* js/merge:
merge-recursive: add/add really is modify/modify with an empty base
Get rid of the dependency on RCS' merge program
merge-file: support -p and -q; fix compile warnings
Add builtin merge-file, a minimal replacement for RCS merge
xdl_merge(): fix and simplify conflict handling
xdl_merge(): fix thinko
xdl_merge(): fix an off-by-one bug
merge-recursive: use xdl_merge().
xmerge: make return value from xdl_merge() more usable.
xdiff: add xdl_merge()
"git push $URL HEAD~6" created a bogus ref HEAD~6 immediately
under $GIT_DIR of the remote repository. While we should keep
refspecs that have arbitrary extended SHA-1 expression on the
source side working (e.g. "HEAD~6:refs/tags/yesterday"), we
should not create bogus ref on the other end.
The documentation talked about "git push $URL tag <tag>" as a short-hand
for refs/tags/<tag>:refs/tags/<tag> for a long time but that was never
the case (the short-hand was for "git fetch"). Instead of fixing the
documentation, just add a bit of code to match it since it is easy to do
and would make it more consistent.
This is just a random hack to work around problems people seem
to be seeing in manpage backend of xmlto (it appears we are
getting ".sp" at the end of line without line break).
99.9999% of the time, the command is used with a single
strategy; after a merge failure, saying "No strategy handled the
merge" is technically correct, but there is no point stressing
we tried and failed all the possibilities the user has given.
Prior to 65ac6e9c3f47807cb603af07a6a9e1a43bc119ae we deleted a file
from the working directory during a merge if the file existed before
the merge started but was renamed by the branch being merged in.
This broke in 65ac6e as git-merge-recursive did not actually update
the working directory on an uncontested rename.
Signed-off-by: Shawn O. Pearce <spearce@spearce.org> Signed-off-by: Junio C Hamano <junkio@cox.net>
Some versions of the SVN libraries cause die() to exit with 255,
and 40cf043389ef4cdf3e56e7c4268d6f302e387fa0 tightened up
test_expect_failure to reject return values >128.
Signed-off-by: Eric Wong <normalperson@yhbt.net> Signed-off-by: Junio C Hamano <junkio@cox.net>
Allow subcommand.color and color.subcommand color configuration
While adding colour to the branch command it was pointed out that a
config option like "branch.color" conflicts with the pre-existing
"branch.something" namespace used for specifying default merge urls and
branches. The suggested solution was to flip the order of the
components to "color.branch", which I did for colourising branch.
This patch does the same thing for
- git-log (color.diff)
- git-status (color.status)
- git-diff (color.diff)
- pager (color.pager)
I haven't removed the old config options; but they should probably be
deprecated and eventually removed to prevent future namespace
collisions. I've done this deprecation by changing the documentation
for the config file to match the new names; and adding the "color.XXX"
options to contrib/completion/git-completion.bash.
Unfortunately git-svn reads "diff.color" and "pager.color"; which I
don't like to change unilaterally.
Signed-off-by: Andy Parkins <andyparkins@gmail.com> Signed-off-by: Junio C Hamano <junkio@cox.net>
* master: (42 commits)
git-svn: correctly handle packed-refs in refs/remotes/
add test case for recursive merge
git-svn: correctly display fatal() error messages
git-svn: allow dcommit to take an alternate head
git-svn: enable logging of information not supported by git
Clarify fetch error for missing objects.
Move Fink and Ports check to after config file
shortlog: fix segfault on empty authorname
shortlog: remove "[PATCH]" prefix from shortlog output
Make sure the empty tree exists when needed in merge-recursive.
Don't use memcpy when source and dest. buffers may overlap
no need to install manpages as executable
Documentation: simpler shared repository creation
shortlog: fix segfault on empty authorname
Add branch.*.merge warning and documentation update
Fix perl/ build.
git-svn: use do_switch for --follow-parent if the SVN library supports it
Fix documentation copy&paste typo
git-svn: extra error check to ensure we open a file correctly
Documentation: update git-clone man page with new behavior
...
Previously dcommit would unconditionally commit all patches
up-to and including the current HEAD. Now if an optional
command-line argument is specified, it will only commit
up to the specified revision.
