of which XML tag is
used to wrap listing block text (delimited with lines of dashes).
e920b56 (Tweak asciidoc output to work with broken docbook-xsl,
2006-03-05) says docbook-xsl 1.68 needs <literallayout>. This
<screen> usages was in the old, 1.72-only section. But since it
is now the "roff-less" section, it probably makes sense to make it
symmetric with the "roff-ful" section.
Testing done with asciidoc 8.3.1 and docbook-xsl 1.74.0.
Signed-off-by: Chris Johnsen <chris_johnsen@pobox.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
Documentation: move quieting params into manpage-base.xsl
Move a couple of XSL parameters that act to silence
informational/warning messages generated when running xmlto from
manpage-1.72.xsl to manpage-base.xsl.
Since unused parameters are silently ignored, there is no problem
if some version of docbook-xsl does not know about these
parameters. The only problem might be if a version of docbook-xsl
uses the parameters for alternate functionality. Since both
parameters have fairly specific names such a situation is
unlikely.
Testing done with asciidoc 8.3.1 and docbook-xsl 1.74.0.
Signed-off-by: Chris Johnsen <chris_johnsen@pobox.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
Documentation: rename docbook-xsl-172 attribute to git-asciidoc-no-roff
It seems that the ability to use raw roff codes in asciidoc.conf
was eliminated by docbook-xsl 1.72.0 _and later_. Unlike the
1.72.0-specific XSLT problem, this behavior was not reverted in
later releases.
This patch aims to make it clear that the affected asciidoc
attribute (flag) can be reasonably used with docbook-xsl versions
other than 1.72.0.
Also, document which make variables should be set for various
versions of asciidoc and docbook-xsl.
Testing done with asciidoc 8.3.1 and docbook-xsl 1.74.0.
Signed-off-by: Chris Johnsen <chris_johnsen@pobox.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
Documentation: move callouts.xsl to manpage-{base,normal}.xsl
Each of manpage-base.xsl and manpage-normal.xsl gets a copy of
the contents of callouts.xsl and the original is removed. The
Makefile is adjusted to refer to manpage-normal.xsl instead of
callouts.xsl. manpage-base.xsl will be later made into a common
base for -normal and -1.72.
Testing done with asciidoc 8.3.1 and docbook-xsl 1.74.0.
Signed-off-by: Chris Johnsen <chris_johnsen@pobox.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
Most shells define the exit value of a pipeline as the exit value
of the last process. For each texi rule, run the DOCBOOK2X_TEXI
tool and the "fixup" script in their own non-pipeline commands so
that make will notice an error exit code.
Signed-off-by: Chris Johnsen <chris_johnsen@pobox.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
* mg/http-auth:
http-push.c: use a faux remote to pass to http_init
Do not name "repo" struct "remote" in push_http.c
http.c: CURLOPT_NETRC_OPTIONAL is not available in ancient versions of cURL
http authentication via prompts
http_init(): Fix config file parsing
http.c: style cleanups
* jc/attributes-checkout:
Add a test for checking whether gitattributes is honored by checkout.
Read attributes from the index that is being checked out
* fg/push-default:
builtin-push.c: Fix typo: "anythig" -> "anything"
Display warning for default git push with no push.default config
New config push.default to decide default behavior for push
* dm/maint-docco:
Documentation: Remove spurious uses of "you" in git-bisect.txt.
Documentation: minor grammatical fix in git-check-ref-format.txt
Documentation: minor grammatical fixes in git-check-attr.txt
Documentation: minor grammatical fixes in git-cat-file.txt
Documentation: minor grammatical fixes and rewording in git-bundle.txt
Documentation: remove some uses of the passive voice in git-bisect.txt
When exporting a subset of commits on a branch that do not go back to a
root commit (e.g. master~2..master), we still want each exported commit to
have the same files in the exported tree as in the original tree.
Previously, when given such a range, we would omit master~2 as a parent of
master~1, but we would still diff against master~2 when selecting the list
of files to include in master~1. This would result in only files that
had changed in the given range showing up in the resulting export. In such
cases, we should diff master~1 against the root instead (i.e. use
diff_root_tree_sha1 instead of diff_tree_sha1).
There's a special case to consider here: incremental exports (i.e. exports
where the --import-marks flag is specified). If master~2 is an imported
mark, then we still want to diff master~1 against master~2 when selecting
the list of files to include.
We can handle all cases, including the special case, by just checking
whether master~2 corresponds to a known object mark when deciding what to
diff against.
Signed-off-by: Elijah Newren <newren@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
The implementation of exec on Windows is just a rough approximation of the
POSIX behavior. In particular, no real process "overlay" happens (a new
process is spawned instead and the parent process waits until the child
terminates). In particular, the process ID cannot be taken by the exec'd
process. But there is one test in t7502-commit.sh that depends on this.
