aboutsummaryrefslogtreecommitdiff
path: root/shell/ash.c (follow)
Commit message (Collapse)AuthorAgeFilesLines
* Merge branch 'busybox' into mergeRon Yorston2018-05-131-28/+64
|\
| * ash: expand: Fix buffer overflow in expandmetaDenys Vlasenko2018-04-141-25/+42
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Upstream commit: Date: Sun, 25 Mar 2018 16:38:00 +0800 expand: Fix buffer overflow in expandmeta The native version of expandmeta allocates a buffer that may be overrun for two reasons. First of all the size is 1 byte too small but this is normally hidden because the minimum size is rounded up to 2048 bytes. Secondly, if the directory level is deep enough, any buffer can be overrun. This patch fixes both problems by calling realloc when necessary. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au> function old new delta expmeta 517 635 +118 expandarg 990 996 +6 mklocal 288 290 +2 ------------------------------------------------------------------------------ (add/remove: 0/0 grow/shrink: 3/0 up/down: 126/0) Total: 126 bytes Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
| * ash: parser: Allow newlines within parameter substitutionDenys Vlasenko2018-04-111-1/+4
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Upstream commit: Date: Thu, 22 Mar 2018 21:41:24 +0800 parser: Allow newlines within parameter substitution On Fri, Mar 16, 2018 at 11:27:22AM +0800, Herbert Xu wrote: > On Thu, Mar 15, 2018 at 10:49:15PM +0100, Harald van Dijk wrote: > > > > Okay, it can be trivially modified to something that does work in other > > shells (even if it were actually executed), but gets rejected at parse time > > by dash: > > > > if false; then > > : ${$+ > > } > > fi > > That's just a bug in dash's parser with ${} in general, because > it bombs out without the if clause too: > > : ${$+ > } This patch fixes the parsing of newlines with parameter substitution. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au> Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
| * hush: do not drop backslash from eval 'echo ok\'Denys Vlasenko2018-04-111-1/+1
| | | | | | | | | | | | | | | | | | | | | | | | newer bash does not drop it, most other shells too function old new delta unbackslash 39 57 +18 parse_stream 2753 2751 -2 ------------------------------------------------------------------------------ (add/remove: 0/0 grow/shrink: 1/1 up/down: 18/-2) Total: 16 bytes Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
| * shell: add comments about [[, no code changesDenys Vlasenko2018-04-101-1/+14
| | | | | | | | Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
| * ash: if "[[" bashism is not supported, do not handle it anywhereDenys Vlasenko2018-04-101-0/+2
| | | | | | | | Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
| * ash: trivial code shrinkDenys Vlasenko2018-04-101-1/+2
| | | | | | | | | | | | | | function old new delta parse_command 1677 1674 -3 Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
* | ash: make loopnest a global variableRon Yorston2018-04-131-1/+2
| | | | | | | | Fixes the test ash-misc/break5.tests.
* | ash: make commandname a global variableRon Yorston2018-04-121-2/+7
| | | | | | | | | | | | | | commandname is static so isn't copied during forkshell. This was causing the arith-tests script in the shell test suite to fail because without commandname some error reports were missing line numbers.
* | ash: make struct forkshell a local variable in evalbackcmdRon Yorston2018-04-101-7/+7
| | | | | | | | | | | | There's no reason for the forkshell structure to be passed between expbackq and evalbackcmd in the backcmd structure. It can be a local variable, as in the other similar cases.
