dynamic configuration with boost program_options - c++

Is there a way to load a dynamic INI file like the one below.
[basic]
number_of_servers=3
[server1]
ip=10.20.30.40
password=sdfslkhf
[server2]
ip=10.20.30.41
password=sdfslkhf
[server3]
ip=10.20.30.42
password=sdfslkhf
Here the idea is that the servers that are defined here is very specific to the software's deployment; so the admin decides how many servers participate in the configuration.
Is there a way to handle this in boost program_options?

Another, potentially more standard way, would be like this:
[basic]
number_of_servers=3
[server]
name=server1
ip=10.20.30.40
password=sdfslkhf
[server]
name=server2
ip=10.20.30.41
password=sdfslkhf
[server]
name=server3
ip=10.20.30.42
password=sdfslkhf
This way you don't need to worry about undefined section names, and I think this style is more widely used as well (definitely it's how QuickFIX does it, in a way very similar to what I outlined).
And you can probably remove the number_of_servers entry, and just use the count() function to find how many server sections there are.

There is an optional bool parameter to allow for unregistered entries in the parse_config_file function. It's set to false by default. See the documentation here:
http://www.boost.org/doc/libs/1_45_0/doc/html/boost/program_options/parse_config_file_id991860.html
If you call this function with true then it will add any unregistered entries into the variables_map as strings. You can check whether they exist with the variables_map::count function.
I hope that helps.

Sure you can. The server sections have a pattern: just load all those matching the pattern into a list of servers.

The challenges I faced while resolving this was to make sure the sections are kept together and are no way mixed up.
In the end I relied on a options_description with the known/finite options and then using the parsed_options that come out of parse_config_file, I had to collect all unrecognized options ( collect_unrecognized ). Then I had to iterate it to pick the options in order.
Thanks every one for their contribution.

Related

C++ PostgreSQL Library, describe SQL, without execution

I have asked this question on Github, issue #641, since it seems like a deficiency or an enhancement question.
Most databases (e.g. Oracle, SQL*Server, Sybase, ... ) send back a description of the resultset, even if there are zero rows. I cannot seem to find this in pqxx, nor can I find a way to describe SQL of any kind prior to execution. The only way I have been able to get the SQL metadata is by collection upon the first row retrieved, but this is limited to an actual row being retrieved. If there are zero rows, I have no way to determine the metadata.
Looking at the v15.x code base for libpq (standard C library), I see that psql has a "workaround" for this very problem. In the file src/bin/psql/common.c there is a function called static bool DescribeQuery(const char *query, double *elapsed_msec) (on or about line #1248).
This function has a rather ingenious solution for this very problem. They create an "unnamed" prepared statement, which they do not execute, instead they use the results of that call, which provides the field info, and subsequently query against the pg_catalog for the metadata descriptions (source below).
Unfortunately, the pqxx library doesn't provide enough (that I can tell) to duplicate this functionality. Does anyone have a solution for this problem?
Credit for resolution goes to Jeroen Vermeulen, on the Github issue #641. He suggested that the predicate AND false be added to the query to force a prototype result. By doing this an empty resultset is returned, with the fields available for metadata collection.

How to install feature based on the property set in custom action?

