I've just managed to get mutation testing working for the first time. My usual testing framework is Codeception but as of writing, it is not compatible with mutation testing (although I believe work is being done on it and it's not far off). I'm using PHPUnit and Infection, neither of which seem easy to work out how to use.
My test suite generated ten mutants. Nine were killed and one escaped. However, I don't know what part of the code or the tests needs to be improved to kill the final mutant.
How do you get information about what code allowed the mutant to escape?
I found in this blog what I couldn't find in Infection's documentation: the results are saved in infection.log.
The log file looks like this:
Escaped mutants:
================
1) <full-path-to-source-file>.php:7 [M] ProtectedVisibility
--- Original
+++ New
## ##
use stdClass;
trait HiddenValue
{
- protected function hidden_value($name = null, $value = null)
+ private function hidden_value($name = null, $value = null)
{
static $data = [];
$keys = array_map(function ($item) {
Timed Out mutants:
==================
Not Covered mutants:
====================
It says that the mutation changed the protected visibility to private and that no tests failed as a result. If this is important, I can now either change the code or write another test to cover this case.
Now that I've found this, I've searched on the Infection website for infection.log and found --show-mutations or -s which will output escaped mutants to the console while running.
I have a delta with some methods in it.
I want to create a schema extract that contains only the methods in the delta in an automated way so that I don't have to create one by hand or using the hateful selection tree in the Jade IDE.
The jadeworld documentation suggests I might be able to do it:
https://www.jadeworld.com/docs/jade-70/content/resources/userguide/chapter_10_-_extracting_and_loading_schemas/extracting_schemas_as_a_non-gui_client_application.htm
When I try, no extract files are created.
This is the command I am running:
jadclient path=E:\Jade63\System\ schema=JadeSchema ini=C:\Jade63\bin\jade.ini app=JadeBatchExtract endJade File d:\temp\delta.scm d:\temp\delta.ddb d:\temp\param.unl delta=TFS3274
Any help would be appreciated.
For 'File' extracts, you need to specify which schema to extract. This is the fourth parameter, after the UNL file, before adding the delta argument. I've added this to the example below, assuming 'Delta' is the schema name.
jadclient path=E:\Jade63\System\ schema=JadeSchema ini=C:\Jade63\bin\jade.ini app=JadeBatchExtract endJade File d:\temp\delta.scm d:\temp\delta.ddb d:\temp\param.unl Delta delta=TFS3274
Unfortunately, I'm not sure if this will extract just the methods that are in the specified delta. Rather, I believe everything specified by the UNL file will be extracted, but where any methods are checked out to a delta, the version in the specified delta will be extracted.
You'll need to experiment to confirm, but in my experience, patches are more suitable for performing extracts without needing to specify what's changed.
Kevin's has answered the question I asked, I'm just adding this bit here for anyone else who happens this way. I was trying to automate creating a UNL file from a delta. The following perl script will generate a UNL file from a schema extract file. So you can create a schema extract from a delta in the IDE, then run this script on it to create a UNL, which you can then use for creating subsequent extracts.
#!/usr/bin/perl
$state="init";
$class="";
$method="";
#result=();
while(<>)
{
if($state eq "init")
{
if(m/typeDefinitions/)
{
$state="inTypes";
}
}
elsif($state eq "inTypes")
{
if(m/[^(]+\(\r/)
{
$state="inClass";
($class=$_) =~ s/\s*(\S+).*\(/$1/;
$class =~ s/[\r\n]//g;
}
elsif(m/inverseDefinitions/)
{
$state="done";
}
}
elsif($state eq "inClass")
{
if(m/jadeMethodDefinitions/)
{
$state="inMethod";
}
elsif(m/^\s*\)\r/)
{
$state="inTypes";
}
}
elsif($state eq "inMethod")
{
if(m/[^(]+[(]/)
{
($method=$_) =~ s/\s*(\S+)\(.*/$1/;
$method =~ s/[\r\n]//g;
$state="inClass";
push #result, "Method $class $method\n";
}
}
}
#result = sort #result;
print #result;
print "\n";
I am writing some code where I have multiple dictionaries for my data. The reason being, I have multiple core objects and multiple smaller assets and the user must be able to choose a smaller asset and have some function off in the distance run the code with the parent noted.
An example of one of the dictionaries: (I'm working in ROBLOX Lua 5.1 but the syntax for the problem should be identical)
local data = {
character = workspace.Stores.NPCs.Thom,
name = "Thom", npcId = 9,
npcDialog = workspace.Stores.NPCs.Thom.Dialog
}
local items = {
item1 = {
model = workspace.Stores.Items.Item1.Main,
npcName = "Thom",
}
}
This is my function:
local function function1(item)
if not items[item] and data[items[item[npcName]]] then return false end
end
As you can see, I try to index the dictionary using a key from another dictionary. Usually this is no problem.
local thisIsAVariable = item[item1[npcName]]
but the method I use above tries to index the data dictionary for data that is in the items dictionary.
Without a ton of local variables and clutter, is there a way to do this? I had an idea to wrap the conflicting dictionary reference in a tostring() function to separate them - would that work?