* | Merge branch 'busybox' into mergeRon Yorston2018-04-091-104/+147
|\|
| * bzip2: fix two crashes on corrupted archivesDenys Vlasenko2018-04-081-1/+0
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | As it turns out, longjmp'ing into freed stack is not healthy... function old new delta unpack_usage_messages - 97 +97 unpack_bz2_stream 369 409 +40 get_next_block 1667 1677 +10 get_bits 156 155 -1 start_bunzip 212 183 -29 bb_show_usage 181 120 -61 ------------------------------------------------------------------------------ (add/remove: 1/0 grow/shrink: 2/3 up/down: 147/-91) Total: 56 bytes Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
| * hush: implement "hush -s"Denys Vlasenko2018-04-051-1/+1
| | | | | | | | | | | | | | | | | | | | function old new delta hush_main 1015 1031 +16 packed_usage 32757 32745 -12 ------------------------------------------------------------------------------ (add/remove: 0/0 grow/shrink: 1/1 up/down: 16/-12) Total: 4 bytes Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
| * ash: unbreak PS1 parsing after "ash: parser: Add syntax stack..." commitDenys Vlasenko2018-04-041-1/+3
| | | | | | | | Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
| * ash: parser: Fix parameter expansion inside inner double quotesDenys Vlasenko2018-04-021-2/+4
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Upstream email: parser: Fix parameter expansion inside inner double quotes The parsing of parameter expansion inside inner double quotes breaks because we never look for ENDVAR while innerdq is true. echo "${x#"${x+''}"''} This patch fixes it by pushing the syntax stack if innerdq is true and we enter a new parameter expansion. This patch also fixes a corner case where a bad substitution error occurs within arithmetic expansion. Reported-by: Denys Vlasenko <vda.linux@googlemail.com> Fixes: ab1cecb40478 (" parser: Add syntax stack for recursive...") Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au> function old new delta readtoken1 2880 2898 +18 Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
| * ash: expand: Fix bugs with words connected to the right of $@Denys Vlasenko2018-04-021-4/+7
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Upstream email: This is actually composed of two bugs. First of all our tracking of quotemark is wrong so anything after "$@" becomes quoted. Once we fix that then the problem is that the first space character after "$@" is not recognised as an IFS. This patch fixes both. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au> Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
| * ash: redir: Fix typo in noclobber codeDenys Vlasenko2018-04-021-1/+1
| | | | | | | | | | | | Upstream commit "redir: Fix typo in noclobber code" Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
| * ash: parser: Fix single-quoted patterns in here-documentsDenys Vlasenko2018-04-021-1/+1
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Upstream commit: From: Herbert Xu <herbert@gondor.apana.org.au> Date: Fri, 9 Mar 2018 23:07:53 +0800 parser: Fix single-quoted patterns in here-documents The script x=* cat <<- EOF ${x#'*'} EOF prints * instead of nothing as it should. The problem is that when we're in sqsyntax context in a here-document, we won't add CTLESC as we should. This patch fixes it: Reported-by: Harald van Dijk <harald@gigawatt.nl> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au> Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
| * ash: parser: Add syntax stack for recursive parsingDenys Vlasenko2018-04-021-97/+134
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | This closes 10821. Upstream patch: From: Herbert Xu <herbert@gondor.apana.org.au> Date: Fri, 9 Mar 2018 00:14:02 +0800 parser: Add syntax stack for recursive parsing Without a stack of syntaxes we cannot correctly these two cases together: "${a#'$$'}" "${a#"${b-'$$'}"}" A recursive parser also helps in some other corner cases such as nested arithmetic expansion with paratheses. This patch adds a syntax stack allocated from the stack using alloca. As a side-effect this allows us to remove the naked backslashes for patterns within double-quotes, which means that EXP_QPAT also has to go. This patch also fixes removes any backslashes that precede right braces when they are present within a parameter expansion context, and backslashes that precede double quotes within inner double quotes inside a parameter expansion in a here-document context. The idea of a recursive parser is based on a patch by Harald van Dijk. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au> var_bash3, var_bash4 and var_bash6 tests are updated with the output given by bash-4.3.43 With this patch, the following tests now pass for ash: dollar_repl_slash_bash2.tests squote_in_varexp2.tests squote_in_varexp.tests var_bash4.tests function old new delta readtoken1 2615 2874 +259 synstack_push - 54 +54 evalvar 574 571 -3 rmescapes 330 310 -20 subevalvar 1279 1258 -21 argstr 1146 1107 -39 ------------------------------------------------------------------------------ (add/remove: 1/0 grow/shrink: 1/4 up/down: 313/-83) Total: 230 bytes Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
* | win32: more backslash to slash conversionsRon Yorston2018-04-081-0/+3
| | | | | | | | | | | | | | | | | | Convert backslashes to slashes: - in the main program's argv[0] - in the shell's file argument Also, fix the code to remove the .exe extension from argv[0] in the main program.
