about summary refs log tree commit diff stats
path: root/src/lua.c
Commit message (Collapse)AuthorAgeFilesLines
...
* .Kartik K. Agaram2021-11-261-2/+2
|
* extract a functionKartik K. Agaram2021-11-251-10/+18
|
* more stack hygieneKartik K. Agaram2021-11-251-0/+2
|
* .Kartik K. Agaram2021-11-251-7/+7
|
* dedup definitions in big-picture viewKartik K. Agaram2021-11-251-4/+40
| | | | We only want to show most recent version of each binding.
* new shortcut: return to big-picture viewKartik K. Agaram2021-11-251-9/+14
|
* teliva is now ASan-clean againKartik K. Agaram2021-11-251-0/+4
| | | | | | | | | At least in short runs. Encouraging that the problem was in a recent commit (5a63a5ca40 from yesterday when I introduced version control). Disabling Address Sanitizer again.
* renameKartik K. Agaram2021-11-241-9/+9
|
* monotonically accumulate versions of definitionsKartik K. Agaram2021-11-241-71/+137
| | | | | | | | | | | One old drawback now has a new look. Before, we loaded definitions in order, so global definitions had to exist before other global definitions that used them. See window and grid in life.tlv. Now we load definitions in reverse order, so initialization needs to change. Worse, if we update window, we need to edit grid just to fix the order. This implies that we can't yet optimize away bindings where there are no new changes.
* .Kartik K. Agaram2021-11-241-3/+3
|
* consistent file opsKartik K. Agaram2021-11-241-6/+6
| | | | Now we're down to 1 real warning and 1 false positive.
* styleKartik K. Agaram2021-11-241-3/+3
|
* renameKartik K. Agaram2021-11-241-7/+7
|
* indentKartik K. Agaram2021-11-241-32/+32
|
* standardize warning flags everywhereKartik K. Agaram2021-11-221-10/+7
| | | | | | | | | | I'd like to enable -Wextra as well, but that creates some false positives. I've at least made my changes clean w.r.t. -Wextra. Now we have 4 remaining warnings with gcc 9.3 that seem genuine. Need to fix those.
* https now working!Kartik K. Agaram2021-11-211-0/+4
| | | | | | | | Still extremely ugly: - I've inlined all the namespaces under ssl, so you need to know that context and config are related to ssl. - luasec comes with its own copy of luasocket. I haven't deduped that yet.
* basic http requests starting to workKartik K. Agaram2021-11-211-0/+10
| | | | | In the process we're starting to load almost all of luasocket by default. And everything is working as expected, no unpleasant surprises.
* now. we. have. JSON.Kartik K. Agaram2021-11-211-0/+2
| | | | Completely unmodified from upstream.
* luasocket now loading properlyKartik K. Agaram2021-11-201-0/+2
| | | | I still haven't tried actually running it.
* extract a helperKartik K. Agaram2021-11-201-8/+17
|
* bugfix: support running from top-levelKartik K. Agaram2021-11-201-1/+1
| | | | Unfortunately we can't currently run teliva from anywhere else :/
* port changes from minimal to maximal versionKartik K. Agaram2021-11-201-7/+19
| | | | From lcurseslib.c to lcurses/ directory.
* report errors when calling non-existent functionsKartik K. Agaram2021-11-201-2/+4
|
* inline in C is not worth the troubleKartik K. Agaram2021-11-201-2/+2
| | | | https://merveilles.town/@akkartik/107310347838372198
* inline lcurses maximally rather than minimallyKartik K. Agaram2021-11-191-0/+1
| | | | | | | | | Until now we had just the bare minimum bindings needed for the demos built so far. Now we have all of lcurses building in place with minimal changes. The changes in this commit can run hanoi.lua when inlined into Lua 5.1, but don't work with Teliva.
* import life.teliva into image formatKartik K. Agaram2021-11-141-0/+1
|
* cleaner error handlingKartik K. Agaram2021-11-141-17/+7
|
* .Kartik K. Agaram2021-11-141-2/+1
|
* remain in editor on errorKartik K. Agaram2021-11-141-5/+20
|
* .Kartik K. Agaram2021-11-141-2/+2
|
* find _some_ way to show error on initial loadKartik K. Agaram2021-11-141-2/+8
|
* more menu cleanupKartik K. Agaram2021-11-141-1/+4
|
* .Kartik K. Agaram2021-11-141-18/+18
|
* drop support for non-image Lua scriptsKartik K. Agaram2021-11-141-35/+8
| | | | We still have no story for error messages. We'll work on that next.
* draw the browse dialog the same wayKartik K. Agaram2021-11-141-9/+45
|
* smarter image browsingKartik K. Agaram2021-11-141-19/+53
| | | | | - distinguish between unused functions and data - don't hardcode a specific convention for the curses window object
* slightly more robust on-disk formatKartik K. Agaram2021-11-141-2/+2
| | | | | | Looks like Lua supports a little bit of programmability in its multi-line string literals. Even though I can't find this documented anywhere.
* janky way to cancel browsingKartik K. Agaram2021-11-141-6/+9
|
* slightly more obvious browse dialogKartik K. Agaram2021-11-141-1/+4
|
* cleaner layout for a function's code and dataKartik K. Agaram2021-11-131-6/+54
|
* .Kartik K. Agaram2021-11-131-0/+1
|
* .Kartik K. Agaram2021-11-131-4/+1
|
* .Kartik K. Agaram2021-11-131-1/+1
|
* clean up after the editorKartik K. Agaram2021-11-131-0/+1
| | | | | Even though we don't need to leave ncurses mode, there's still a process boundary to clean up before.
* .Kartik K. Agaram2021-11-131-1/+2
|
* .Kartik K. Agaram2021-11-131-1/+1
|
* replace Kilo's raw terminal handling with ncursesKartik K. Agaram2021-11-131-1/+0
| | | | | | This was on my todo list. What made it urgent was finding that calling getch() even once while in ncurses caused Kilo to stop detecting arrow keys. No need to debug that sort of nonsense.
* avoid side-effects on the Lua stackKartik K. Agaram2021-11-131-0/+5
|
* stop leaking memoryKartik K. Agaram2021-11-131-2/+3
|
* memory corruption bugKartik K. Agaram2021-11-131-1/+3
| | | | | | | | | I was saving an address on the stack to a global, and it was getting clobbered later. This is the sort of thing I completely eliminated in https://github.com/akkartik/mu :/ Now I'm taking a leaf out of the Mu playbook and leaking a little bit of memory every time I switch definitions.