git-commit-vandalism/t/t9502-gitweb-standalone-parse-output.sh

208 lines
5.9 KiB
Bash
Raw Normal View History

gitweb: Document current snapshot rules via new tests Add t9502-gitweb-standalone-parse-output test script, which runs gitweb as a CGI script from the commandline and checks that it produces the correct output. Currently this test script contains only tests of snapshot naming (proposed name of snapshot file) and snapshot prefix (prefix of files in the archive / snapshot). It defines and uses 'tar' snapshot format, without compression, for easy checking of snapshot prefix. Testing is done using check_snapshot function. Gitweb uses the following format for snapshot filenames: <sanitized project name>-<hash parameter><snapshot suffix> where <sanitized project name> is project name with '.git' or '/.git' suffix stripped, unless '.git' is the whole project name. For snapshot prefix it uses simply: <sanitized project name>/ Disadvantages of current snapshot rules: * There exists convention that <basename>.<suffix> archive unpacks to <basename>/ directory (<basename>/ is prefix of archive). Gitweb does not respect it * Snapshot links generated by gitweb use full SHA-1 id as a value of 'h' / $hash parameter. With current rules it leads to long file names like e.g. repo-1005c80cc11c531d327b12195027cbbb4ff9e3cb.tgz * For handcrafted URLs, where 'h' / $hash parameter is a symbolic 'volatile' revision name such as "HEAD" or "next" snapshot name doesn't tell us what exact version it was created from * Proposed filename in Content-Disposition header should not contain any directory path information, which means that it should not contain '/' (see RFC2183)... which means that snapshot naming is broken for $hash being e.g. hirearchical branch name such as 'xx/test' This would be improved in next commit. Signed-off-by: Jakub Narebski <jnareb@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2009-11-07 16:13:28 +01:00
#!/bin/sh
#
# Copyright (c) 2009 Mark Rada
#
test_description='gitweb as standalone script (parsing script output).
This test runs gitweb (git web interface) as a CGI script from the
commandline, and checks that it produces the correct output, either
in the HTTP header or the actual script output.'
. ./gitweb-lib.sh
# ----------------------------------------------------------------------
# snapshot file name and prefix
cat >>gitweb_config.perl <<\EOF
$known_snapshot_formats{'tar'} = {
'display' => 'tar',
'type' => 'application/x-tar',
'suffix' => '.tar',
'format' => 'tar',
};
$feature{'snapshot'}{'default'} = ['tar'];
EOF
# Call check_snapshot with the arguments "<basename> [<prefix>]"
#
# This will check that gitweb HTTP header contains proposed filename
# as <basename> with '.tar' suffix added, and that generated tarfile
# (gitweb message body) has <prefix> as prefix for al files in tarfile
#
# <prefix> default to <basename>
check_snapshot () {
basename=$1
prefix=${2:-"$1"}
echo "basename=$basename"
grep "filename=.*$basename.tar" gitweb.headers >/dev/null 2>&1 &&
"$TAR" tf gitweb.body >file_list &&
! grep -v -e "^$prefix$" -e "^$prefix/" -e "^pax_global_header$" file_list
gitweb: Document current snapshot rules via new tests Add t9502-gitweb-standalone-parse-output test script, which runs gitweb as a CGI script from the commandline and checks that it produces the correct output. Currently this test script contains only tests of snapshot naming (proposed name of snapshot file) and snapshot prefix (prefix of files in the archive / snapshot). It defines and uses 'tar' snapshot format, without compression, for easy checking of snapshot prefix. Testing is done using check_snapshot function. Gitweb uses the following format for snapshot filenames: <sanitized project name>-<hash parameter><snapshot suffix> where <sanitized project name> is project name with '.git' or '/.git' suffix stripped, unless '.git' is the whole project name. For snapshot prefix it uses simply: <sanitized project name>/ Disadvantages of current snapshot rules: * There exists convention that <basename>.<suffix> archive unpacks to <basename>/ directory (<basename>/ is prefix of archive). Gitweb does not respect it * Snapshot links generated by gitweb use full SHA-1 id as a value of 'h' / $hash parameter. With current rules it leads to long file names like e.g. repo-1005c80cc11c531d327b12195027cbbb4ff9e3cb.tgz * For handcrafted URLs, where 'h' / $hash parameter is a symbolic 'volatile' revision name such as "HEAD" or "next" snapshot name doesn't tell us what exact version it was created from * Proposed filename in Content-Disposition header should not contain any directory path information, which means that it should not contain '/' (see RFC2183)... which means that snapshot naming is broken for $hash being e.g. hirearchical branch name such as 'xx/test' This would be improved in next commit. Signed-off-by: Jakub Narebski <jnareb@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2009-11-07 16:13:28 +01:00