Signed-off-by: Eric Wong <normalperson@yhbt.net> Signed-off-by: Junio C Hamano <junkio@cox.net>
git-svn: enable logging of information not supported by git
The changes are now tracked in
$GIT_DIR/svn/$GIT_SVN_ID/untracked.log
Information in the untracked.log include:
* the addition and removal of empty directories
(changes of these will also warn the user)
* file and directory property changes, including (but not
limited to) svk:merge and svn:externals
* revision properties (revprops) are also tracked
* users will be warned of 'absent' file and directories
(if users are forbidden access)
Fields in entries are separated by spaces; "unsafe" characters
are URI-encoded so that each entry takes exactly one line.
There is currently no automated parser for dealing with the data
in untracked.log, but it should be possible to write one to
create empty directories on checkout and manage
externals/subprojects.
Signed-off-by: Eric Wong <normalperson@yhbt.net> Signed-off-by: Junio C Hamano <junkio@cox.net>
The old code looked backwards from the email address to parse
the name, allowing an arbitrary number of spaces between the
two. However, in the case of no name, we looked back too far to
the 'author' (or 'Author:') header.
The bug was triggered by commit febf7ea4bed from linux-2.6.
Jeff King originally fixed it by looking back only one
character; Johannes Schindelin pointed out that we could try
harder while at it to cope with commits with broken headers.
shortlog: remove "[PATCH]" prefix from shortlog output
Originally noticed by Nicolas Pitre; the real cause was the code
was prepared to deal with [PATCH] (and [PATCH n/m whatever])
prefixes but forgot that the string can be indented while acting
as a filter.
* maint:
Make sure the empty tree exists when needed in merge-recursive.
Don't use memcpy when source and dest. buffers may overlap
no need to install manpages as executable
Make sure the empty tree exists when needed in merge-recursive.
There are some baseless merge cases where git-merge-recursive will
try to compare one of the branches against the empty tree. However
most projects won't have the empty tree object in their object database
as Git does not normally create empty tree objects. If the empty tree
object is missing then the merge process will die, as it cannot load the
object from the database. The error message may make the user think that
their database is corrupt when its actually not.
So instead we should just create the empty tree object whenever it is
needed. If the object already exists as a loose object then no harm
done. Otherwise that loose object will be pruned away later by either
git-prune or git-prune-packed.
Thanks goes to Junio for suggesting this fix.
Signed-off-by: Shawn O. Pearce <spearce@spearce.org> Signed-off-by: Junio C Hamano <junkio@cox.net>
Take Johannes Schindelin's suggestions for a further simplification of
the shared repository creation using git --bare init-db --shared, and
for a simplified cvsimport using an existing CVS working directory.
Also insert more man page references.
Signed-off-by: J. Bruce Fields <bfields@citi.umich.edu>
cvs-migration.txt | 27 ++++++++++++++-------------
1 file changed, 14 insertions(+), 13 deletions(-) Signed-off-by: Junio C Hamano <junkio@cox.net>
The old code looked backwards from the email address to parse the name,
allowing an arbitrary number of spaces between the two. However, in the case
of no name, we looked back too far to the 'author' (or 'Author:') header.
Instead, remove at most one space between name and address.
The bug was triggered by commit febf7ea4bed from linux-2.6.
Signed-off-by: Jeff King <peff@peff.net> Signed-off-by: Junio C Hamano <junkio@cox.net>
Add branch.*.merge warning and documentation update
This patch clarifies the meaning of the branch.*.merge option.
Previously, if branch.*.merge was specified but did not match any
ref, the message "No changes." was not really helpful regarding
the misconfiguration. This patch adds a warning for this.
Signed-off-by: Josef Weidendorfer <Josef.Weidendorfer@gmx.de> Signed-off-by: Junio C Hamano <junkio@cox.net>
An earlier commit f848718a broke the build in perl/ directory by
allowing the Makefile.PL to overwrite the now-tracked Makefile.
Fix this by forcing Makefile.PL to produce its output in
perl.mak as the broken commit originally intended.
git-svn: use do_switch for --follow-parent if the SVN library supports it
do_switch works with the SVN Perl bindings after r22312 in the
Subversion trunk. Since no released version of SVN currently
supports it; we'll just autodetect it and enable its usage
when a user has a recent-enough version of SVN.