We have to skip it on Windows.
t7005-editor: Use $SHELL_PATH in the editor scripts
The test sets up various shell scripts and uses them as commit message
editors. On Windows, we need a shebang line in order to recognize the
files as executable shell scripts. This adds it.
diff --no-index: Do not generate patch output if other output is requested
Previously, 'git diff --no-index --stat a b' generated patch output in
addition to the --stat output (or whatever other output format was
requested). Now only the requested output is generated, and patch
output remains the default.
Signed-off-by: Johannes Sixt <j6t@kdbg.org> Signed-off-by: Junio C Hamano <gitster@pobox.com>
* maint:
Increase the size of the die/warning buffer to avoid truncation
close_sha1_file(): make it easier to diagnose errors
avoid possible overflow in delta size filtering computation
Add warning about known issues to documentation of cvsimport
The described issues are compiled from the tests by Michael Haggerty and me.
Because it is not apparent that these can be fixed anytime soon at least warn
unwary users not to rely on the inbuilt cvsimport to much.
Signed-off-by: Heiko Voigt <hvoigt@hvoigt.net> Signed-off-by: Junio C Hamano <gitster@pobox.com>
completion: add --thread=deep/shallow to format-patch
Signed-off-by: Stephen Boyd <bebarino@gmail.com> Trivially-Acked-By: Shawn O. Pearce <spearce@spearce.org> Signed-off-by: Junio C Hamano <gitster@pobox.com>
completion: add --cc and --no-attachment option to format-patch
Signed-off-by: Stephen Boyd <bebarino@gmail.com> Trivially-Acked-By: Shawn O. Pearce <spearce@spearce.org> Signed-off-by: Junio C Hamano <gitster@pobox.com>
Signed-off-by: Stephen Boyd <bebarino@gmail.com> Trivially-Acked-By: Shawn O. Pearce <spearce@spearce.org> Signed-off-by: Junio C Hamano <gitster@pobox.com>
We already skip over loose refs under $GIT_DIR/refs if the name
ends with ".lock", so creating a branch named "foo.lock" will not
appear in the output of "git branch", "git for-each-ref", nor will
its commit be considered reachable by "git rev-list --all".
In the latter case this is especially evil, as it may cause
repository corruption when objects reachable only through such a
ref are deleted by "git prune".
It should be reasonably safe to deny use of ".lock" as a ref suffix.
In prior versions of Git such branches would be "phantom branches";
you can create it, but you can't see it in "git branch" output.
Signed-off-by: Shawn O. Pearce <spearce@spearce.org> Signed-off-by: Junio C Hamano <gitster@pobox.com>
Increase the size of the die/warning buffer to avoid truncation
Long messages like those from lockfile.c when a lock can't be
obtained truncate with only 256 bytes in the message buffer.
Bump it to 1024 to give more space for these longer cases.
Signed-off-by: Shawn O. Pearce <spearce@spearce.org> Signed-off-by: Junio C Hamano <gitster@pobox.com>
MinGW: Quote arguments for subprocesses that contain a single-quote
Before a process can be spawned by mingw_spawnve, arguments must be
surrounded by double-quotes if special characters are present. This is
necessary because the startup code of the spawned process will expand
arguments that look like glob patterns. "Normal" Windows command line
utilities expand only * and ?, but MSYS programs, including bash, are
different: They also expand braces, and this has already been taken care
of by compat/mingw.c:quote_arg().
But MSYS programs also treat single-quotes in a special way: Arguments
between single-quotes are spliced together (with spaces) into a word.
With this patch this treatment is avoided by quoting arguments that contain
single-quotes.
This lets t4252 pass on Windows.
Signed-off-by: Johannes Sixt <j6t@kdbg.org> Signed-off-by: Junio C Hamano <gitster@pobox.com>
* js/windows-tests:
t0060: fix whitespace in "wc -c" invocation
t5503: GIT_DEBUG_SEND_PACK is not supported on MinGW
t7004: Use prerequisite tags to skip tests that need gpg
Use prerequisites to skip tests that need unzip
t3700: Skip a test with backslashes in pathspec
Skip tests that require a filesystem that obeys POSIX permissions
t0060: Fix tests on Windows
Use prerequisite tags to skip tests that depend on symbolic links
t9100, t9129: Use prerequisite tags for UTF-8 tests
t5302: Use prerequisite tags to skip 64-bit offset tests
Skip tests that fail if the executable bit is not handled by the filesystem
t3600: Use test prerequisite tags
test-lib: Infrastructure to test and check for prerequisites
t0050: Check whether git init detected symbolic link support correctly
Tests on Windows: $(pwd) must return Windows-style paths
test-lib: Work around missing sum on Windows
test-lib: Work around incompatible sort and find on Windows
avoid possible overflow in delta size filtering computation
On a 32-bit system, the maximum possible size for an object is less than
4GB, while 64-bit systems may cope with larger objects. Due to this
limitation, variables holding object sizes are using an unsigned long
type (32 bits on 32-bit systems, or 64 bits on 64-bit systems).