}
test_expect_success setup '
test_commit first foo &&
git branch xx/test &&
FULL_ID=$(git rev-parse --verify HEAD) &&
SHORT_ID=$(git rev-parse --verify --short=7 HEAD)
'
test_debug '
echo "FULL_ID = $FULL_ID"
echo "SHORT_ID = $SHORT_ID"
'
test_expect_success 'snapshot: full sha1' '
gitweb_run "p=.git;a=snapshot;h=$FULL_ID;sf=tar" &&
2009-11-07 16:13:29 +01:00
check_snapshot ".git-$SHORT_ID"
gitweb: Document current snapshot rules via new tests Add t9502-gitweb-standalone-parse-output test script, which runs gitweb as a CGI script from the commandline and checks that it produces the correct output. Currently this test script contains only tests of snapshot naming (proposed name of snapshot file) and snapshot prefix (prefix of files in the archive / snapshot). It defines and uses 'tar' snapshot format, without compression, for easy checking of snapshot prefix. Testing is done using check_snapshot function. Gitweb uses the following format for snapshot filenames: <sanitized project name>-<hash parameter><snapshot suffix> where <sanitized project name> is project name with '.git' or '/.git' suffix stripped, unless '.git' is the whole project name. For snapshot prefix it uses simply: <sanitized project name>/ Disadvantages of current snapshot rules: * There exists convention that <basename>.<suffix> archive unpacks to <basename>/ directory (<basename>/ is prefix of archive). Gitweb does not respect it * Snapshot links generated by gitweb use full SHA-1 id as a value of 'h' / $hash parameter. With current rules it leads to long file names like e.g. repo-1005c80cc11c531d327b12195027cbbb4ff9e3cb.tgz * For handcrafted URLs, where 'h' / $hash parameter is a symbolic 'volatile' revision name such as "HEAD" or "next" snapshot name doesn't tell us what exact version it was created from * Proposed filename in Content-Disposition header should not contain any directory path information, which means that it should not contain '/' (see RFC2183)... which means that snapshot naming is broken for $hash being e.g. hirearchical branch name such as 'xx/test' This would be improved in next commit. Signed-off-by: Jakub Narebski <jnareb@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2009-11-07 16:13:28 +01:00
'
test_debug 'cat gitweb.headers && cat file_list'
test_expect_success 'snapshot: shortened sha1' '
gitweb_run "p=.git;a=snapshot;h=$SHORT_ID;sf=tar" &&
2009-11-07 16:13:29 +01:00
check_snapshot ".git-$SHORT_ID"
'
test_debug 'cat gitweb.headers && cat file_list'
test_expect_success 'snapshot: almost full sha1' '
ID=$(git rev-parse --short=30 HEAD) &&
gitweb_run "p=.git;a=snapshot;h=$ID;sf=tar" &&
check_snapshot ".git-$SHORT_ID"
gitweb: Document current snapshot rules via new tests Add t9502-gitweb-standalone-parse-output test script, which runs gitweb as a CGI script from the commandline and checks that it produces the correct output. Currently this test script contains only tests of snapshot naming (proposed name of snapshot file) and snapshot prefix (prefix of files in the archive / snapshot). It defines and uses 'tar' snapshot format, without compression, for easy checking of snapshot prefix. Testing is done using check_snapshot function. Gitweb uses the following format for snapshot filenames: <sanitized project name>-<hash parameter><snapshot suffix> where <sanitized project name> is project name with '.git' or '/.git' suffix stripped, unless '.git' is the whole project name. For snapshot prefix it uses simply: <sanitized project name>/ Disadvantages of current snapshot rules: * There exists convention that <basename>.<suffix> archive unpacks to <basename>/ directory (<basename>/ is prefix of archive). Gitweb does not respect it * Snapshot links generated by gitweb use full SHA-1 id as a value of 'h' / $hash parameter. With current rules it leads to long file names like e.g. repo-1005c80cc11c531d327b12195027cbbb4ff9e3cb.tgz * For handcrafted URLs, where 'h' / $hash parameter is a symbolic 'volatile' revision name such as "HEAD" or "next" snapshot name doesn't tell us what exact version it was created from * Proposed filename in Content-Disposition header should not contain any directory path information, which means that it should not contain '/' (see RFC2183)... which means that snapshot naming is broken for $hash being e.g. hirearchical branch name such as 'xx/test' This would be improved in next commit. Signed-off-by: Jakub Narebski <jnareb@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2009-11-07 16:13:28 +01:00