Thank you.
As I see it, your issue is that:
data[items[item[npcName]]]
is looking for data[“Thom”] ... but you do not have such a key in the data table. You have a “name” key that has a “Thom” value. You could reverse the name key and value in the data table. “Thom” = name
I have been using Postman so much and built a lot of useful things in it, but now need to implement one more thing.
Briefly:
Need to create test cases to test a correct records counting for every institution.
Now I handled it like that
Structure:
---------Collection
-------Folder
---Districts(folder)
-some folders with tests
---Colleges(folder)
-some folders with tests
---Schools(folder)
-Principal(Folder with test in Schools folder)
where these requests:
POST: Create a list
where in Tests:
var jList = JSON.parse(responseBody);
postman.setEnvironmentVariable("list_id", jList.data.id);
POST: Add some filters to list
GET: lists/{{list_id}} whete in "Test" code:
```
var allLists = JSON.parse(responseBody);
pm.test("test count", function () {
const value = allLists.data.count;
pm.expect(typeof value === 'number').to.eql(true);
pm.expect(value > 0 && value < 999999).to.eql(true);
}); // I check if the number(made in step№2 in a range we need
```
DELETE: delete this list
I guess I do 'overhead work' adding POST(create a list) and -DELETE endpoints in every folder, can I somehow bring it out to environment or variable and execute it before every POST: Add some filters to list and -DELETE endpoint after that
Something like general BeforeEach and AfterEach
Perhaps can I even bring this(below) out in one place for every separate test?
pm.test("test count", function () {
const value = allLists.data.count;
pm.expect(typeof value === 'number').to.eql(true);
pm.expect(value > 0 && value < 999999).to.eql(true);
});
Here is an example how it looks like if it not difficult, give me a piece of advice! Thanks
I have a need to crawl a set of data of indeterminate size and build a table key/value index of it. Since I don't know the dimensions in advance, seems I have to use a recursive function. My Lua skills are very new and superficial. I'm having difficulty understanding how to deal with returning a table from the function call.
Note this is for a Lua 5.1 script processor
API = function(tbl)
local table_api = {}
-- do stuff here with target data and add to table_api
table_api["key"] = value
-- then later there is a need to recurse deeper into target data
table_api["lower"] = API(var)
return table_api
end
result = API(source_data)
In most every other language I know there would be some way to make the recurse line table_api["lower"] = API(var) work but since Lua does table variables by reference, my return sub-table just keeps getting overwritten and my result is a tiny last bit of what it should be.
Just for background on my purpose: there's a commercial application I'm working with that has a weakly documented Lua scripting interface. It's running Lua 5.1 and the API is not well documented and frequently updated. As I understand it, everything is stored in _G, so I wanted to write something to reverse engineer the API. I have a working recursive function (not shown here) that enumerates all of _G. For that return value, it just builds up an annotated string and progressively builds on the string. That all works fine and is really useful, but it shows much more that the API interface; all the actual data elements are included, so I have to sift through like 30,000 records to determine an API set of about 500 terms. In order to determine the API, I am trying to use this sub-table return value recursive function being discussed in this question. The code I'm showing here is just a small distilled subset of the larger function.
I'll go ahead and include the full code here. I was hoping to incrementally build a large'ish table of each API level, any sublevels, and finally whatever keys used at the lowest level.
In the end, I was expecting to have a table that I could address like this:
result["api"]["label"]["api"]["sublabel"]["value"]["valuename"]
Full code:
tableAPIShow = function(tbl, table_track)
table_track = table_track or {}
local table_api = {}
if type(tbl) == 'table' then
-- Check if values are tables.
local parent_table_flag = true
for ind,val in pairs(tbl) do
if type(val) ~= 'table' then
parent_table_flag = false
break
end
end
-- If all children are table type, check each of them for subordinate commonality
local api_flag = false
if parent_table_flag == true then
local child_table = {}
local child_table_flag = false
api_flag = true
for ind,val in pairs(tbl) do
-- For each child table, store the names of the indexes.
for sub_ind,sub_val in pairs(val) do
if child_table_flag == false then -- First time though, create starting template view of typical child table.
child_table[sub_ind] = true -- Store the indexes as a template table.
elseif child_table[sub_ind] == nil then -- Otherwise, test this child table compared to the reference template.
api_flag = false
break
end
end
if api_flag == false then -- need to break out of nested loop
break
end
child_table_flag = true
end
end
if api_flag == true then
-- If everything gets to here, then this level is an API with matching child tables below.
for ind,val in pairs(tbl) do
if table_api["api"] == nil then
table_api["api"] = {}
end
table_api["api"][ind] = tableAPIShow(val, table_track)
end
else
-- This level is not an API level, determine how to process otherwise.
for ind,val in pairs(tbl) do
if type(val) == 'table' then
if table_track[val] ~= nil then -- Have we already recursed this table?
else
table_track[val] = true
if table_api["table"] == nil then
table_api["table"] = {}
end
table_api["table"][ind] = tableAPIShow(val, table_track)
end
else -- The children are not tables, they are values
if table_api["value"] == nil then
table_api["value"] = {}
end
table_api["value"][ind] = val
end
end
end
else
-- It's not a table, just return it.