Signed-off-by: Eric Wong <normalperson@yhbt.net> Signed-off-by: Junio C Hamano <junkio@cox.net>
Documentation: update git-clone man page with new behavior
Update git-clone man page to reflect recent changes
(--use-separate-remote default and use of .git/config instead of
remotes files), and rewrite introduction.
Signed-off-by: J. Bruce Fields <bfields@citi.umich.edu> Signed-off-by: Junio C Hamano <junkio@cox.net>
At some point between v5.6 and 5.8 Perl started to assume its input,
output and filehandles are UTF-8. This breaks the counting of bytes
for the CVS protocol, resulting in the client expecting less data
than we actually send, and storing truncated files.
Signed-off-by: Martin Langhoff <martin@catalyst.net.nz> Signed-off-by: Junio C Hamano <junkio@cox.net>
Currently the treshold is 5000. The likelihood of this value to ever be
crossed for a single push is really small making it not really useful.
The optimal treshold for a pure space saving on a filesystem with 4kb
blocks is 3. However this is likely to create many small packs
concentrating a large number of files in a single directory compared to
the same objects which are spread over 256 directories when loose. This
means we would need 512 objects per pack on average to approximagte the
same directory cost (a pack has 2 files because of the index).
But 512 is a really high value just like 5000 since most pushes are
unlikely to have that many objects. So let's try with a value of 100
which should have a good balance between small pushes going to be
exploded into loose objects and large pushes kept as whole packs.
This is not a replacement for periodic repacks of course.
Signed-off-by: Nicolas Pitre <nico@cam.org> Signed-off-by: Junio C Hamano <junkio@cox.net>
Modify cvs-migration.txt so it explains first how to develop against a
shared repository, then how to set up a shared repository, then how to
import a repository from cvs. Though this seems chronologically
backwards, it's still readable in this order, and it puts the more
commonly needed material closer to the front.
Remove the annotate/pickaxe section; perhaps it can find a place elsewhere
in the future. Remove most of the "why git is better than cvs" stuff from
the introduction.
Add some minor clarifications, including two that have come up several
times on the mailing list:
1. Recommend committing any changes before running pull.
2. Note that changes must be commited before they can be pushed.
Update the clone discussion to reflect the new --use-separate-remotes
default, and add a brief mention of git-cvsserver.
Signed-off-by: J. Bruce Fields <bfields@citi.umich.edu> Signed-off-by: Junio C Hamano <junkio@cox.net>
git-merge: fix "fix confusion between tag and branch" for real
An earlier commit 3683dc5a broke the merge message generation with
a careless use of && where it was not needed, breaking the merge
message for cases where non branches are given.
git-svn: avoid network timeouts for long-running fetches
Long-running fetches run inside children to avoid memory leaks.
When we refork, the connection in the parent can be idle for a
long time; attempting to reuse it in the next child can result
in timeouts.
Signed-off-by: Eric Wong <normalperson@yhbt.net> Signed-off-by: Junio C Hamano <junkio@cox.net>
gitweb: Allow PNG, GIF, JPEG images to be displayed in "blob" view
Allow images in one of web formats (PNG, GIF, JPEG) - actually files
with mimetype of image/png, image/git, image/jpeg - to be displayed in
"blob" view using <img /> element, instead of using "blob_plain" view
for them, like for all other files except also text/* mimetype files.
This makes possible to easily go to file history, to HEAD version of
the file, to appropriate commit etc; all of those are not available
in "blob_plain" (raw) view.
Only text files can have "blame" view link in the formats part of
navbar.
Signed-off-by: Jakub Narebski <jnareb@gmail.com> Signed-off-by: Junio C Hamano <junkio@cox.net>
While deciding if the new style command line argument is a tag
or a branch, we checked it with "git show-ref -s --verify" to
see if results in an error, but when it is not a branch, the
check leaked the error message out, which was not needed to be
shown to the end user.
An earlier commit a9cb3c6e changed git-commit to use the
contents of MERGE_MSG even when we do not have MERGE_HEAD (the
rationale is in its log message).
However, the change tricks the following sequence to include a
merge message in a completely unrelated commit:
$ git pull somewhere
: oops, the conflicts are too much. forget it.
$ git reset --hard
: work work work
$ git commit
To fix this confusion, this patch makes "git reset" to remove
the leftover MERGE_MSG that was prepared when the user abandoned
the merge.