* | ps: indicate forkshell processes in listingRon Yorston2018-04-051-3/+4
| |
* | ps: obtain applet names from other BusyBox processesRon Yorston2018-04-041-7/+0
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Remove the code which passed applet names to child processes using environment variables. This only allowed ps to display names for its ancestors. Instead attempt to read applet names from the memory of unrelated processes. The Microsoft documentation alone wasn't enough to figure out how to do this. Additional hints from: https://stackoverflow.com/questions/4298331/exe-or-dll-image-base-address https://stackoverflow.com/questions/14467229/get-base-address-of-process
* | ash: reinstate applet name variable after forkshellRon Yorston2018-04-031-0/+7
| | | | | | | | | | | | When 'sh --forkshell' is invoked the BB_APPLET_<pid> environment variable is set, but then it's cleared when the environment of the child process is restored. Reinstate it in reinitvar.
* | ash: minor forkshell fixesRon Yorston2018-04-021-7/+5
| | | | | | | | | | | | | | | | - update forkshell_evalbackcmd to match the non-MinGW code; - increase the size of the buffer to hold the forkshell handle for 64-bit machines; - use memcpy to copy the globals_var structure to avoid copying bits by hand.
* | Merge branch 'busybox' into mergeRon Yorston2018-04-021-53/+49
|\|
| * parser: Fix backquote support in here-document EOF markDenys Vlasenko2018-04-011-0/+6
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Upstream commit: Author: Herbert Xu <herbert@gondor.apana.org.au> Date: Thu Mar 15 18:27:30 2018 +0800 parser: Fix backquote support in here-document EOF mark Currently using backquotes in a here-document EOF mark is broken because dash tries to do command substitution on it. This patch fixes it by checking whether we're looking for an EOF mark during tokenisation. Reported-by: Harald van Dijk <harald@gigawatt.nl> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au> With added fix for quoted-ness of the EOF mark. Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
| * ash,hush: handle a few more bkslash-newline casesDenys Vlasenko2018-04-011-1/+1
| | | | | | | | Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
| * ash: use pgetc_eatbnl() in more placesDenys Vlasenko2018-03-311-35/+12
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Part of upstream commit: Date: Thu Mar 8 08:37:11 2018 +0100 Author: Harald van Dijk <harald@gigawatt.nl> parser: use pgetc_eatbnl() in more places dash has a pgetc_eatbnl function in parser.c which skips any backslash-newline combinations. It's not used everywhere it could be. There is also some duplicated backslash-newline handling elsewhere in parser.c. Replace most of the calls to pgetc() with calls to pgetc_eatbnl() and remove the duplicated backslash-newline handling. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au> Not adding "readtoken1(pgetc_eatbnl(), DQSYNTAX..." changes, since readtoken1() handles the "starts with backslash + newline" case itself. Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
| * ash: 'nolog' and 'debug' options cause "$-" to wreak havocMartijn Dekker2018-03-311-1/+1
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Upstream commit: Date: Tue Mar 6 17:40:37 2018 +0000 expand: 'nolog' and 'debug' options cause "$-" to wreak havoc Op 29-03-17 om 20:02 schreef Martijn Dekker: > Bug: if either the 'nolog' or the 'debug' option is set, trying to > expand "$-" silently aborts parsing of an entire argument. > > $ dash -o nolog -c 'set -fuC; echo "|$- are the options|"; set +o nolog; echo "|$- are the options|"' > | > |uCf are the options| > $ dash -o debug -c 'set -fuC; echo "|$- are the options|"; set +o debug; echo "|$- are the options|"' > | > |uCf are the options| This turned out to be easy to fix. The routine producing the "$-" expansion failed to skip options for which there is no option letter, but only a long-form name. In dash, 'nolog' and 'debug' are currently the only two such options. Patch below. - Martijn Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au> In bbox ash, pipefail is the option which exhibited this. Signed-off-by: Martijn Dekker <martijn@inlv.org> Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
