Refactor, into a common function, the version and capability negotiation
done when invoking a long-running process as a clean or smudge filter.
This will be useful for other Git code that needs to interact similarly
with a long-running process.
As you can see in the change to t0021, this commit changes the error
message reported when the long-running process does not introduce itself
with the expected "server"-terminated line. Originally, the error
message reports that the filter "does not support filter protocol
version 2", differentiating between the old single-file filter protocol
and the new multi-file filter protocol - I have updated it to something
more generic and useful.
Signed-off-by: Jonathan Tan <jonathantanmy@google.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
* ls/filter-process-delayed:
convert: add "status=delayed" to filter process protocol
convert: refactor capabilities negotiation
convert: move multiple file filter error handling to separate function
convert: put the flags field before the flag itself for consistent style
t0021: write "OUT <size>" only on success
t0021: make debug log file name configurable
t0021: keep filter log files on comparison
Some `clean` / `smudge` filters may require a significant amount of
time to process a single blob (e.g. the Git LFS smudge filter might
perform network requests). During this process the Git checkout
operation is blocked and Git needs to wait until the filter is done to
continue with the checkout.
Teach the filter process protocol, introduced in edcc8581 ("convert: add
filter.<driver>.process option", 2016-10-16), to accept the status
"delayed" as response to a filter request. Upon this response Git
continues with the checkout operation. After the checkout operation Git
calls "finish_delayed_checkout" which queries the filter for remaining
blobs. If the filter is still working on the completion, then the filter
is expected to block. If the filter has completed all remaining blobs
then an empty response is expected.
Git has a multiple code paths that checkout a blob. Support delayed
checkouts only in `clone` (in unpack-trees.c) and `checkout` operations
for now. The optimization is most effective in these code paths as all
files of the tree are processed.
Signed-off-by: Lars Schneider <larsxschneider@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The code to negotiate long running filter capabilities was very
repetitive for new capabilities. Replace the repetitive conditional
statements with a table-driven approach. This is useful for the
subsequent patch 'convert: add "status=delayed" to filter process
protocol'.
Suggested-by: Junio C Hamano <gitster@pobox.com>
Signed-off-by: Lars Schneider <larsxschneider@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
When using the hashmap a common need is to have access to caller provided
data in the compare function. A couple of times we abuse the keydata field
to pass in the data needed. This happens for example in patch-ids.c.
This patch changes the function signature of the compare function
to have one more void pointer available. The pointer given for each
invocation of the compare function must be defined in the init function
of the hashmap and is just passed through.
Documentation of this new feature is deferred to a later patch.
This is a rather mechanical conversion, just adding the new pass-through
parameter. However while at it improve the naming of the fields of all
compare functions used by hashmaps by ensuring unused parameters are
prefixed with 'unused_' and naming the parameters what they are (instead
of 'unused' make it 'unused_keydata').
Signed-off-by: Stefan Beller <sbeller@google.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Refactoring the filter error handling is useful for the subsequent patch
'convert: add "status=delayed" to filter process protocol'.
In addition, replace the parentheses around the empty "if" block with a
single semicolon to adhere to the Git style guide.
Signed-off-by: Lars Schneider <larsxschneider@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Fix configuration codepath to pay proper attention to commondir
that is used in multi-worktree situation, and isolate config API
into its own header file.
* bw/config-h:
config: don't implicitly use gitdir or commondir
config: respect commondir
setup: teach discover_git_directory to respect the commondir
config: don't include config.h by default
config: remove git_config_iter
config: create config.h
Stop including config.h by default in cache.h. Instead only include
config.h in those files which require use of the config system.
Signed-off-by: Brandon Williams <bmwill@google.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Enable sub-processes to gracefully handle when the process dies by
updating subprocess_read_status to return an error on EOF instead of
dying.
Update apply_multi_file_filter to take advantage of the revised
subprocess_read_status.
Signed-off-by: Ben Peart <benpeart@microsoft.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Move the sub-proces functions into sub-process.h/c. Add documentation
for the new module in Documentation/technical/api-sub-process.txt
Signed-off-by: Ben Peart <benpeart@microsoft.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Do a mechanical rename of the functions that will become the reusable
sub-process module.
Signed-off-by: Ben Peart <benpeart@microsoft.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Update all functions that are going to be moved into a reusable module
so that they only work with the reusable data structures. Move code
that is specific to the filter out into the filter specific functions.