'
test_debug 'cat gitweb.headers && cat file_list'
test_expect_success 'snapshot: HEAD' '
gitweb_run "p=.git;a=snapshot;h=HEAD;sf=tar" &&
2009-11-07 16:13:29 +01:00
check_snapshot ".git-HEAD-$SHORT_ID"
gitweb: Document current snapshot rules via new tests Add t9502-gitweb-standalone-parse-output test script, which runs gitweb as a CGI script from the commandline and checks that it produces the correct output. Currently this test script contains only tests of snapshot naming (proposed name of snapshot file) and snapshot prefix (prefix of files in the archive / snapshot). It defines and uses 'tar' snapshot format, without compression, for easy checking of snapshot prefix. Testing is done using check_snapshot function. Gitweb uses the following format for snapshot filenames: <sanitized project name>-<hash parameter><snapshot suffix> where <sanitized project name> is project name with '.git' or '/.git' suffix stripped, unless '.git' is the whole project name. For snapshot prefix it uses simply: <sanitized project name>/ Disadvantages of current snapshot rules: * There exists convention that <basename>.<suffix> archive unpacks to <basename>/ directory (<basename>/ is prefix of archive). Gitweb does not respect it * Snapshot links generated by gitweb use full SHA-1 id as a value of 'h' / $hash parameter. With current rules it leads to long file names like e.g. repo-1005c80cc11c531d327b12195027cbbb4ff9e3cb.tgz * For handcrafted URLs, where 'h' / $hash parameter is a symbolic 'volatile' revision name such as "HEAD" or "next" snapshot name doesn't tell us what exact version it was created from * Proposed filename in Content-Disposition header should not contain any directory path information, which means that it should not contain '/' (see RFC2183)... which means that snapshot naming is broken for $hash being e.g. hirearchical branch name such as 'xx/test' This would be improved in next commit. Signed-off-by: Jakub Narebski <jnareb@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2009-11-07 16:13:28 +01:00
'
test_debug 'cat gitweb.headers && cat file_list'
test_expect_success 'snapshot: short branch name (master)' '
gitweb_run "p=.git;a=snapshot;h=master;sf=tar" &&
2009-11-07 16:13:29 +01:00
ID=$(git rev-parse --verify --short=7 master) &&
check_snapshot ".git-master-$ID"
'
test_debug 'cat gitweb.headers && cat file_list'
test_expect_success 'snapshot: short tag name (first)' '
gitweb_run "p=.git;a=snapshot;h=first;sf=tar" &&
ID=$(git rev-parse --verify --short=7 first) &&
check_snapshot ".git-first-$ID"
'
test_debug 'cat gitweb.headers && cat file_list'
test_expect_success 'snapshot: full branch name (refs/heads/master)' '
gitweb_run "p=.git;a=snapshot;h=refs/heads/master;sf=tar" &&
ID=$(git rev-parse --verify --short=7 master) &&
check_snapshot ".git-master-$ID"
'
test_debug 'cat gitweb.headers && cat file_list'
test_expect_success 'snapshot: full tag name (refs/tags/first)' '
gitweb_run "p=.git;a=snapshot;h=refs/tags/first;sf=tar" &&
check_snapshot ".git-first"
gitweb: Document current snapshot rules via new tests Add t9502-gitweb-standalone-parse-output test script, which runs gitweb as a CGI script from the commandline and checks that it produces the correct output. Currently this test script contains only tests of snapshot naming (proposed name of snapshot file) and snapshot prefix (prefix of files in the archive / snapshot). It defines and uses 'tar' snapshot format, without compression, for easy checking of snapshot prefix. Testing is done using check_snapshot function. Gitweb uses the following format for snapshot filenames: <sanitized project name>-<hash parameter><snapshot suffix> where <sanitized project name> is project name with '.git' or '/.git' suffix stripped, unless '.git' is the whole project name. For snapshot prefix it uses simply: <sanitized project name>/ Disadvantages of current snapshot rules: * There exists convention that <basename>.