Signed-off-by: Junio C Hamano <junkio@cox.net> Acked-by: Luben Tuikov <ltuikov@yahoo.com
Rename the section titles to make the "how-to" content of the section
obvious. Also clarify that changes have to be commited before they can
be pushed.
Suppose you have changes in new1 to the original lines 10-20,
and changes in new2 to the original lines 15-25, then the
changes to 10-25 conflict. But it is possible that the next
changes in new1 still overlap with this change to new2.
So, in the next iteration we have to look at the same change
to new2 again.
The old code tried to be a bit too clever. The new code is
shorter and more to the point: do not fiddle with the ranges
at all.
Also, xdl_append_merge() tries harder to combine conflicts.
This is necessary, because with the above simplification,
some conflicts would not be recognized as conflicts otherwise:
In the above scenario, it is possible that there is no other
change to new1. Absent the combine logic, the change in new2
would be recorded _again_, but as a non-conflict.
Signed-off-by: Johannes Schindelin <Johannes.Schindelin@gmx.de>
Document git diff options -b / --ignore-space-change and
-w / --ignore-all-space, introduced by Johannes Schindelin
in commit 0d21efa5, "Teach diff about -b and -w flags".
The description of options is taken from GNU diff man page and
GNU Diffutils info documentation.
Signed-off-by: Jakub Narebski <jnareb@gmail.com> Signed-off-by: Junio C Hamano <junkio@cox.net>
unpack-trees: make sure "df_conflict_entry.name" is NUL terminated.
The structure that ends with a flexible array member (or 0
length array with older GCC) "char name[FLEX_ARRAY]" is
allocated on the stack and we use it after clearing its entire
size with memset. That does not guarantee that "name" is
properly NUL terminated as we intended on platforms with more
forgiving structure alignment requirements.
git-fetch: ignore dereferenced tags in expand_refs_wildcard
There was a little bug in the brace expansion which should remove
the ^{} from the tagname. It used ${name#'^{}'} instead of $(name%'^{}'},
the difference is that '#' will remove the given pattern only from the
beginning of a string and '%' only from the end of a string.
Signed-off-by: Michael Loeffler <zvpunry@zvpunry.de> Signed-off-by: Junio C Hamano <junkio@cox.net>
git-clone: Rename --use-immingled-remote option to --no-separate-remote
With making --use-separate-remote default when creating non-bare
clone, there was need for the flag which would turn off this behavior.
It was called --use-immingled-remote.
Immingle means to blend, to combine into one, to intermingle, but it
is a bit obscure word. I think it would be better to use simply
--no-separate-remote as the opposite to --use-separate-remote
option to git clone.
Signed-off-by: Jakub Narebski <jnareb@gmail.com> Signed-off-by: Junio C Hamano <junkio@cox.net>
Linus recommended this, otherwise any renames cause the
diffstat output to be ridiculous in some circumstances.
Because the corresponding "git-pull" done when the requestee
actually makes pull shows the stat with rename detection
enabled, it makes sense to match what the request message
includes to that output, to make the result easier to verify.
Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Junio C Hamano <junkio@cox.net>
git-mv: search more precisely for source directory in index
A move of a directory should find the entries in the index by
searching for the name _including_ the slash. Otherwise, the
directory can be shadowed by a file when it matches the prefix
and is lexicographically smaller, e.g. "ab.c" shadows "ab/".
Noticed by Sergey Vlasov.
[jc: added Sergey's original reproduction recipe as a test case
at the end of t7001.]
Signed-off-by: Johannes Schindelin <Johannes.Schindelin@gmx.de> Signed-off-by: Junio C Hamano <junkio@cox.net>
If one side's block (of changed lines) ends later than the other
side's block, the former should be tested against the next block
of the other side, not vice versa.
Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de> Signed-off-by: Junio C Hamano <junkio@cox.net>
This new function implements the functionality of RCS merge, but
in-memory. It returns < 0 on error, otherwise the number of conflicts.
Finding the conflicting lines can be a very expensive task. You can
control the eagerness of this algorithm:
- a level value of 0 means that all overlapping changes are treated
as conflicts,
- a value of 1 means that if the overlapping changes are identical,
it is not treated as a conflict.
