The following code produces an endless loop of repeated child_added, though I never add a child. The nodes are removed. I just loop forever. How can I fix this?
//notifications branch is pre-populated when I get here:
notifications.startAt(null).on('child_added', processNotification, logError);
function processNotification(notification)
{
removeNotification(notification);
}
function removeNotification(notification)
{
var child = notifications.child(notification.name());
console.log("REMOVING", child.toString());
child.remove();
}
What I log is the same four nodes (A,B,C,D) being removed:
REMOVING https://my.firebaseio.com/experimental/-JVMi0_4poXfOWUt5g49 - A
REMOVING https://my.firebaseio.com/experimental/-JVMi1Y_bFwZAkffRel4 - B
REMOVING https://my.firebaseio.com/experimental/-JVMi2lyhKj8z27ik71x - C
REMOVING https://my.firebaseio.com/experimental/-JVMhzazdgHYAstqxu8L - D
REMOVING https://my.firebaseio.com/experimental/-JVMi0_4poXfOWUt5g49 - A
REMOVING https://my.firebaseio.com/experimental/-JVMi1Y_bFwZAkffRel4 - B
REMOVING https://my.firebaseio.com/experimental/-JVMi2lyhKj8z27ik71x - C
REMOVING https://mu.firebaseio.com/experimental/-JVMhzazdgHYAstqxu8L - D
...
I was able to reproduce the problem. I think it is caused by the fact that the Query tries to keep items in its "window" when you remove one.
You can work around this problem by using a once('value' instead of on('child_added', like this:
notifications.startAt(null).once('value', processNotifications);
function processNotifications(notifications)
{
notifications.forEach(function(notification) {
notification.ref().remove();
});
}
Since the value only fires once, it will not be interfering with the remove operation.
Ultimately, I just dropped the query and went with normal processing, dropping the "startAt(null)":
notifications.on('child_added', processNotification, logError);
It's not obvious that the behavior will be different from the documentation, but the startAt(null) does not produce the same behavior (though in thoery it should). I would love to see this better documented.
Related
Pls help me
A few weeks ago it came out of gamemaker 2.3, practically in the gamemaker language they changed the scripts into functions, but now after converting the files to be able to reopen them, I double-checked all the scripts and etc but anyway when I start it it remains a black screen, however it doesn't give me any compilation errors or whatever, what could be the problem?
Ps.
I might sound stupid, but if someone has the same program as me I can pass the project to them so they can see the scripts for themselves, so basically it's just the base and there is only the script to make the player walk and for collisions, I know that no one would want to waste time, but I ask the same
Its possible that your code is stuck in an infinite loop, here's an example of what that might look like:
var doloop = true
while(doloop == true){
x += 1
y += 1
}
the "doloop" variable is never changed within the while loop, so it is always equal to true and the loop never ends. Because the code never finishes looping, it can never get around to drawing anything, so you end up with a black screen. The easiest way to check for these is to put a breakpoint/debugging point at the beginning and just after every while/for/do/ect loop and debug it. e.g. (I am using asterisks "*" to represent breakpoints)
var doloop = true
* while(doloop == true){
x += 1
y += 1
}
*
When you get to one of the loops remove the first breakpoint and hit the "continue" button in the debugger. If it (it being the computer) takes an longer than it should to hit the second breakpoint (as in, you wait for a ten seconds to or two minutes (depends on how complex the code is) and it still hasn't hit the second breakpoint), then you should replace the breakpoint at the beginning of the loop to check and make sure it is still in there. If it is still in the loop, then that is likely where the code is getting stuck. Review the loop and everywhere any associated variables are set/changed, and you should be able to find the problem (even if it takes a while).
Majestic_Monkey_ and the commentors are correct: use the debugger. It's easy and it's your friend. Just place a red circle on the very first line of code that runs, and click the little bug icon and you can step through your code easily.
But to address your specific issue (or if anyone in the future has this issue): scripts have changed into files that can have many functions. Where you used to have
//script_name
var num = argument0 + argument1;
return num;
You would now have
function script_name(a, b) {
var num = a + b;
return num;
}
All you have to do is create a decleration for your new function:
function my_function_name(argument_names, etc...)
Then wrap all your old code in { }, and replace all those ugly "argument0" things with actual names. It's that easy. Plus you can have more than one function per script!
Is there a way to get the next or previous tag in AwesomeWM Lua Config?
Reason :- I want to get the next or previous tag to set the focused clients tag to it and then move to the tag.
I know there is awful.tag.viewnext which moves to the next tag but unsure about how to get the tag it will move to to set this on a client.
Thanks in advance
Well... yes, there is a way, but there is not a completely trivial one.
First, let's look at the default config. What existing functionality has to get next/previous tags? That is the mouse wheel on the taglist:
https://github.com/awesomeWM/awesome/blob/7a759432d3100ff6870e0b2b427e3352bf17c7cc/awesomerc.lua#L160-L161
So, how does awful.tag.viewnext and viewprev get the next tag? These functions just call viewidx with an idx of 1 or -1, e.g.:
https://github.com/awesomeWM/awesome/blob/7a759432d3100ff6870e0b2b427e3352bf17c7cc/lib/awful/tag.lua#L1494
So, awful.tag.viewidx can do what we want. How does it do it?