Signed-off-by: Ben Peart <benpeart@microsoft.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
To enable future reuse of the filter.<driver>.process infrastructure,
split the cmd2process structure into two separate parts.
subprocess_entry will now contain the generic data required to manage
the creation and tracking of the child process in a hashmap.
cmd2process is a filter protocol specific structure that is used to
track the negotiated capabilities of the filter.
Signed-off-by: Ben Peart <benpeart@microsoft.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
To enable future reuse of the filter.<driver>.process infrastructure,
split start_multi_file_filter() into two separate parts.
start_multi_file_filter() will now only contain the generic logic to
manage the creation and tracking of the child process in a hashmap.
start_multi_file_filter_fn() is a protocol specific initialization
function that will negotiate the multi-file-filter interface version
and capabilities.
Signed-off-by: Ben Peart <benpeart@microsoft.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Add packet_writel() which writes multiple lines in a single call and
then calls packet_flush_gently(). Update convert.c to use the new
packet_writel() function from pkt-line.
Signed-off-by: Ben Peart <benpeart@microsoft.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
start_multi_file_filter() and apply_multi_file_filter() currently test
for errno == EPIPE but treating EPIPE as an error is already happening
from one of the packet_write() functions.
Signed-off-by: Ben Peart <benpeart@microsoft.com>
Found/Fixed-by: Jeff King <peff@peff.net>
Acked-by: Lars Schneider <larsxschneider@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The remaining callers are all simple "I have N attributes I am
interested in. I'll ask about them with various paths one by one".
After this step, no caller to git_check_attrs() remains. After
removing it, we can extend "struct attr_check" struct with data
that can be used in optimizing the query for the specific N
attributes it contains.
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Signed-off-by: Stefan Beller <sbeller@google.com>
Signed-off-by: Brandon Williams <bmwill@google.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The traditional API to check attributes is to prepare an N-element
array of "struct git_attr_check" and pass N and the array to the
function "git_check_attr()" as arguments.
In preparation to revamp the API to pass a single structure, in
which these N elements are held, rename the type used for these
individual array elements to "struct attr_check_item" and rename
the function to "git_check_attrs()".
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Signed-off-by: Stefan Beller <sbeller@google.com>
Signed-off-by: Brandon Williams <bmwill@google.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Fix a corner case in merge-recursive regression that crept in
during 2.10 development cycle.
* jc/renormalize-merge-kill-safer-crlf:
convert: git cherry-pick -Xrenormalize did not work
merge-recursive: handle NULL in add_cacheinfo() correctly
cherry-pick: demonstrate a segmentation fault
Working with a repo that used to be all CRLF. At some point it
was changed to all LF, with `text=auto` in .gitattributes.
Trying to cherry-pick a commit from before the switchover fails:
$ git cherry-pick -Xrenormalize <commit>
fatal: CRLF would be replaced by LF in [path]
Commit 65237284 "unify the "auto" handling of CRLF" introduced
a regression:
Whenever crlf_action is CRLF_TEXT_XXX and not CRLF_AUTO_XXX,
SAFE_CRLF_RENORMALIZE was feed into check_safe_crlf(). This is
wrong because here everything else than SAFE_CRLF_WARN is treated as
SAFE_CRLF_FAIL.
Call check_safe_crlf() only if checksafe is SAFE_CRLF_WARN or
SAFE_CRLF_FAIL.
Reported-by: Eevee (Lexy Munroe) <eevee@veekun.com>
Signed-off-by: Torsten Bögershausen <tboegi@web.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The smudge/clean filter API expect an external process is spawned
to filter the contents for each path that has a filter defined. A
new type of "process" filter API has been added to allow the first
request to run the filter for a path to spawn a single process, and
all filtering need is served by this single process for multiple
paths, reducing the process creation overhead.
* ls/filter-process:
contrib/long-running-filter: add long running filter example
convert: add filter.<driver>.process option
convert: prepare filter.<driver>.process option
convert: make apply_filter() adhere to standard Git error handling
pkt-line: add functions to read/write flush terminated packet streams
pkt-line: add packet_write_gently()
pkt-line: add packet_flush_gently()
pkt-line: add packet_write_fmt_gently()
pkt-line: extract set_packet_header()
pkt-line: rename packet_write() to packet_write_fmt()
run-command: add clean_on_exit_handler
run-command: move check_pipe() from write_or_die to run_command
convert: modernize tests
convert: quote filter names in error messages
Mark error messages about CRLF for translation.
Update test to reflect changes.