- If you set level to 2, overlapping changes will be analyzed, so that
almost identical changes will not result in huge conflicts. Rather,
only the conflicting lines will be shown inside conflict markers.
With each increasing level, the algorithm gets slower, but more accurate.
Note that the code for level 2 depends on the simple definition of
mmfile_t specific to git, and therefore it will be harder to port that
to LibXDiff.
Signed-off-by: Johannes Schindelin <Johannes.Schindelin@gmx.de> Signed-off-by: Junio C Hamano <junkio@cox.net>
git-svn: avoid fetching files twice in the same revision
SVN is not entirely consistent in returning log information and
sometimes returns file information when adding subdirectories,
and sometimes it does not (only returning information about the
directory that was added). This caused git-svn to occasionally
add a file to the list of files to be fetched twice. Now we
change the data structure to be hash to avoid repeated fetches.
As of now (in master), this only affects repositories fetched
without deltas enabled (file://, and when manually overriden
with GIT_SVN_DELTA_FETCH=0); so this bug mainly affects users of
1.4.4.1 and maint.
Thanks to Florian Weimer for reporting this bug.
[jc: backported for maint]
Signed-off-by: Eric Wong <normalperson@yhbt.net> Signed-off-by: Junio C Hamano <junkio@cox.net>
The fact that git has previously used symbolic links for representing
symbolic refs doesn't seem relevant to the current function of
git-symbolic-ref. This patch makes less of a big deal about the
symbolic link history and instead focuses on what git does now.
Signed-off-by: Andy Parkins <andyparkins@gmail.com> Signed-off-by: Junio C Hamano <junkio@cox.net>
git-diff: Introduce --index and deprecate --cached.
'git diff --cached' still works, but its use is discouraged
in the documentation. 'git diff --index' does the same thing
and is consistent with how 'git apply --index' works.
Signed-off-by: Andreas Ericsson <ae@op5.se> Signed-off-by: Junio C Hamano <junkio@cox.net>
Without this patch "git commit file.c file2.c" produces the not
so stellar output:
error: pathspec 'file.c' did not match any.
error: pathspec 'file2.c' did not match any.
With this patch, the output is changed to:
error: pathspec 'file.c' did not match any file(s) known to git.
error: pathspec 'file2.c' did not match any file(s) known to git.
Did you forget to 'git add'?
Signed-off-by: Andreas Ericsson <ae@op5.se> Signed-off-by: Junio C Hamano <junkio@cox.net>
* branch 'maint':
Document git-repo-config --bool/--int options.
tutorial: talk about user.name early and don't start with commit -a
git-blame: fix rev parameter handling.
tutorial: talk about user.name early and don't start with commit -a
Introducing yourself to git early would be a good idea; otherwise
the user may not find the mistake until much later when "git log"
is learned.
Teaching "commit -a" without saying that it is a shortcut for
listing the paths to commit leaves the user puzzled. Teach the
form with explicit paths first.
We lacked "--" termination in the underlying init_revisions() call
which made it impossible to specify a revision that happens to
have the same name as an existing file.
* jc/globfetch:
fetch-pack: do not barf when duplicate re patterns are given
git-fetch: allow forcing glob pattern in refspec
git-fetch: allow glob pattern in refspec
git-fetch: fix dumb protocol transport to fetch from pack-pruned ref
git-fetch: reuse ls-remote result.
git blame -C: fix output format tweaks when crossing file boundary.
We used to get the case that more than two paths came from the
same commit wrong when computing the output width and deciding
to turn on --show-name option automatically. When we find that
lines that came from a path that is different from what we
started digging from, we should always turn --show-name on, and
we should count the name length for all files involved.
After the bugfix to connect to repositories where the user has
limited read permissions, multi-init was broken due to our
SVN::Ra connection being limited to working in a subdirectory;
so we now create a new Ra connection for init-ing branches
and another for tags
Along with that fix, allow the user to use the command-line
option flags for multi-init (--revision being the most notable;
but also --no-auth-cache, --config-dir, --username (for passing
to SVN), and --shared/--template for passing to git-init-db
Signed-off-by: Eric Wong <normalperson@yhbt.net> Signed-off-by: Junio C Hamano <junkio@cox.net>
* match LESS environment settings to those in pager.c
* parse diff.color and pager.color settings in the
config file, and pass --color to git-log
* --color and --pager= settings are supported
Signed-off-by: Eric Wong <normalperson@yhbt.net> Signed-off-by: Junio C Hamano <junkio@cox.net>
Commit 35e65ecc broke completion of local refs, e.g. "git pull . fo<tab>"
no longer would complete to "foo". Instead it printed out an internal
git error ("fatal: Not a git repository: '.'").
