m-chrzan.xyz
aboutsummaryrefslogtreecommitdiff
path: root/src/password-store.sh
AgeCommit message (Collapse)Author
2014-03-19Simplify exports.Jason A. Donenfeld
2014-03-19Copyright noticeJason A. Donenfeld
2014-03-19Team pass: enable multiple keys and per directoryJason A. Donenfeld
The .gpg-id file may now have multiple keys in it, one per line. If a .gpg-id file exists inside a subdirectory, passwords inside that directory are encrypted to that/those ids. The init command has learned a -p/--path option for writing such a sub directory .gpg-id and now can take several arguments for ids.
2014-03-18Shred shm files.Jason A. Donenfeld
2014-03-18Do not compress passwords.Jason A. Donenfeld
According to a forthcoming paper by Alfredo Pironti, OpenPGP compression can reveal entropy levels. We thus disable compression. Existing password stores can be reencrypted without compression using the "--reencrypt" flag for "init". Reported-by: Alfredo Pironti <alfredo.pironti@inria.fr> Signed-off-by: Jason A. Donenfeld <Jason@zx2c4.com>
2013-09-14Use a glob in the pattern match instead of using two comparisons for [yY].Chris Down
2013-09-14Rephrase awkward/confusing message about enabling echo during password entry.Chris Down
2013-09-14Fix directory traversal for reencryption when $PREFIX is a symlinkBrian Shore
2013-05-22check if a passfile exists before checking for directoryBrian Mattern
2013-04-04Simplify sed to not use replacement.Jason A. Donenfeld
2013-02-23Tree should follow symlinks.Marvin Killing
Make show/ls/list follow links by passing -l to tree.
2012-10-18Bump version.Jason A. Donenfeld
2012-10-17Do not interpret backslashes when reading passwordsguns
The `read` builtin accepts backslash notation for common non-printing characters by default, like `\t` and `\n`. This requires that any literal backslashes must also be escaped as `\\`. Given that `gpg -e` does not interpret input, the `read` invocations are changed to do the same. Also, the right hand side of an `==` comparison within `[[ ]]` must be quoted in order to suppress pattern metacharacter expansion. Quoting the bash manual: When the == and != operators are used, the string to the right of the operator is considered a pattern and matched according to the rules described below under Pattern Matching.
2012-10-09Restore exclusivity of options in multiline/echo check.Jason A. Donenfeld
2012-10-05Trim trailing slash in tree display.Jason A. Donenfeld
2012-10-03Make --no-echo default, and rename to --echo.Jason A. Donenfeld
Reported-by: Paul Wise <pabs@debian.org>
2012-09-30Check error code of gpg before copying to clipboard.Jason A. Donenfeld
2012-09-30Explicitly use gpg2 rather than relying on symlink.Jason A. Donenfeld
2012-09-24Style fixes.Jason A. Donenfeld
2012-09-22Normalize negation and bashism.Jason A. Donenfeld
2012-09-22Fix directory removal.Jason A. Donenfeld
2012-09-22Add man page comment to usage message.Jason A. Donenfeld
2012-09-22Bump version to 1.4.Jason A. Donenfeld
2012-09-22Normalize copyright headers.Jason A. Donenfeld
2012-09-22Allow git repo to be elsewhere via env var.Jason A. Donenfeld
2012-09-21Add support for PASSWORD_STORE_KEY env var.Jason A. Donenfeld
2012-09-21Add option to init to reencrypt all passwords.Jason A. Donenfeld
Reported-by: Simon KP <si@eskp.net>
2012-09-20Quit if pwgen returns nothing.Jason A. Donenfeld
Reported-by: Brian Mattern <rephorm@rephorm.com>
2012-09-20Prompt before overwriting password in generate, unless --force is provided.Jason A. Donenfeld
2012-09-19Make qdbus ugliness more segmented.Jason A. Donenfeld
2012-09-19Since we're using batch mode, we can read directly.Jason A. Donenfeld
2012-09-19Clean up git handling.Jason A. Donenfeld
Get rid of push/pull shortcuts, as they weren't widely used. Add contents to repo on git init. Centralize git add logic, and make it less error prone.
2012-09-19Fix git regression.Jason A. Donenfeld
2012-09-18Quote the program name.Jason A. Donenfeld
2012-09-17Abstract potentially platform specific commands into their own commands.Jason A. Donenfeld
2012-09-16Support recursive and forced removal.Jason A. Donenfeld
Laurent asked for this. Reported-by: Laurent Ghigonis <laurent@p1sec.com>
2012-09-16Small stylistic things.Jason A. Donenfeld
Reported-by: Laurent Ghigonis <laurent@p1sec.com>
2012-09-14Bump to 1.3.1.Jason A. Donenfeld
2012-09-13Typo.Jason A. Donenfeld
2012-09-13Bump version.Jason A. Donenfeld
2012-09-13Allow avoiding the prompt for overwriting, with --force flag.Jason A. Donenfeld
2012-09-13Split out common gpg options into variable.Jason A. Donenfeld
Reported-by: Brian Mattern <rephorm@rephorm.com>
2012-09-13use gpg --batch for quieter outputBrian Mattern
Right now, every time I call pass to decrypt a key, I get output like: You need a passphrase to unlock the secret key for user: "User Name <user@domain.com>" 2048-bit ELG-E key, ID XXXXXXNX, created 2012-04-20 (main key ID NNXXNNNX) password This patch cleans it up so that only the password is output. I use pinentry-gtk-2 to enter the gpg passphrase, so this information is redundant to me. I haven't tried other pinentry versions to see if they repeat the information as well.
2012-09-13prevent <ctrl-c> on password entry from wiping out fileBrian Mattern
Currently, if you hit ctrl-c at the standard 'Enter password' prompt, since it is piped directly to gpg, the entry gets cleared. Trying to read from that entry results in: gpg: [don't know]: 1st length byte missing This patch fixes this. Tweaked by Jason A. Donenfeld <Jason@zx2c4.com> to add GNU readline features by using -e in read.
2012-09-13Confirm before overwriting an entryBrian Mattern
This asks before inserting a password when one already exists at that location (instead of just overwriting it).
2012-09-13Use bash's read command better.Jason A. Donenfeld
2012-09-13Add --version switch, per request.Jason A. Donenfeld
2012-09-13Be more careful about cleaning up.Jason A. Donenfeld
2012-09-12Use getopt to regularize optionsBrian Mattern
I like being able to specify command line options (like -c) at the end of the line (which is usually when I think of the fact that I need them). The attached patch uses getopt(1) to regularize the option list so that lazy people like me can specify the options in any order.
2012-09-12Don't quote ${EDITOR:-vi}rupa
Attached is a trivial patch that removes the quotes from the line that invokes $EDITOR. It's perfectly cromulent to set $EDITOR to something with spaces in it, so when we evaluate this one we want it evaluated bare. For example security nerds might want EDITOR='vim -n' if they are scared of swap files, and that breaks if we quote there.