<suffix> archive unpacks to <basename>/ directory (<basename>/ is prefix of archive). Gitweb does not respect it * Snapshot links generated by gitweb use full SHA-1 id as a value of 'h' / $hash parameter. With current rules it leads to long file names like e.g. repo-1005c80cc11c531d327b12195027cbbb4ff9e3cb.tgz * For handcrafted URLs, where 'h' / $hash parameter is a symbolic 'volatile' revision name such as "HEAD" or "next" snapshot name doesn't tell us what exact version it was created from * Proposed filename in Content-Disposition header should not contain any directory path information, which means that it should not contain '/' (see RFC2183)... which means that snapshot naming is broken for $hash being e.g. hirearchical branch name such as 'xx/test' This would be improved in next commit. Signed-off-by: Jakub Narebski <jnareb@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2009-11-07 16:13:28 +01:00
'
test_debug 'cat gitweb.headers && cat file_list'
2009-11-07 16:13:29 +01:00
test_expect_success 'snapshot: hierarchical branch name (xx/test)' '
gitweb: Document current snapshot rules via new tests Add t9502-gitweb-standalone-parse-output test script, which runs gitweb as a CGI script from the commandline and checks that it produces the correct output. Currently this test script contains only tests of snapshot naming (proposed name of snapshot file) and snapshot prefix (prefix of files in the archive / snapshot). It defines and uses 'tar' snapshot format, without compression, for easy checking of snapshot prefix. Testing is done using check_snapshot function. Gitweb uses the following format for snapshot filenames: <sanitized project name>-<hash parameter><snapshot suffix> where <sanitized project name> is project name with '.git' or '/.git' suffix stripped, unless '.git' is the whole project name. For snapshot prefix it uses simply: <sanitized project name>/ Disadvantages of current snapshot rules: * There exists convention that <basename>.<suffix> archive unpacks to <basename>/ directory (<basename>/ is prefix of archive). Gitweb does not respect it * Snapshot links generated by gitweb use full SHA-1 id as a value of 'h' / $hash parameter. With current rules it leads to long file names like e.g. repo-1005c80cc11c531d327b12195027cbbb4ff9e3cb.tgz * For handcrafted URLs, where 'h' / $hash parameter is a symbolic 'volatile' revision name such as "HEAD" or "next" snapshot name doesn't tell us what exact version it was created from * Proposed filename in Content-Disposition header should not contain any directory path information, which means that it should not contain '/' (see RFC2183)... which means that snapshot naming is broken for $hash being e.g. hirearchical branch name such as 'xx/test' This would be improved in next commit. Signed-off-by: Jakub Narebski <jnareb@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2009-11-07 16:13:28 +01:00
gitweb_run "p=.git;a=snapshot;h=xx/test;sf=tar" &&
! grep "filename=.*/" gitweb.headers
'
test_debug 'cat gitweb.headers'
gitweb: Restructure projects list generation Extract filtering out forks (which is done if 'forks' feature is enabled) into filter_forks_from_projects_list subroutine, and searching projects (via projects search form, or via content tags) into search_projects_list subroutine. Both are now run _before_ displaying projects, and not while printing; this allow to know upfront if there were any found projects. Gitweb now can and do print 'No such projects found' if user searches for phrase which does not correspond to any project (any repository). This also would allow splitting projects list into pages, if we so desire. Filtering out forks and marking repository (project) as having forks is now consolidated into one subroutine (special case of handling forks in git_get_projects_list only for $projects_list being file is now removed). Forks handling is also cleaned up and simplified. $pr->{'forks'} now contains un-filled list of forks; we can now also detect situation where the way for having forks is prepared, but there are no forks yet. Sorting projects got also refactored in a very straight way (just moving code) into sort_projects_list subroutine. The interaction between forks, content tags and searching is now made more explicit: searching whether by tag, or via search form turns off fork filtering (gitweb searches also forks, and will show all results). If 'ctags' feature is disabled, then searching by tag is too. The t9500 test now includes some basic test for 'forks' and 'ctags' features; the t9502 includes test checking if gitweb correctly filters out forks. Generating list of projects by scanning given directory is now also a bit simplified wrt. handling filtering; it is byproduct of extracting filtering forks to separate subroutine. While at it we now detect that there are no projects and respond with "404 No projects found" also for 'project_index' and 'opml' actions. Helped-by: Jonathan Nieder <jrnieder@gmail.com> Signed-off-by: Jakub Narebski <jnareb@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-04-29 19:51:56 +02:00