The break occurred when I tried to improve performance by switching from
git-peek-remote to git-for-each-ref. Apparently git-peek-remote will
drop into directory "$1/.git" (where $1 is its first parameter) if it
is given a repository with a working directory. This allowed the bash
completion code to work properly even though it was not handing over
the true repository directory.
So now we do a stat in bash to see if we need to add "/.git" to the
path string before running any command with --git-dir.
I also tried to optimize away two "git rev-parse --git-dir" invocations
in common cases like "git log fo<tab>" as typically the user is in the
top level directory of their project and therefore the .git subdirectory
is in the current working directory. This should make a difference on
systems where fork+exec might take a little while.
Signed-off-by: Shawn O. Pearce <spearce@spearce.org> Signed-off-by: Junio C Hamano <junkio@cox.net>
git-svn: fix output reporting from the delta fetcher
There was nothing printed in the code originally because I left
out a pair of parentheses. Nevertheless, the affected code has
been replaced with a more efficient version that respects the -q
flag as well as requiring less bandwidth.
We save some bandwidth by not requesting changed paths
information when calling get_log() since we're using the delta
fetcher.
Signed-off-by: Eric Wong <normalperson@yhbt.net> Signed-off-by: Junio C Hamano <junkio@cox.net>
git-svn: error out when the SVN connection fails during a fetch
finish_report does seem to return a useful value indicating success
or failure, so we'll just set a flag when close_edit is called
(it is not called on failures, nor is abort_edit) and check
the flag before proceeding.
Thanks to Pazu for pointing this out.
Signed-off-by: Eric Wong <normalperson@yhbt.net> Signed-off-by: Junio C Hamano <junkio@cox.net>
Don't force the user to specify more than one revision parameter,
thus making git-shortlog behave more like git-log.
'git-shortlog master' will now produce the expected results; the
other end of the range simply is the (oldest) root commit.
Signed-off-by: Rene Scharfe <rene.scharfe@lsrfire.ath.cx> Signed-off-by: Junio C Hamano <junkio@cox.net>
git-svn: enable delta transfers during fetches when using SVN:: libs
This should drastically reduce bandwidth used for network
transfers. This is not enabled for file:// repositories by
default because of the increased CPU usage and I/O needed.
GIT_SVN_DELTA_FETCH may be set to a true value to enable or
false (0) to disable delta transfers regardless of the
repository type.
Signed-off-by: Eric Wong <normalperson@yhbt.net> Signed-off-by: Junio C Hamano <junkio@cox.net>
Use .git/config for storing "origin" shortcut repository
Rather than use a separate config .git/remotes/ for remote shortcuts, this
patch adds the analagous definitions to .git/config using git-repo-config
calls.
For example what was previously .git/remotes/origin
URL: proto://host/path
Pull: refs/heads/master:refs/heads/origin
Is now added to .git/config as
[remote "origin"]
url = proto://host/path
fetch = refs/heads/master:refs/heads/origin
Signed-off-by: Andy Parkins <andyparkins@gmail.com> Signed-off-by: Junio C Hamano <junkio@cox.net>
sha1_object_info(): be consistent with read_sha1_file()
We used to try loose objects first with sha1_object_info(), but packed
objects first with read_sha1_file(). Now, prefer packed objects over loose
ones with sha1_object_info(), too.
Usually the old behaviour would pose no problem, but when you tried to fix
a fscked up repository by inserting a known-good pack,
git cat-file $(git cat-file -t <sha1>) <sha1>
could fail, even when
git cat-file blob <sha1>
would _not_ fail. Worse, a repack would fail, too.
Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de> Signed-off-by: Junio C Hamano <junkio@cox.net>
This should not change any functionality, but just makes it readable by
having a space between syntactic construct keyword and open parenthesis
(e.g. "if (expr", not "if(expr") and between close parenthesis and open
brace (e.g. "if (expr) {" not "if (expr){").