Commit Graph

89 Commits

Author SHA1 Message Date
ridiculousfish
dc8cc39bc8 Don't autosuggest in builtin_read
https://github.com/fish-shell/fish-shell/issues/29
2012-11-18 02:43:35 -08:00
ridiculousfish
d76f880faf Support for fish_right_prompt
Fixes https://github.com/fish-shell/fish-shell/issues/80
2012-11-07 19:59:20 -08:00
ridiculousfish
e52cf09bc1 Simplify handling of escape sequences in echo -e 2012-10-17 17:08:45 -07:00
ridiculousfish
57de1388e8 Make the \c special character in echo suppress the newline too 2012-10-17 02:59:43 -07:00
ridiculousfish
7ee0ce745e Implement -e option to echo (to interpret special characters), and -E to not interpret them
https://github.com/fish-shell/fish-shell/issues/337
2012-10-17 02:56:03 -07:00
Ian Munsie
ffc03735e8 Don't barf when ending a FAKE block
After updating, fish would immediately die when started in an xterm from
my window manager (wmii) with a message like:

fish: builtin.cpp:3357: int builtin_end(parser_t&, wchar_t**): Assertion `false' failed.
fish: Job 1, “~/code/fish-shell/fish” terminated by signal SIGABRT (Abort)

Starting fish from an existing shell worked, and running it on other
machines that I was only SSHing into worked as well. I did discover that
on one machine starting git from one specific directory
(/home/ian/bml/kernel) would cause this failure, but starting from any
other directory worked as normal. I'm not entirely sure what
circumstances contribute to this failure - starting from a copy of the
kernel directory works fine.

The failure only started with the following commit, which introduced the
assert(false):

commit d788c84440
Made type property of block_t constant and private
Further work towards cleaning up block_t hierarchy

Looking at this fail in gdb:

(gdb) r
Starting program: /home/ian/code/fish-shell/fish
[Thread debugging using libthread_db enabled]
Using host libthread_db library "/lib/x86_64-linux-gnu/libthread_db.so.1".
fish: builtin.cpp:3357: int builtin_end(parser_t&, wchar_t**): Assertion `false' failed.

Program received signal SIGABRT, Aborted.
0x00007ffff6c82475 in *__GI_raise (sig=<optimized out>) at ../nptl/sysdeps/unix/sysv/linux/raise.c:64
64      ../nptl/sysdeps/unix/sysv/linux/raise.c: No such file or directory.
(gdb) bt
    function=0x548520 "int builtin_end(parser_t&, wchar_t**)") at assert.c:81
(gdb) up 3
3357                    assert(false); //should never get here
(gdb) p parser.current_block->type()
$1 = FAKE

So this happens when we run an end command for a FAKE block.

The below patch adds an empty case for FAKE blocks to avoid hitting the
assert. I would need to study the code in more detail to understand if
we should even be executing this code if we are in a FAKE block, but
this patch seems to solve the issue for me.

Signed-off-by: Ian Munsie <darkstarsword@gmail.com>
2012-09-22 23:51:41 -07:00
ridiculousfish
1a59346b51 Changed "elseif" to "else if" 2012-09-03 13:24:01 -07:00
ridiculousfish
ff124465fd Clean up some warnings and some unused if-related code 2012-09-01 12:29:00 -07:00
ridiculousfish
cc1395797e First stab at elseif implementation 2012-09-01 01:46:14 -07:00
ridiculousfish
95de6cf5a7 Migrated function_data_t out of base block class
Removed an auto_ptr (yay)
2012-08-26 23:30:23 -07:00
ridiculousfish
d788c84440 Made type property of block_t constant and private
Further work towards cleaning up block_t hierarchy
2012-08-26 23:16:20 -07:00
ridiculousfish
96046ffd30 First stab at getting rid of the ugly state1, state2 properties of block_t 2012-08-26 22:42:29 -07:00
ridiculousfish
61686aff34 Adopt posix_spawn (!)
Rewrite IO chains to be a vector of pointers, instead of a linked list
Removed io_transmogrify
2012-08-15 00:57:56 -07:00
ridiculousfish
84729c4dfa Additional warnings cleanup, effective C++ violations, dead code removal 2012-08-05 13:24:33 -07:00
ridiculousfish
c67702a498 Cleaned up lots of typecasts, simplified some string handling 2012-08-04 15:11:43 -07:00
ridiculousfish
b904aa78e8 Additional warning cleanup and switching from int to size_t where appropriate 2012-08-04 11:34:45 -07:00
ridiculousfish
8185bee4b8 Lots of work towards making fish build without warnings on Mountain Lion, mostly in terms of using size_t instead of int 2012-08-04 11:34:45 -07:00
ridiculousfish
e7cbcc83a4 Implemented history deletion from fish_config
Fixes https://github.com/fish-shell/fish-shell/issues/250
2012-07-27 00:31:00 -07:00
ridiculousfish
62c49f13ce Switch from std::list to std::vector in a few places to reduce compiled code size 2012-07-24 22:32:11 -07:00
ridiculousfish
261bf12c91 Lots of miscellaneous cleanup. Unified the path_get_cd_path, path_allocate_cd_path, etc. functions 2012-07-20 22:11:05 -07:00
maxfl
0c03b6ddc7 Allow 'contains' builtin to return index
Add an option -i/--index to the contains builtin to print the index
of first coincidence.
2012-07-10 16:21:00 -07:00
maxfl
232ed91ec7 Fix read default prompt. It's now single line. 2012-07-06 15:31:45 -07:00
maxfl
ea4b37d5c5 Fixed case behaviour
* case no properly handles -h and --help flags, i.e. treats it as
  pattern
* fixed case escaping:
The following expressions now work correctly:
switch '*'
  echo '*'
  echo Match any string
end
switch '*'
  echo '\*'
  echo Match asterix
end
switch '\\'
  echo '\\\\'
  echo Match slash
end