Signed-off-by: Vasco Almeida <vascomalmeida@sapo.pt>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Git's clean/smudge mechanism invokes an external filter process for
every single blob that is affected by a filter. If Git filters a lot of
blobs then the startup time of the external filter processes can become
a significant part of the overall Git execution time.
In a preliminary performance test this developer used a clean/smudge
filter written in golang to filter 12,000 files. This process took 364s
with the existing filter mechanism and 5s with the new mechanism. See
details here: https://github.com/github/git-lfs/pull/1382
This patch adds the `filter.<driver>.process` string option which, if
used, keeps the external filter process running and processes all blobs
with the packet format (pkt-line) based protocol over standard input and
standard output. The full protocol is explained in detail in
`Documentation/gitattributes.txt`.
A few key decisions:
* The long running filter process is referred to as filter protocol
version 2 because the existing single shot filter invocation is
considered version 1.
* Git sends a welcome message and expects a response right after the
external filter process has started. This ensures that Git will not
hang if a version 1 filter is incorrectly used with the
filter.<driver>.process option for version 2 filters. In addition,
Git can detect this kind of error and warn the user.
* The status of a filter operation (e.g. "success" or "error) is set
before the actual response and (if necessary!) re-set after the
response. The advantage of this two step status response is that if
the filter detects an error early, then the filter can communicate
this and Git does not even need to create structures to read the
response.
* All status responses are pkt-line lists terminated with a flush
packet. This allows us to send other status fields with the same
protocol in the future.
Helped-by: Martin-Louis Bright <mlbright@gmail.com>
Reviewed-by: Jakub Narebski <jnareb@gmail.com>
Signed-off-by: Lars Schneider <larsxschneider@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Refactor the existing 'single shot filter mechanism' and prepare the
new 'long running filter mechanism'.
Signed-off-by: Lars Schneider <larsxschneider@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
apply_filter() returns a boolean that tells the caller if it
"did convert or did not convert". The variable `ret` was used throughout
the function to track errors whereas `1` denoted success and `0`
failure. This is unusual for the Git source where `0` denotes success.
Rename the variable and flip its value to make the function easier
readable for Git developers.
Signed-off-by: Lars Schneider <larsxschneider@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Git filter driver commands with spaces (e.g. `filter.sh foo`) are hard
to read in error messages. Quote them to improve the readability.
Signed-off-by: Lars Schneider <larsxschneider@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
When a non-reversible CRLF conversion is done in "git add",
a warning is printed on stderr (or Git dies, depending on checksafe)
The function commit_chk_wrnNNO() in t0027 was written to test this,
but did the wrong thing: Instead of looking at the warning
from "git add", it looked at the warning from "git commit".
This is racy because "git commit" may not have to do CRLF conversion
at all if it can use the sha1 value from the index (which depends on
whether "add" and "commit" run in a single second).
Correct t0027 and replace the commit for each and every file with a commit
of all files in one go.
The function commit_chk_wrnNNO() should be renamed in a separate commit.
Now that t0027 does the right thing, it detects a bug in covert.c:
This sequence should generate the warning `LF will be replaced by CRLF`,
but does not:
$ git init
$ git config core.autocrlf false
$ printf "Line\r\n" >file
$ git add file
$ git commit -m "commit with CRLF"
$ git config core.autocrlf true
$ printf "Line\n" >file
$ git add file
"git add" calls crlf_to_git() in convert.c, which calls check_safe_crlf().
When has_cr_in_index(path) is true, crlf_to_git() returns too early and
check_safe_crlf() is not called at all.
Factor out the code which determines if "git checkout" converts LF->CRLF
into will_convert_lf_to_crlf().
Update the logic around check_safe_crlf() and "simulate" the possible
LF->CRLF conversion at "git checkout" with help of will_convert_lf_to_crlf().
Thanks to Jeff King <peff@peff.net> for analyzing t0027.
Reported-By: Johannes Schindelin <Johannes.Schindelin@gmx.de>
Signed-off-by: Torsten Bögershausen <tboegi@web.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Before this change,
$ echo "* text=auto" >.gitattributes
$ echo "* eol=crlf" >>.gitattributes
would have the same effect as
$ echo "* text" >.gitattributes
$ git config core.eol crlf
Since the 'eol' attribute had higher priority than 'text=auto', this may
corrupt binary files and is not what most users expect to happen.