| * ash: fix "char == CTLfoo" comparison signedness bugDenys Vlasenko2018-03-301-3/+3
| | | | | | | | | | | | | | It usually does not bite since bbox forces -funsigned-char build. But for some reason void linux people disabled that. Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
| * ash: use F_DUPFD_CLOEXEC and O_CLOEXECDenys Vlasenko2018-03-301-9/+16
| | | | | | | | | | | | | | | | | | | | | | | | function old new delta setjobctl 371 367 -4 setinputfile 226 220 -6 ------------------------------------------------------------------------------ (add/remove: 0/0 grow/shrink: 0/2 up/down: 0/-10) Total: -10 bytes Based on patch by Mark Marshall <mark.marshall@omicronenergy.com> Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
| * ash,hush: fix "saved" redirected fds still visible in childrenDenys Vlasenko2018-03-281-5/+11
| | | | | | | | | | | | | | | | | | | | Based on a patch by Mark Marshall <mark.marshall@omicronenergy.com> function old new delta dup_CLOEXEC - 49 +49 fcntl_F_DUPFD 46 - -46 Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
* | ash: add some debug for the forkshell data blockRon Yorston2018-04-011-0/+87
| |
* | ash: align funcblocksize handling with upstreamRon Yorston2018-04-011-68/+76
| | | | | | | | | | | | Upstream BusyBox removed the global funcblocksize variable. Instead a local variable with the same name is passed to and returned from all functions that calculate item sizes. Do the same here.
* | ash: align funcstring handling with upstreamRon Yorston2018-04-011-27/+30
| | | | | | | | | | | | | | | | | | Upstream BusyBox removed the funcstringsize variable. String sizes are included in funcblocksize then strings are copied into the data block starting from the end. Do the same here. Add a function to calculate string lengths, allowing for NULL string pointers (as nodeckstrup does already for MinGW).
* | ash: reorder items in forkshell data blockRon Yorston2018-04-011-48/+47
| | | | | | | | | | | | Put the nodeptr array before funcblock and funcstring. Since nodeptr now immediately follows the forkshell structure in the data block we can declare it in the structure and do away with nodeptr_offset.
* | ash: revise handling of builtin environment variables during forkRon Yorston2018-03-311-32/+24
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Commit db03ff5ea made some improvements to the handling of builtin environment variables during forkshell. However, more can be done. Currently the calculated size of varinit doesn't match what's actually copied. It overestimates the amount of string space and the number of pointers. The text of each builtin environment variable is also copied twice: once as part of vartab and again from varinit. Remove the code to copy varinit via the forkshell block. Instead use information from vartab and varinit_data to reinitialise varinit and fix the links in vartab. In addition, ensure that the LINENO variable is properly initialised in the child.
* | ash: copy environment variables when importing themRon Yorston2018-03-301-0/+4
| | | | | | | | | | | | | | | | | | Clearing the environment on Microsoft Windows results in pointers used for shell variables becoming invalid. Take copies when the environment is imported. This only affects NOFORK applets invoked by tryexec, but the number of NOFORK applets has increased considerably in recent years.
* | ash: use SIGTERM to kill processesRon Yorston2018-03-271-1/+9
| | | | | | | | | | When Ctrl-C is pressed try killing processes with SIGTERM initially. If they're still running after a short pause use sterner measures.