# ----------------------------------------------------------------------
# forks of projects
test_expect_success 'forks: setup' '
git init --bare foo.git &&
echo file > file &&
git --git-dir=foo.git --work-tree=. add file &&
git --git-dir=foo.git --work-tree=. commit -m "Initial commit" &&
echo "foo" > foo.git/description &&
git clone --bare foo.git foo.bar.git &&
echo "foo.bar" > foo.bar.git/description &&
git clone --bare foo.git foo_baz.git &&
echo "foo_baz" > foo_baz.git/description &&
rm -fr foo &&
mkdir -p foo &&
(
cd foo &&
git clone --shared --bare ../foo.git foo-forked.git &&
echo "fork of foo" > foo-forked.git/description
)
'
test_expect_success 'forks: not skipped unless "forks" feature enabled' '
gitweb_run "a=project_list" &&
grep -q ">\\.git<" gitweb.body &&
grep -q ">foo\\.git<" gitweb.body &&
grep -q ">foo_baz\\.git<" gitweb.body &&
grep -q ">foo\\.bar\\.git<" gitweb.body &&
grep -q ">foo_baz\\.git<" gitweb.body &&
grep -q ">foo/foo-forked\\.git<" gitweb.body &&
grep -q ">fork of .*<" gitweb.body
'
test_expect_success 'enable forks feature' '
cat >>gitweb_config.perl <<-\EOF
$feature{"forks"}{"default"} = [1];
EOF
'
gitweb: Restructure projects list generation Extract filtering out forks (which is done if 'forks' feature is enabled) into filter_forks_from_projects_list subroutine, and searching projects (via projects search form, or via content tags) into search_projects_list subroutine. Both are now run _before_ displaying projects, and not while printing; this allow to know upfront if there were any found projects. Gitweb now can and do print 'No such projects found' if user searches for phrase which does not correspond to any project (any repository). This also would allow splitting projects list into pages, if we so desire. Filtering out forks and marking repository (project) as having forks is now consolidated into one subroutine (special case of handling forks in git_get_projects_list only for $projects_list being file is now removed). Forks handling is also cleaned up and simplified. $pr->{'forks'} now contains un-filled list of forks; we can now also detect situation where the way for having forks is prepared, but there are no forks yet. Sorting projects got also refactored in a very straight way (just moving code) into sort_projects_list subroutine. The interaction between forks, content tags and searching is now made more explicit: searching whether by tag, or via search form turns off fork filtering (gitweb searches also forks, and will show all results). If 'ctags' feature is disabled, then searching by tag is too. The t9500 test now includes some basic test for 'forks' and 'ctags' features; the t9502 includes test checking if gitweb correctly filters out forks. Generating list of projects by scanning given directory is now also a bit simplified wrt. handling filtering; it is byproduct of extracting filtering forks to separate subroutine. While at it we now detect that there are no projects and respond with "404 No projects found" also for 'project_index' and 'opml' actions. Helped-by: Jonathan Nieder <jrnieder@gmail.com> Signed-off-by: Jakub Narebski <jnareb@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-04-29 19:51:56 +02:00
test_expect_success 'forks: forks skipped if "forks" feature enabled' '
gitweb_run "a=project_list" &&
grep -q ">\\.git<" gitweb.body &&
grep -q ">foo\\.git<" gitweb.body &&
grep -q ">foo_baz\\.git<" gitweb.body &&
grep -q ">foo\\.bar\\.git<" gitweb.body &&
grep -q ">foo_baz\\.git<" gitweb.body &&
grep -v ">foo/foo-forked\\.git<" gitweb.body &&
grep -v ">fork of .*<" gitweb.body
'
test_expect_success 'forks: "forks" action for forked repository' '
gitweb_run "p=foo.git;a=forks" &&
grep -q ">foo/foo-forked\\.git<" gitweb.body &&
grep -q ">fork of foo<" gitweb.body
'
test_expect_success 'forks: can access forked repository' '
gitweb_run "p=foo/foo-forked.git;a=summary" &&
grep -q "200 OK" gitweb.headers &&
grep -q ">fork of foo<" gitweb.body
'
