summaryrefslogtreecommitdiff
Commit message (Collapse)AuthorAgeFilesLines
* Add NEWS file.HEADv1.0.8masterDaniel Silverstone2017-08-021-0/+16
| | | | | | | This should be updated whenever a new feature is added to Supple. It is not meant to be a full changelog, so only put headline items in this file please. The first time a change happens after a new version, please remember to add the new title.
* security: clear environmentRichard Ipsum2017-07-301-6/+5
| | | | Motivation: https://www.dwheeler.com/secure-programs/Secure-Programs-HOWTO/environment-variables.html
* Remove supple_paths.h because it was confusing dh-luaDaniel Silverstone2017-07-233-11/+5
| | | | | | | | | | | | Sadly, the approach Richard took to introduce supple_paths.h broke the Debian packaging because dh-lua doesn't use the Makefile. I looked at ways to use it, but because dh-lua builds for all the versions of Lua at once, it can't easily work as-is. Trying to add support for the different Lua versions in Makefile turned out to be a pain, so instead I merged the defines support into the rest of the defines we use. It's less elegant but should be equivalent.
* Allow baking of LUA_PATH and LUA_CPATH env varsRichard Ipsum2017-07-203-4/+31
| | | | | | These vars get unset for security reasons, but on NixOS these vars are needed to determine the location of dependent libs, since they're not available in their usual fhs positions.
* Allow lua link option to be overriddenRichard Ipsum2017-07-201-2/+3
|
* Squash compiler warning of ignored ret valueRichard Ipsum2017-07-081-3/+8
|
* setuid only if rootRichard Ipsum2017-07-081-2/+4
| | | | | | Running the sandbox without setuid is less secure, but it will still work, it is not possible for nix to make binaries setuid when being used as a pkg manager outside of NixOS, i.e. on Debian.
* Do not link capi.so against the lua interpreterdsilvers/link-fixupDaniel Silverstone2017-07-081-1/+1
|
* Move wrapper path from supple.host into supple.capiv1.0.7Daniel Silverstone2016-10-173-3/+6
|
* Block GC during make_call to be more safev1.0.6Daniel Silverstone2016-10-141-0/+2
|
* Fix a bit of bad documentationv1.0.5Daniel Silverstone2016-05-301-1/+1
|
* Change return ABI for supple.host.run()Daniel Silverstone2016-05-301-1/+6
|
* Add CoCDaniel Silverstone2016-05-301-0/+20
|
* Add user documentation. Skipping internal docs for nowDaniel Silverstone2015-11-024-2/+146
|
* Better support testing supple during things like package buildsv1.0Daniel Silverstone2015-10-153-7/+16
|
* Support debugging the simple example a bit more, and ensure we hit ram limitsDaniel Silverstone2015-10-151-2/+8
|
* Force some garbage collection so that we don't get confused ordering in repliesDaniel Silverstone2015-10-151-1/+5
|
* Allow tests to run in the absence of luacovDaniel Silverstone2015-10-158-8/+8
|
* Also clean html docsDaniel Silverstone2015-07-031-0/+1
|
* Very beginnings of ldoc supportDaniel Silverstone2015-07-033-0/+11
|
* Fix indented code sectionDaniel Silverstone2015-07-031-2/+2
|
* Support 5.2 style load() as well as 5.1's loadstring()/setfenv() comboDaniel Silverstone2015-02-031-2/+7
|
* Add fake __gc and support newer library registrationDaniel Silverstone2015-02-031-0/+14
|
* Don't use luaconf.h defines, they're not stable, also switch away from ↵Daniel Silverstone2015-02-031-4/+7
| | | | lua_cpcall
* Install the supple binary into libDaniel Silverstone2013-04-251-2/+2
|
* PREFIX, not INST_ROOT, pleaseDaniel Silverstone2013-04-141-1/+3
|
* Submodule instructionsDaniel Silverstone2012-09-201-0/+5
|
* Move luacov out to a submoduleDaniel Silverstone2012-09-0913-986/+3
|
* TEST: Ensure we don't lock down during sandbox.run() testDaniel Silverstone2012-08-211-0/+2
|
* TEST: Remove dubious 'not optimised' assertDaniel Silverstone2012-08-201-1/+0
|
* OBJECTS: Don't use tostring() because it might loop via __tostring infinitelyDaniel Silverstone2012-08-161-2/+2
|
* Merge branch 'diagnosis'Daniel Silverstone2012-08-1511-39/+289
|\
| * OBJECTS: Refcount tags (proxies) so that we only pass __gc across when we ↵diagnosisDaniel Silverstone2012-08-131-4/+14
| | | | | | | | really mean it
| * TRACK: Cope with negative depths and nil tracksDaniel Silverstone2012-08-131-1/+6
| |
| * SANDBOX: Update sandbox and wrapper so that we write to stderr (dead, but ↵Daniel Silverstone2012-08-132-2/+18
| | | | | | | | who cares) if an error propagates all the way out
| * TEST: Add missing requireDaniel Silverstone2012-08-131-1/+1
| |
| * SUPPLE: Lots of diagnosis stuffDaniel Silverstone2012-08-1210-35/+254
|/
* Credits fileDaniel Silverstone2012-08-111-0/+11
|
* EXAMPLE: Add demonstration for supple.host.loadstring()Daniel Silverstone2012-08-081-0/+16
|
* HOST: Simple loadstring for the sandboxDaniel Silverstone2012-08-081-0/+5
|
* SANDBOX: Ensure the first wrappered load is direct and the rest are wrapperedDaniel Silverstone2012-08-081-2/+9
|
* SANDBOX: Ensure we use the same globals for every loadstringDaniel Silverstone2012-08-081-5/+3
|
* HOST: Allow set_globals to define a set of globals to pass to all supple ↵Daniel Silverstone2012-08-081-8/+18
| | | | sandboxes
* HOST: Ensure sandbox's FDs are more likely to be closed cleanlyDaniel Silverstone2012-08-061-2/+3
|
* EXAMPLE: Show pairs, ipairs and next working, and allow lprint() to show ↵Daniel Silverstone2012-08-051-1/+47
| | | | simple tabular returns too
* HOST: Ensure we simplify return values to local types so we can clean down ↵Daniel Silverstone2012-08-051-0/+28
| | | | if the function returns tables etc
* COMMS: if read() fails, propagate the errorDaniel Silverstone2012-08-051-1/+4
|
* SUPPLE: Ensure __next is set on proxied tables by defaultDaniel Silverstone2012-08-052-0/+6
|
* SANDBOX: Wrapper unpack() so it works transparentlyDaniel Silverstone2012-08-051-1/+14
|
* SANDBOX: Use the CAPI variants of ipairs,pairs,next for transparent proxy ↵Daniel Silverstone2012-08-051-1/+4
| | | | behaviour
View Lorry Controller Status