summaryrefslogtreecommitdiff
path: root/src (follow)
Commit message (Collapse)AuthorAgeFilesLines
* darwin: fix mktemp argument orderJason A. Donenfeld2014-06-291-1/+1
|
* darwin: properly ejects ramdisksJason A. Donenfeld2014-06-292-15/+20
| | | | | | | In the move away from extensive global variables and improved cleanup routines, we forgot to fix the darwin platform file, which means temporary ramdisks never got unmounted. This patch cleans up the general cleanup trap logic routines.
* Allow options in EDITOR variableJason A. Donenfeld2014-06-291-1/+1
| | | | | | | | This is required by most graphical editors, as the shell command they provide needs to be told to wait until the file has been closed. This reverts '311c98c9140205f32be2720db9261c67da658c4c'. We thus use PATH so that the tests still pass in directories with spaces.
* Update for new URL.Jason A. Donenfeld2014-05-271-1/+1
|
* Bump version1.6.2Jason A. Donenfeld2014-05-141-1/+1
|
* Don't define tmpdir on git init.Jason A. Donenfeld2014-05-141-4/+3
|
* Disable secmem warning when discovering keysmilki2014-05-141-1/+1
| | | | | Key discovery parses stderr output of gnupg. The secmem warning interferes with parsing.
* edit: Quote editor so tests work in spaced directories.Jason A. Donenfeld2014-05-101-1/+1
|
* reencryption: Properly escape input to sedJason A. Donenfeld2014-05-101-1/+1
|
* git: use secure tmp directoryJason A. Donenfeld2014-05-101-4/+11
|
* configure git to decrypt gpg files automaticallySamuel Le Thiec2014-05-101-0/+5
| | | | | - usefull for git commands showing diff, e.g.: pass git log -p - from what I can see, not 'cleartext' temporary files are created
* completion: add new generate flagsJason A. Donenfeld2014-05-083-2/+7
|
* inplace: mutually exclusive with forceJason A. Donenfeld2014-05-081-2/+2
|
* usage: tab to spacesJason A. Donenfeld2014-05-081-1/+1
|
* generate: use nice ansi colors instead.Jason A. Donenfeld2014-05-081-3/+2
| | | | | | | | | Revert "Mute git-commit messages to make pass insert readable" This reverts commit f30ce6374d554e704162d5fa8e49acd9c6fd0ecc. I decided I like the git output. Instead highlight generated passwords using nice terminal output instead.
* zsh: posix compatible sed fix for zsh-completionJason A. Donenfeld2014-05-081-1/+1
| | | | | This reverts commit 56381287a16792b4c6410f07db68e02f3574c213, and further fixes things.
* Force sane sort order.Jason A. Donenfeld2014-05-061-2/+2
|
* generate: add --in-place optionJason A. Donenfeld2014-05-061-7/+21
|
* Mute git-commit messages to make pass insert readableMikhail Gusarov2014-05-061-1/+1
| | | | | | | | | | | | | | | | | | | `pass generate' prints 10 lines of output, with only 2 related to the operation at hand: % pass generate xyz 1 2 You need a passphrase to unlock the secret key for 3 user: USER <EMAIL> 4 4096-bit RSA key, ID XYZ, created 1234-56-78 5 6 [master 12345] Added generated password for xyz to store. 7 1 file changed, 0 insertions(+), 0 deletions(-) 8 create mode 100644 xyz.gpg 9 The generated password to xyz is: 10 Some password 123 % This patch takes care of 3 of 8 noise lines.