test_expect_success 'forks: project_index lists all projects (incl. forks)' '
cat >expected <<-\EOF &&
gitweb: Restructure projects list generation Extract filtering out forks (which is done if 'forks' feature is enabled) into filter_forks_from_projects_list subroutine, and searching projects (via projects search form, or via content tags) into search_projects_list subroutine. Both are now run _before_ displaying projects, and not while printing; this allow to know upfront if there were any found projects. Gitweb now can and do print 'No such projects found' if user searches for phrase which does not correspond to any project (any repository). This also would allow splitting projects list into pages, if we so desire. Filtering out forks and marking repository (project) as having forks is now consolidated into one subroutine (special case of handling forks in git_get_projects_list only for $projects_list being file is now removed). Forks handling is also cleaned up and simplified. $pr->{'forks'} now contains un-filled list of forks; we can now also detect situation where the way for having forks is prepared, but there are no forks yet. Sorting projects got also refactored in a very straight way (just moving code) into sort_projects_list subroutine. The interaction between forks, content tags and searching is now made more explicit: searching whether by tag, or via search form turns off fork filtering (gitweb searches also forks, and will show all results). If 'ctags' feature is disabled, then searching by tag is too. The t9500 test now includes some basic test for 'forks' and 'ctags' features; the t9502 includes test checking if gitweb correctly filters out forks. Generating list of projects by scanning given directory is now also a bit simplified wrt. handling filtering; it is byproduct of extracting filtering forks to separate subroutine. While at it we now detect that there are no projects and respond with "404 No projects found" also for 'project_index' and 'opml' actions. Helped-by: Jonathan Nieder <jrnieder@gmail.com> Signed-off-by: Jakub Narebski <jnareb@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-04-29 19:51:56 +02:00
.git
foo.bar.git
foo.git
foo/foo-forked.git
foo_baz.git
EOF
gitweb_run "a=project_index" &&
sed -e "s/ .*//" <gitweb.body | sort >actual &&
test_cmp expected actual
'
xss() {
echo >&2 "Checking $*..." &&
gitweb_run "$@" &&
if grep "$TAG" gitweb.body; then
echo >&2 "xss: $TAG should have been quoted in output"
return 1
fi
return 0
}
test_expect_success 'xss checks' '
TAG="<magic-xss-tag>" &&
xss "a=rss&p=$TAG" &&
gitweb: escape URLs generated by href() There's a cross-site scripting problem in gitweb, where it will print URLs generated by its href() helper without further quoting. This allows an attacker to point a victim to a specially crafted gitweb URL and inject arbitrary HTML into the resulting page (which the victim sees as coming from gitweb). The base of the URL comes from evaluate_uri(), which pulls the value of $REQUEST_URI via the CGI module. It tries to strip off $PATH_INFO, but fails to do so in some cases (including ones that contain special characters, like "+"). Most of the uses of the URL end up being passed to "$cgi->a(-href = href())", which will get quoted properly by the CGI module. But in a few places, we output them ourselves as part of manually-generated HTML, and whatever was in the original URL will appear unquoted in the output. Given that all of the nearby variables placed into this manual HTML _are_ quoted, it seems like the authors assumed that these URLs would not need quoting. So it's possible that the bug is actually in evaluate_uri(), which should be doing a more careful job of stripping $PATH_INFO. There's some discussion in a comment in that function, as well as the commit message in 81d3fe9f48 (gitweb: fix wrong base URL when non-root DirectoryIndex, 2009-02-15). But I'm not sure I understand it. Regardless, it's a good idea to quote these values at the point of insertion into the HTML output: 1. Even if there is a bug in evaluate_uri(), this would give us belt-and-suspenders protection. 2. evaluate_uri() is only handling the base. Some generated URLs will also mention arbitrary refs or filenames in the repositories, and these should be quoted anyway. 3. It should never _hurt_ to quote (and that's what all of the $cgi->a() calls are doing already). So there may be further work here, but this patch at least prevents the XSS vulnerability, and shouldn't make anything worse. The test here covers the calls in print_feed_meta(), but I manually audited every call to href() to see how its output was used, and quoted appropriately. Most of them are esc_attr(), as they're used in tag attributes, but I used esc_html() when the URLs were printed bare. The distinction is largely academic, as one is implemented as a wrapper for the other. Reported-by: NAKAYAMA DAISUKE <nakyamad@icloud.com> Signed-off-by: Jeff King <peff@peff.net> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2019-11-15 10:06:07 +01:00