* | win32: add a function to convert backslashes to slashesRon Yorston2018-03-181-6/+2
| |
* | ash: remove special treatment of device filesRon Yorston2018-03-151-20/+0
| | | | | | | | | | Since device files are now handled in mingw_open there's no need for any special treatment in ash redirection.
* | Merge branch 'busybox' into mergeRon Yorston2018-03-011-6/+17
|\|
| * ash : fix double-quoted "\z" handlingDenys Vlasenko2018-02-131-6/+17
| | | | | | | | | | | | | | function old new delta readtoken1 2602 2608 +6 Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
* | Remove fake signal-handling codeRon Yorston2018-03-011-3/+10
| | | | | | | | | | | | | | | | | | Microsoft Windows has only limited support for signals. busybox-w32 initially papered over this fact by adding definitions for unsupported signals and signal-handling functions. Remove this fake code and deal with the consequences by excluding anything that fails to compile as a result.
* | ash: changes to command resolution, execution and displayRon Yorston2018-02-281-23/+55
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | ash separates searching for commands from running them. Searching is performed in find_command(). The result of the search is stored in a hash table but not as the full path: the path is reconstructed in shellexec() by combining the directory and command name. In Windows the command name is insufficient, as the executable name may also include an extension. To resolve this we must ensure that extensions are processed in the same order in find_command, shellexec and any other places where command names are resolved into file names. The order used matches that of spawnve: .com, .exe, .bat, .cmd. Finally the bare filename with no additional extension The order has been made consistent: - in find_command for absolute paths and the Windows-specific path search; - in tryexec (which is called by shellexec) where an additional test that the file is executable has been added; - in the type and hash built-ins where the output has been modified to include the extension. Additionally, the code in tryexec to handle ENOEXEC errors from execve has been excluded. The rationale for the exec functions in POSIX [1] suggests the requirement to retry is a workaround for the lack of '#!' support on some platforms. [1] http://pubs.opengroup.org/onlinepubs/009695399/functions/exec.html
* | win32: move detection of file formats to stat(2)Ron Yorston2018-02-261-5/+1
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Move the code to detect shell scripts and binary executables from mingw_access to a separate function, has_exec_format. Call this function in do_lstat to decide whether to set the executable bits in the file mode. This will slow down stat but has a couple of advantages: - shell scripts are highlighted in ls output - the test applet can use stat(2) to detect executable files The new function is used to handle another corner case in spawnveq: binary executables without the usual .exe extension are only run by spawnve if the file name ends with '.'. Two minor changes: - file_is_win32_executable has been renamed add_win32_extension to clarify what it does - a call to file_is_executable has been removed from find_command in ash as it resulted in unhelpful error messages.
* | win32: add support for batch filesRon Yorston2018-02-261-18/+22
| | | | | | | | | | | | | | | | Support batch files with .bat and .cmd extensions, similar to what's done for .exe and .com. Check extensions in the same order as Windows' spawn function: .com, .exe, .bat, .cmd.
* | ash: tidy up /dev/null redirectionRon Yorston2018-02-221-11/+6
| | | | | | | | | | Remove some duplicated code and raise an error for unsupported devices.
* | ash: use correct size for tblentryRon Yorston2018-02-191-2/+2
| | | | | | | | | | | | | | | | | | | | As noted in cmdlookup the size of a tblentry instance, *t, is sizeof(struct tblentry) + strlen(t->cmdname) because the structure includes one byte for the cmdname array. Apply the same logic in tblentry_size and tblentry_copy.
* | ash: allow waitpid_child to blockRon Yorston2018-02-171-1/+1
| | | | | | | | | | The wrong sort of boolean operator resulted in waitpid_child always being nonblocking.
generated by cgit v1.2.3-55-g6feb (git 2.39.1) at 2025-03-19 19:59:16 +0000