Make the 'eol' attribute to obey 'text=auto' and now
$ echo "* text=auto" >.gitattributes
$ echo "* eol=crlf" >>.gitattributes
behaves the same as
$ echo "* text=auto" >.gitattributes
$ git config core.eol crlf
In other words,
$ echo "* text=auto eol=crlf" >.gitattributes
has the same effect as
$ git config core.autocrlf true
and
$ echo "* text=auto eol=lf" >.gitattributes
has the same effect as
$ git config core.autocrlf input
Signed-off-by: Torsten Bögershausen <tboegi@web.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
When the ident attributes is set, get_stream_filter() did not obey
core.autocrlf=true, and the file was checked out with LF.
Change the rule when a streaming filter can be used:
- if an external filter is specified, don't use a stream filter.
- if the worktree eol is CRLF and "auto" is active, don't use a stream filter.
- Otherwise the stream filter can be used.
Add test cases in t0027.
Signed-off-by: Torsten Bögershausen <tboegi@web.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
df747b81 (convert.c: refactor crlf_action, 2016-02-10) introduced a
bug to "git ls-files --eol".
The "text" attribute was shown as "text eol=lf" or "text eol=crlf",
depending on core.autocrlf or core.eol.
Correct this and add test cases in t0027.
Signed-off-by: Torsten Bögershausen <tboegi@web.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Simplify the statistics:
lonecr counts the CR which is not followed by a LF,
lonelf counts the LF which is not preceded by a CR,
crlf counts CRLF combinations.
This simplifies the evaluation of the statistics.
Signed-off-by: Torsten Bögershausen <tboegi@web.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Refactor the determination and usage of crlf_action.
Today, when no "crlf" attribute are set on a file, crlf_action is set to
CRLF_GUESS. Use CRLF_UNDEFINED instead, and search for "text" or "eol" as
before.
After searching for line ending attributes, save the value in
struct conv_attrs.crlf_action attr_action,
so that get_convert_attr_ascii() is able report the attributes.
Replace the old CRLF_GUESS usage:
CRLF_GUESS && core.autocrlf=true -> CRLF_AUTO_CRLF
CRLF_GUESS && core.autocrlf=false -> CRLF_BINARY
CRLF_GUESS && core.autocrlf=input -> CRLF_AUTO_INPUT
Save the action in conv_attrs.crlf_action (as before) and change
all callers.
Make more clear, what is what, by defining:
- CRLF_UNDEFINED : No attributes set. Temparally used, until core.autocrlf
and core.eol is evaluated and one of CRLF_BINARY,
CRLF_AUTO_INPUT or CRLF_AUTO_CRLF is selected
- CRLF_BINARY : No processing of line endings.
- CRLF_TEXT : attribute "text" is set, line endings are processed.
- CRLF_TEXT_INPUT: attribute "input" or "eol=lf" is set. This implies text.
- CRLF_TEXT_CRLF : attribute "eol=crlf" is set. This implies text.
- CRLF_AUTO : attribute "auto" is set.
- CRLF_AUTO_INPUT: core.autocrlf=input (no attributes)
- CRLF_AUTO_CRLF : core.autocrlf=true (no attributes)
Signed-off-by: Torsten Bögershausen <tboegi@web.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Clean/smudge filters defined in a configuration file of lower
precedence can now be overridden to be a pass-through no-op by
setting the variable to an empty string.
* ls/clean-smudge-override-in-config:
convert: treat an empty string for clean/smudge filters as "cat"
Add a helper function to find out, which line endings text files
should get at checkout, depending on core.autocrlf and core.eol
configuration variables.
Signed-off-by: Torsten Bögershausen <tboegi@web.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Integrate the code of input_crlf_action() into convert_attrs(),
so that ca.crlf_action is always valid after calling convert_attrs().
Keep a copy of crlf_action in attr_action, this is needed for
get_convert_attr_ascii().
Remove eol_attr from struct conv_attrs, as it is now used temporally.
Signed-off-by: Torsten Bögershausen <tboegi@web.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Some functions get a parameter path, but don't use it.
Remove the unused parameter.
Signed-off-by: Torsten Bögershausen <tboegi@web.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Once a lower-priority configuration file defines a clean or smudge
filter, there is no convenient way to override it to produce as-is
output. Even though the configuration mechanism implements "the
last one wins" semantics, you cannot set them to an empty string and
expect them to work, as apply_filter() would try to run the empty
string as an external command and fail. The conversion is not done,
but the function would still report a failure to convert.
Even though resetting the variable to "cat" (i.e. pass the data back
as-is and report success) is an obvious and a viable way to solve
this, it is wasteful to spawn an external process just as a
workaround.
Instead, teach apply_filter() to treat an empty string as a no-op
filter that always returns successfully its input as-is without
conversion.