xss "a=rss&p=foo.git&f=$TAG" &&
xss "" "$TAG+"
'
gitweb: Restructure projects list generation Extract filtering out forks (which is done if 'forks' feature is enabled) into filter_forks_from_projects_list subroutine, and searching projects (via projects search form, or via content tags) into search_projects_list subroutine. Both are now run _before_ displaying projects, and not while printing; this allow to know upfront if there were any found projects. Gitweb now can and do print 'No such projects found' if user searches for phrase which does not correspond to any project (any repository). This also would allow splitting projects list into pages, if we so desire. Filtering out forks and marking repository (project) as having forks is now consolidated into one subroutine (special case of handling forks in git_get_projects_list only for $projects_list being file is now removed). Forks handling is also cleaned up and simplified. $pr->{'forks'} now contains un-filled list of forks; we can now also detect situation where the way for having forks is prepared, but there are no forks yet. Sorting projects got also refactored in a very straight way (just moving code) into sort_projects_list subroutine. The interaction between forks, content tags and searching is now made more explicit: searching whether by tag, or via search form turns off fork filtering (gitweb searches also forks, and will show all results). If 'ctags' feature is disabled, then searching by tag is too. The t9500 test now includes some basic test for 'forks' and 'ctags' features; the t9502 includes test checking if gitweb correctly filters out forks. Generating list of projects by scanning given directory is now also a bit simplified wrt. handling filtering; it is byproduct of extracting filtering forks to separate subroutine. While at it we now detect that there are no projects and respond with "404 No projects found" also for 'project_index' and 'opml' actions. Helped-by: Jonathan Nieder <jrnieder@gmail.com> Signed-off-by: Jakub Narebski <jnareb@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-04-29 19:51:56 +02:00
gitweb: Document current snapshot rules via new tests Add t9502-gitweb-standalone-parse-output test script, which runs gitweb as a CGI script from the commandline and checks that it produces the correct output. Currently this test script contains only tests of snapshot naming (proposed name of snapshot file) and snapshot prefix (prefix of files in the archive / snapshot). It defines and uses 'tar' snapshot format, without compression, for easy checking of snapshot prefix. Testing is done using check_snapshot function. Gitweb uses the following format for snapshot filenames: <sanitized project name>-<hash parameter><snapshot suffix> where <sanitized project name> is project name with '.git' or '/.git' suffix stripped, unless '.git' is the whole project name. For snapshot prefix it uses simply: <sanitized project name>/ Disadvantages of current snapshot rules: * There exists convention that <basename>.<suffix> archive unpacks to <basename>/ directory (<basename>/ is prefix of archive). Gitweb does not respect it * Snapshot links generated by gitweb use full SHA-1 id as a value of 'h' / $hash parameter. With current rules it leads to long file names like e.g. repo-1005c80cc11c531d327b12195027cbbb4ff9e3cb.tgz * For handcrafted URLs, where 'h' / $hash parameter is a symbolic 'volatile' revision name such as "HEAD" or "next" snapshot name doesn't tell us what exact version it was created from * Proposed filename in Content-Disposition header should not contain any directory path information, which means that it should not contain '/' (see RFC2183)... which means that snapshot naming is broken for $hash being e.g. hirearchical branch name such as 'xx/test' This would be improved in next commit. Signed-off-by: Jakub Narebski <jnareb@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2009-11-07 16:13:28 +01:00
test_done