When large objects are encountered, and/or people play with large delta
depth values, it is possible for the maximum allowed delta size
computation to overflow, especially on a 32-bit system. When this
occurs, surviving result bits may represent a value much smaller than
what it is supposed to be, or even zero. This prevents some objects
from being deltified although they do get deltified when a smaller depth
limit is used. Fix this by always performing a 64-bit multiplication.
Signed-off-by: Nicolas Pitre <nico@cam.org> Signed-off-by: Junio C Hamano <gitster@pobox.com>
Some people and foreign SCM converter may have named their branches
as frotz@24 and we still want to keep supporting it.
However, "git branch frotz@{24}" is a disaster. It cannot even
checked out because "git checkout frotz@{24}" will interpret it as
"detach the HEAD at twenty-fourth reflog entry of the frotz branch".
(2) a refname that ends with a dot.
We already reject a path component that begins with a dot, primarily
to avoid ambiguous range interpretation. If we allowed ".B" as a
valid ref, it is unclear if "A...B" means "in dot-B but not in A" or
"either in A or B but not in both".
But for this to be complete, we need also to forbid "A." to avoid "in
B but not in A-dot". This was not a problem in the original range
notation, but we should have added this restriction when three-dot
notation was introduced.
Unlike "no dot at the beginning of any path component" rule, this
rule does not have to be "no dot at the end of any path component",
because you cannot abbreviate the tail end away, similar to you can
say "dot-B" to mean "refs/heads/dot-B".
For these reasons, it is not likely people created branches with these
names on purpose, but we have allowed such names to be used for quite some
time, and it is possible that people created such branches by mistake or
by accident.
To help people with branches with such unfortunate names to recover,
we still allow "branch -d 'bad.'" to delete such branches, and also allow
"branch -m bad. good" to rename them.
http-push: using error() and warning() as appropriate
Change three occurences of using inconsistent error/warning reporting by
using the relevant error() / warning() calls to be consitent with the
rest of the code.
Signed-off-by: Miklos Vajna <vmiklos@frugalware.org> Signed-off-by: Junio C Hamano <gitster@pobox.com>
The command is supposed to rename the branch we were on before switched
from to a new name, but was not aware of the short-hand notation we added
recently.
These allow you to say "git checkout @{-2}" to switch to the branch two
"branch switching" ago by pretending as if you typed the name of that
branch. As it is likely that we will be introducing more short-hands to
write the name of a branch without writing it explicitly, rename the
functions from "nth_last_branch" to more generic "branch_name", to prepare
for different semantics.
will create two messages with files 1 and 2 attached respectively.
There is no effect when using --numbered-files and --stdout together
without an --attach or --inline, the --numbered-files option will be
ignored. Add a test to show this.
Signed-off-by: Stephen Boyd <bebarino@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
format-patch: --attach/inline uses filename instead of SHA1
Currently when format-patch is used with --attach or --inline the patch
attachment has the SHA1 of the commit for its filename. This replaces
the SHA1 with the filename used by format-patch when outputting to
files.
Fix tests relying on the SHA1 output and add a test showing how the
--suffix option affects the attachment filename output.
Signed-off-by: Stephen Boyd <bebarino@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
We use the commit to generate the patch filename in reopen_stdout()
before we redirect stdout. The cover letter codepath creates a dummy
commit with the desired subject line 'cover letter'.
Signed-off-by: Stephen Boyd <bebarino@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
format-patch: construct patch filename in one function
reopen_stdout() usually takes the oneline subject of a commit,
appends the patch suffix, prepends the output directory (if any) and
then reopens stdout as the resulting file. Now the patch filename (the
oneline subject and the patch suffix) is created in
get_patch_filename() and passed to reopen_stdout() which prepends the
output directory and reopens stdout as that file.
The original function to get the oneline description,
get_oneline_for_filename(), has been renamed to get_patch_filename() to
reflect its new functionality.
Signed-off-by: Stephen Boyd <bebarino@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
pretty.c: add %f format specifier to format_commit_message()
This specifier represents the sanitized and filename friendly subject
line of a commit. No checks are made against the length of the string,
so users may need to trim the result to the desired length if using as a
filename. This is commonly used by format-patch to massage commit
subjects into filenames and output patches to files.
Signed-off-by: Stephen Boyd <bebarino@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
Documentation: minor grammatical fixes and rewording in git-bundle.txt
This commit also converts all reference specifications to a monospaced font,
as the embedded ~ character used in some of the references sometimes causes
the text up to the next ~ to be displayed incorrectly as a subscript when the
HTML pages are generated. This was tested with asciidoc 8.2.5.
Signed-off-by: David J. Mellor <dmellor@whistlingcat.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>