-- Probably never use this portion because it's caught on upper level recurse and not called
return tbl
end
return table_api
end
And I was calling this function in the main script like this:
local str = tableAPIShow(_G)
I've got another function that recursively shows a table so I can look inside my results and see I only get a return value that contains only the values of the top-level of _G (I have it excluded built-in Lua functions/values because I'm only interested in the Application API):
{
[value] = table: 00000000F22CB700 {
[value][_VERSION] = Application/5.8.1 (x86_64; Windows NT 10.0.16299),
[value][tableAPIShow] = "function: 00000000F22C6DE0, defined in (121-231) C:\\Users\\user\\AppData\\Local\\Temp\\APP\\/~mis00002690 ",
[value][_FINAL_VERSION] = true,
[value][Path] = ./Scripts/Database/elements/,
[value][class] = "function: 00000000F1953C40, defined in (68-81) Scripts/Common/Class.lua ",
[value][db_path] = ./Scripts/Database/,
[value][merge_all_units] = "function: 00000000F20D20C8, defined in (2242-2250) Scripts/Database/db_merge.lua ",
}
You just need to localize the variable you store your table in and it will work as you expect:
local table_api = {}
(note that you are passing table variable that conflicts with the global table variable and is not currently used in the function.)
I am inclined to believe that your tableAPIShow function code is working correctly and the
another function that recursively shows a table
fails to serialize your table fully. Hence you don't see the deeper levels of the table returned by tableAPIShow().
I got your initial and current code (tableAPIShow) to work with my simple table serialize function: the whole _Global table is exported completely and formatted as you implemented it in your tableAPIShow().
Code for testing:
apiMassiveTable = {
api = {
get = {
profile = {"profileID", "format"},
name = {"profileID", "encoding"},
number = {"profileID", "binary"}
},
set = {
name = {"apikey", "profileID", "encoding", "newname"},
number = {"apikey", "profileID", "binary", "newnumber"}
},
retrieve = {}
},
metadata = {version="1.4.2", build="nightly"}
}
-- tableAPIShow implemenation here
table.serialize = dofile("serialize.lua")
print(table.serialize("myNameForAHugeTable", tableAPIShow(_G)))
PS: Whatever serialize function you're using, it should enquote strings like Application/5.8.1 (x86_64; Windows NT 10.0.16299) which it does not.
Like #Paul Kulchenko said, you need to learn to use locals (https://www.lua.org/pil/4.2.html). Global variable post-exist until a new lua_State is loaded (a new environment, could be a new process depending on what interpreter you are using). So a tip is to always use local variables for anything you don't want to leave the function or leave the compilation unit.
Think of tables like dictionaries: a word is attached to a definition. Thusly the definition is the data.
What I think you are trying to do is serialization of a table of data. However, that isn't really necessary. You can either shadow copy or deep copy the given table. A shadow copy is when you don't delve into the depths of tables found in the keys, etc. A deep copy is when you copy tables in keys of tables in keys of tables... etc.
local shallow_copy = function(tab)
local rep_tab = {}
for index, value in pairs(tab)do
rep_tab[index] = value
end
return rep_tab
end
-- because local variable is not defined or declared immediately on a 1-liner,
-- a declaration has to exist so that deep_copy can be used
-- lets metatable execute functions
local deep_copy
deep_copy = function(tab)
local rep_tab = {}
for index, value in pairs(tab)do
if(type(value) == "table")then
rep_tab[index] = deep_copy(value)
else
rep_tab[index] = value
end
end
return rep_tab
end
Deco's deepcopy.lua
https://gist.github.com/Deco/3985043
You can also index tables using periods:
local tab = {}
tab.abc = 123
tab["def"] = 456
print(tab.abc, tab["def"])
To serialize the entire _G, you would just filter out the junk you don't need and recurse each table encountered. Watch out for _G, package, and _ENV because if defined it will recurse back to the start.
-- use cap as a whitelist
enumerate_dir = function(dir, cap)
local base = {}
for i, v in pairs(dir) do
-- skip trouble
if(i ~= "_G" and i ~= "_ENV" and i ~= "package")then
if(type(v) == "table")then -- if we have a table
base[i] = enumerate_dir(v, cap)
else
for k, n in pairs(cap) do
if(type(v) == n)then -- if whitelisted
base[i] = tostring(v)
end
end
end
end
end
return base
end
-- only functions and tables from _G please
local enumeration = enumerate_dir(_G, {"function", "table"})
-- do some cool quips to get a basic tree system
prefix = ""
recursive_print = function(dir)
for i, v in pairs(dir)do
if(type(v) == "table")then
print(prefix, i, v)
prefix = prefix .. ">"
recursive_print(v)
else
print(prefix, i, v)
end
end
if(#prefix > 0)then
prefix = prefix:sub(1, #prefix-1)
end
end
recursive_print(test)