The same for '?' sign
2012-07-05 14:17:25 -07:00
ridiculousfish
7bbc7a61ce Clarify the function name deferral in functions_def
Prohibit making a function with an empty name
2012-07-01 15:33:50 -07:00
maxfl
ab62fe6496 Update funced
* Editor mode is no the default
* Use -i or --interactive or -e fish to edit function in interactive
  mode
* tmpname is now created with random number added and check that file
  do not already exist
* check $TMPDIR existence and put /tmp if it does not exist
* There is an undocumented feature to use functions, started with dash.
  Introduce necessary changes to funced, functions, def_function()
  in order to make it work properly.
* Delete editor guessing. Use $EDITOR variable or -e key
2012-07-01 14:20:43 -07:00
Kevin Ballard
adfd3c1098 Don't blow away $status in switch statements
Tweak the switch and case builtins to not blow away $status
inappropriately. Fixes issue #161.
2012-06-24 13:43:44 -07:00
ridiculousfish
f69489a4fc Mostly fix https://github.com/fish-shell/fish-shell/issues/72 2012-06-21 16:58:26 -07:00
Siteshwar Vashisht
25f9105a97 Made history --help show history man page and history is now saved only once while deleting items 2012-06-12 14:20:07 +05:30
Siteshwar Vashisht
ca61d0ee8b Fixed crash in history builtin for inputs like : history --search -prefix "echo" 2012-06-10 01:08:19 +05:30
ridiculousfish
7698553c3a Made a bare history call end with a newline 2012-06-04 23:59:04 -07:00
Siteshwar Vashisht
ec34f2527a Adding history builtin 2012-06-05 09:54:42 +05:30
ridiculousfish
d09c9fba02 Added a -s option to echo to mean "don't output spaces"
Replaced default prompt to stop calling printf. I opened a bug to replace the default prompt entirely with one that's cheaper.
2012-05-19 16:59:56 -07:00
ridiculousfish
da6295c428 More work towards getting function.h off of shared_ptr 2012-05-17 19:46:08 -07:00
ridiculousfish
86645b32e1 Work towards getting function.h off of shared_ptr 2012-05-17 19:38:08 -07:00
ridiculousfish
9bcc7df96f Lots of modernization changed
Removed wcsdupcat
2012-05-09 02:33:42 -07:00
ridiculousfish
3be07cc645 Fix to start re-respecting CDPATH 2012-05-05 14:30:20 -07:00
ridiculousfish
0bc644abf0 Fix lots of bugs related to the static analyzer
Improved how screen.cpp interacts with output_set_writer()
2012-03-26 01:21:10 -07:00
ridiculousfish
31b7d076b7 Remove some dead variables.
Fix screwy output for invalid tilde expansion in expand.cpp
Some cleanup per clang static analyzer
2012-03-25 23:31:03 -07:00
ridiculousfish
a11687fc5c Make the functions builtin have a bit nicer output
Stop autosuggesting things with newlines
Make webconfig a little nicer
2012-03-25 22:41:22 -07:00
ridiculousfish
f92b24221a Made pwd a builtin
Fixed a thread error when autosuggesting with a command substitution
2012-03-09 20:16:26 -08:00
ridiculousfish
d173bb6e0a A bunch of changes working towards eliminating all memory allocation after fork() 2012-03-08 23:21:07 -08:00
ridiculousfish
ce859c9e92 Added some missing calls to setup_fork_guards() in utilities.
Made echo a builtin
2012-03-07 11:35:22 -08:00
ridiculousfish
6788c2710f Improved error reporting for builtin_test 2012-03-07 01:02:46 -08:00
ridiculousfish
ed7c10f366 Implemented test as a builtin (!) 2012-03-07 00:54:01 -08:00
ridiculousfish
8a46931e34 Remove string_buffer_t (!) 2012-03-03 22:08:34 -08:00
ridiculousfish
3ead99b088 Put fish on a diet. Tracked down the biggest memory hogs and fixed them. Total allocations down by a factor of 3 or so, live allocations a few KB. 2012-03-03 15:22:03 -08:00
ridiculousfish
2da8df6202 Rearrange wopen and stat to avoid a race and hopefully improve performance a little bit. 2012-03-02 11:12:08 -08:00
Peter Ammon
8b26d0104c Some initial changes to use CLO_EXEC, with an eye towards some day using it correctly. 2012-03-02 00:27:40 -08:00
ridiculousfish
fdfa5c0602 Some initial work towards resolving nasty fork/pthread issues, and to having a per-parser job list 2012-02-27 18:43:24 -08:00
ridiculousfish
38e40862fe More work towards autosuggesting completions 2012-02-25 18:54:49 -08:00