It gets a table of all un-hidden tags, finds the index of the currently selected tag and then uses gears.math.cycle to compute the index of the tag with the requested offset.
https://github.com/awesomeWM/awesome/blob/7a759432d3100ff6870e0b2b427e3352bf17c7cc/lib/awful/tag.lua#L1452-L1469
For you, something like the following should do the trick:
function get_tag_at_offset(i, s)
s = screen[s or awful.screen.focused()]
local tags = s.tags
local showntags = {}
for _, t in pairs(tags) do
if not awful.tag.getproperty(t, "hide") then
table.insert(showntags, t)
end
end
for k, t in ipairs(showntags) do
if t == sel then
return showntags[gears.math.cycle(#showntags, k + i)]
end
end
end
The above function can be used like get_tag_at_offset(1) to get the next tag and with argument -1 for the previous. Where possible, I would recommend to also pass in a screen via the second argument.
Also, all of this is completely untested and just written on the spot. There could very well be typos and other mistakes in here.
I'm having difficulty with possible empty relationships in gremlin3.
If I apply a filter on step (hasLabel) and the tranverser contains at least one label, everything runs fine. If not, transverse stops.
If I use optional, my condition gets ignored.
This is my query. The problem is on .hasLabel (within ($ {included})). I can get to work if I remove them.
result = await this.repository`g.V (${organization_uuid}).out()
.until(filter{{it.get().value("name")
.contains(${concept_name})}))
.repeat(inE().otherV().simplePath().timeLimit
(100)).dedup().as('target')
.bothE().as('relation').otherV()
.hasLabel(within(${included}))
.as('descendant').union(select('target',
'relation', 'descendant')
.by(valueMap(true)),outE('related').as
('direct').otherV().as('other')
.select ('direct','other').by(valueMap(true))) `;
I have a fill-in with the following code, made using the AppBuilder
DEFINE VARIABLE fichNoBuktiTransfer AS CHARACTER FORMAT "N(18)":U
LABEL "No.Bukti Transfer"
VIEW-AS FILL-IN NATIVE
SIZE 37.2 BY 1 NO-UNDO.
Since the format is N, it blocks the user from entering non-alphanumeric entries. However, it does not prevent the user from copypasting such entries into the fill-in. I have an error checking like thusly to prevent such entries using the on leave trigger:
IF LENGTH(SELF:Screen-value) > 18 THEN DO:
SELF:SCREEN-VALUE = ''.
RETURN NO-APPLY.
END.
vch-list = "!,*, ,#,#,$,%,^,&,*,(,),-,+,_,=".
REPEAT vinl-entry = 1 TO NUM-ENTRIES(vch-list):
IF INDEX(SELF:SCREEN-VALUE,ENTRY(vinl-entry,vch-list) ) > 0 THEN DO:
SELF:SCREEN-VALUE = ''.
RETURN NO-APPLY.
END.
END.
However, after the error handling kicked in, when the user inputs any string and triggers on leave, error 632 occurs:
error 632 occurs
Is there any way to disable the error message? Or should I approach the error handling in a different way?
EDIT: Forgot to mention, I am running on Openedge version 10.2B
You didn't mention the version, but I'll assume you have a version in which the CLIPBOARD system handle already exists.
I've simulated your program and I believe it shouldn't behave that way. It seems to me the error flag is raised anyway. My guess is even though those symbols can't be displayed, they are assigned to the screen value somehow.
Conjectures put aside, I've managed to suppress it by adding the following code:
ON CTRL-V OF FILL-IN-1 IN FRAME DEFAULT-FRAME
DO:
if index(clipboard:value, vch-list) > 0 then
return no-apply.
END.
Of course this means vch-list can't be scoped to your trigger anymore, in case it is, because you'll need the value before the leave. So I assigned the special characters list as an INIT value to the variable.
After doing this, I didn't get the error anymore. Hope it helps.
To track changes in a fill-in I always use at first this code:
ON VALUE-CHANGED OF FILL-IN-1 IN FRAME DEFAULT-FRAME
DO:
/* proofing part */
if ( index( clipboard:value, vch-list ) > 0 ) then do:
return no-apply.
end.
END.
You could add some mouse or developer events via AppBuilder to track changes in a fill-in.
I'm using Emacs 24 on Windows to write some R code. Up until about 30 minutes ago, whenever I would write a new function, ESS would automatically indent the lines following the function declaration and pressing the tab key on a new blank line would jump me to the appropriately indented starting position inside the declaration.
EG:
foo <- function() {
first line started here
second line here. .etc
}
Now, it is hard wrapping everything to the left, and not responding by automatically indenting after the function declaration or when I hit the tab key.
foo <- function() {
first line
second line
}
I've googled, but my google-fu is failing me on this. Anyone know how to restore default tab behavior to ESS in Emacs?
just for the record. Whenever such things happens, select the whole buffer C-x h and press C-M-\ to indent the whole region. This will show unambiguously the syntax error.
Try to add a space after "#".
I don't think ESS-mode handles # as a comment unless you have space after it.
I just came across the same problem you describe.
None of the above seemed to work, but I narrowed it down to using a carriage return and then an open parenthesis inside a string, like so:
### indent ( <tab> ) working fine up to here
s1 <- "string
(then this in brackets)"
### now indent does nothing!
The fact that it's balanced later doesn't help. I think EMACS reads this as opening a new expression/ block in spite of the fact that it occurs in a quoted string. This seems to apply also to expression openers { and [. It only seems to happen when the 'open expression' symbol appears at the start of the line...
In my case the string was part of a plot label, so the solution was to use \n instead.