* zsh-completion: regression when using BSD sedKyle Marek-Spartz2014-05-061-1/+1
| | | | | | | | | | | | | Made sed expression GNU and BSD compatible. A recent change in pass.zsh-completion broke autocompletion when using BSD sed. I’ve made the relevant sed expression compatible with GNU and BSD sed. Previous change with regression: http://lists.zx2c4.com/pipermail/password-store/2014-April/000773.html http://git.zx2c4.com/password-store/commit/src/completion/pass.zsh-completion?id=f82e9d6cf3bc3a12bdfce89bf319d76f79e66efc
* edit: pass local variable to trap functionJason A. Donenfeld2014-04-291-2/+6
|
* base64: better justificationJason A. Donenfeld2014-04-271-3/+3
|
* Use die function instead of boring if else everywhere.Jason A. Donenfeld2014-04-241-69/+23
|
* version: bump1.6.1Jason A. Donenfeld2014-04-241-1/+1
|
* show: error out if password store is empty.Jason A. Donenfeld2014-04-241-1/+4
|
* find: proper segmenting of terms1.6Jason A. Donenfeld2014-04-241-3/+3
|
* Don't prompt for questions when stdin is not a tty.Jason A. Donenfeld2014-04-241-2/+3
|
* Allow overridable bash.Jason A. Donenfeld2014-04-241-1/+1
|
* Turns out aliases were a bad idea.Jason A. Donenfeld2014-04-243-27/+26
|
* Make gpg_opts an array.Jason A. Donenfeld2014-04-241-12/+12
|
* Shellcheck errors.Jason A. Donenfeld2014-04-241-6/+6
|
* git: the grammar nazis winJason A. Donenfeld2014-04-241-10/+10
|
* find: support tree 1.7.0Jason A. Donenfeld2014-04-231-14/+2
|
* grep: we need to not throw away the return valueJason A. Donenfeld2014-04-231-2/+2
|
* Use aliases instead of variables for command portability.Jason A. Donenfeld2014-04-233-26/+27
|
* reencrypt: More concise hack.Jason A. Donenfeld2014-04-231-1/+1
|
* reencrypt: Don't bother to save IFSJason A. Donenfeld2014-04-231-4/+1
|
* While loops don't make variables local.Jason A. Donenfeld2014-04-231-2/+3
|
* Elide more variables.Jason A. Donenfeld2014-04-231-26/+16
|
* reencrypt: support bash 3Jason A. Donenfeld2014-04-231-9/+5
|
* Elide variable declarations.Jason A. Donenfeld2014-04-221-29/+8
|
* reencryption: take into account groupsJason A. Donenfeld2014-04-221-10/+26
| | | | Signed-off-by: Jason A. Donenfeld <Jason@zx2c4.com>
* zsh-completion: only remove next char if /Jason A. Donenfeld2014-04-221-3/+2
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | From Kevin: I found a small bug in the zsh completions. Basically when the PASSWORD_STORE_DIR ends in a slash the first character of the result is eaten, making completion essentially useless. (It does this before determining matches). This can be fixed by changing what is line 106 in my version from: _values -C 'passwords' $(find -L "$prefix" \( -name .git -o -name .gpg-id \) -prune -o $@ -print | sed -e "s#${prefix}.##" -e 's#\.gpg##' | sort) to _values -C 'passwords' $(find -L "$prefix" \( -name .git -o -name .gpg-id \) -prune -o $@ -print | sed -e "s#${prefix}/\\?##" -e 's#\.gpg##' | sort) The difference is the first sed regex expression. The original version assumed that the next character was a slash and removed it while the new version only removes it if it is a slash. "s#${prefix}.##" -> "s#${prefix}/\\?##" Signed-off-by: Jason A. Donenfeld <Jason@zx2c4.com> Reported-by: Kevin Cox <kevincox@kevincox.ca>
* insert: exit with error if EOFJason A. Donenfeld2014-04-221-2/+2
|
* insert: Go to top of loop if invalid read.Jason A. Donenfeld2014-04-211-2/+2
|
* Use sort -u instead of uniqJason A. Donenfeld2014-04-201-2/+2
|
* Don't use subshells for reading from find.Jason A. Donenfeld2014-04-201-4/+4
|
* platform: Auto-detect local platform fileJason A. Donenfeld2014-04-191-1/+1
| | | | | | | In fact, if we're running from the source directory, just auto-detect the platform file in the first place. Signed-off-by: Jason A. Donenfeld <Jason@zx2c4.com>
* platform: Allow custom platform file before installJason A. Donenfeld2014-04-191-1/+1
| | | | | | | | | | | The new environment variable, PASSWORD_STORE_PLATFORM_FILE is now used for loading custom platform files while pass lives as src/password-store.sh. After it is installed using 'make install', this environment variable is no longer used, and either no platform file is loaded (if the default platform is acceptable), or a hardcoded also-installed platform file is referenced. Signed-off-by: Jason A. Donenfeld <Jason@zx2c4.com>
* reencrypt: cleaner temp file declarationJason A. Donenfeld2014-04-181-4/+4
|