Signed-off-by: Lars Schneider <larsxschneider@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
When working in a cross-platform environment, a user may want to
check if text files are stored normalized in the repository and
if .gitattributes are set appropriately.
Make it possible to let Git show the line endings in the index and
in the working tree and the effective text/eol attributes.
The end of line ("eolinfo") are shown like this:
"-text" binary (or with bare CR) file
"none" text file without any EOL
"lf" text file with LF
"crlf" text file with CRLF
"mixed" text file with mixed line endings.
The effective text/eol attribute is one of these:
"", "-text", "text", "text=auto", "text eol=lf", "text eol=crlf"
git ls-files --eol gives an output like this:
i/none w/none attr/text=auto t/t5100/empty
i/-text w/-text attr/-text t/test-binary-2.png
i/lf w/lf attr/text eol=lf t/t5100/rfc2047-info-0007
i/lf w/crlf attr/text eol=crlf doit.bat
i/mixed w/mixed attr/ locale/XX.po
to show what eol convention is used in the data in the index ('i'),
and in the working tree ('w'), and what attribute is in effect,
for each path that is shown.
Add test cases in t0027.
Helped-By: Eric Sunshine <sunshine@sunshineco.com>
Signed-off-by: Torsten Bögershausen <tboegi@web.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
We sometimes sprintf into fixed-size buffers when we know
that the buffer is large enough to fit the input (either
because it's a constant, or because it's numeric input that
is bounded in size). Likewise with strcpy of constant
strings.
However, these sites make it hard to audit sprintf and
strcpy calls for buffer overflows, as a reader has to
cross-reference the size of the array with the input. Let's
use xsnprintf instead, which communicates to a reader that
we don't expect this to overflow (and catches the mistake in
case we do).
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
We are explicitly ignoring SIGPIPE, as we fully expect that the
filter program may not read our output fully. Ignore EPIPE that
may come from writing to it as well.
A new test was stolen from Jeff's suggestion.
Helped-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
When running a required clean filter, we do not have to mmap the
original before feeding the filter. Instead, stream the file
contents directly to the filter and process its output.
* sp/stream-clean-filter:
sha1_file: don't convert off_t to size_t too early to avoid potential die()
convert: stream from fd to required clean filter to reduce used address space
copy_fd(): do not close the input file descriptor
mmap_limit: introduce GIT_MMAP_LIMIT to allow testing expected mmap size
memory_limit: use git_env_ulong() to parse GIT_ALLOC_LIMIT
config.c: add git_env_ulong() to parse environment variable
convert: drop arguments other than 'path' from would_convert_to_git()
The data is streamed to the filter process anyway. Better avoid mapping
the file if possible. This is especially useful if a clean filter
reduces the size, for example if it computes a sha1 for binary data,
like git media. The file size that the previous implementation could
handle was limited by the available address space; large files for
example could not be handled with (32-bit) msysgit. The new
implementation can filter files of any size as long as the filter output
is small enough.
The new code path is only taken if the filter is required. The filter
consumes data directly from the fd. If it fails, the original data is
not immediately available. The condition can easily be handled as
a fatal error, which is expected for a required filter anyway.
If the filter was not required, the condition would need to be handled
in a different way, like seeking to 0 and reading the data. But this
would require more restructuring of the code and is probably not worth
it. The obvious approach of falling back to reading all data would not
help achieving the main purpose of this patch, which is to handle large
files with limited address space. If reading all data is an option, we
can simply take the old code path right away and mmap the entire file.
The environment variable GIT_MMAP_LIMIT, which has been introduced in
a previous commit is used to test that the expected code path is taken.
A related test that exercises required filters is modified to verify
that the data actually has been modified on its way from the file system
to the object store.
Signed-off-by: Steffen Prohaska <prohaska@zib.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Most struct child_process variables are cleared using memset first after
declaration. Provide a macro, CHILD_PROCESS_INIT, that can be used to
initialize them statically instead. That's shorter, doesn't require a
function call and is slightly more readable (especially given that we
already have STRBUF_INIT, ARGV_ARRAY_INIT etc.).
Helped-by: Johannes Sixt <j6t@kdbg.org>
Signed-off-by: Rene Scharfe <l.s.r@web.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Jonathan Tan
Junio C Hamano
Lars Schneider
Stefan Beller
Brandon Williams
Ben Peart
Torsten Bögershausen
Vasco Almeida
Jeff King
Steffen Prohaska
René Scharfe