I am trying to install one from two features based on the value that should be set inside of the custom action.
Firstly, I set the value of a property:
UINT __stdcall ConfigurationCheckAction(MSIHANDLE hInstall)
{
HRESULT hr = S_OK;
UINT er = ERROR_INSTALL_FAILURE;
hr = WcaInitialize(hInstall, "ConfigurationCheckAction");
if (condition) {
MsiSetProperty( hInstall, TEXT("STREAM"), TEXT("RED") );
}
else {
MsiSetProperty( hInstall, TEXT("STREAM"), TEXT("BLUE") );
}
return WcaFinalize(er);
}
Secondly, I make two conditions per two features:
<Feature Id='Complete' Level='1'>
<Feature Id="Red" ConfigurableDirectory="TARGETDIR" Title="F1" Level="0">
<Condition Level="1">STREAM</Condition>
</Feature>
<Feature Id="Blue" ConfigurableDirectory="TARGETDIR" Title="F2" Level="0">
<Condition Level="1">NOT STREAM</Condition>
</Feature>
</Feature>
Note that I don't define property inside of the wxs file previously, as I would like to set it from the custom action.
My custom action is called before InstallInitialize and Execute is immediate.
From the installation log I have confirmation that the property is set.
However, my conditional installation does not work, as it seems like what is in the condition is always evaluated as false.
I tried evaluating conditions:
STREAM, STREAM=RED, STREAM="RED", < ![CDATA[STREAM=RED]]>
I am running out of ideas what to do and would appreciate help.
Too late to test all of this, but here goes with some information. I will check back tomorrow. Essentially I think the problem is your custom action sequencing. Try before Costing.
Some things to consider:
Custom action sequencing: you need to sequence your custom action right and it needs to be present in both silent and interactive installation modes.
Did you try to sequence the set property custom action before CostInitialize? You state you set it before InstallInitialize, but try it before CostInitialize instead (you might have tried).
And did you remember to insert this custom action in the InstallUISequence as well as the InstallExecuteSequence? You need to insert in both sequences in case the setup runs in silent mode. Before CostInitialize in both sequences I believe.
Feature Level: manipulating features via the feature level and INSTALLLEVEL is just one way to do feature control, you can also set features via the command line or using a custom action.
Setting a feature level to 0 should hide the feature from view in the setup's custom dialog.
Setting a feature level higher than the setup's INSTALLLEVEL will deselect the feature from installation.
And the other way around setting a feature level lower or equal to the setup's INSTALLLEVEL will select the feature for installation.
The conditional syntax allowed is quite flexible, and could provide the functionality you need outright - but I have never used them properly. Here is an example from the Installshield forum.
ADDLOCAL & REMOVE: you can manipulate the feature selection by changing the values of the ADDLOCAL and REMOVE properties from a custom action (technically also REINSTALL and ADVERTISE) - and these properties can be set via the command line as well.
Win32: you can also use the Win32 functions MsiGetFeatureState and MsiSetFeatureState - from a C++ custom action - to set feature selection.
Frankly it is a bit mad the whole thing. Also keep in mind that there are feature action states (what is going to happen to a feature) and feature installed states (what state it is in). The Win32 function documentation should explain.
Cross-linking for easy retrieval:
Unselected Feature Being Installed
I have done something similar, but we ended up controlling this at a component level(adding the condition to the <Component/> elements instead of the feature element using a transform during heat). But our condition utilizes CDATA while also using double quotes for the value, which you don't list in what you've tried. So first I'd try the following conditions in your features:
<Condition><![CDATA[STREAM="RED"]]></Condition>
<Condition><![CDATA[STREAM="BLUE"]]></Condition>
If that still does not work, I would try the following:
Add the STREAM property with a default value to your WiX. Then test it with that default value to see if having the default value set to begin with makes it work. That could mean you need to set the property sooner, possibly off a UI event. <Property Id="STREAM" Value="RED"/>
As a last resort, you could add the conditions to each component as I did, but we only did that for very specific reasons, hopefully you can get the conditional feature to work with the above suggestions!
I hope the above fixes your problem, or at least leads you to the answer!
Thank you for your replies. In the end, a combination of your suggestions helped me.
I want to state what did and what did not work:
Adding property to WiX with a default value was not necessary (as well with adding property of this property Secure='yes')
Calling custom action before CostInitialize did not solve the problem on its own, but I believe it was one of the factors that resolved an issue.
Conditional sintax was corrected by:
a) Putting condition inside of CDATA and adding quotes to the value of property as suggested: <Condition><![CDATA[STREAM="RED"]]></Condition>
b) Reversing condition levels so feature has condition level 1 and condition has level 0. This means that feature is always installed, unless the condition expression is false.
Concerning the correct order of the custom actions, the description of the custom action type 51 contains the decisive hint:
"To affect a property used in a condition on a component or feature, the custom action must come before the CostFinalize action in the action sequence."

What is the difference between kubectl apply and kubectl replace

I am learning Kubernetes recently, and I am not very clear about the difference between "kubectl apply" and "kubectl replace". Is there any situation that we can only use one of them?
I have written up a thorough explanation of the differences between apply, replace, and patch: Kubernetes Apply vs. Replace vs. Patch. It includes an explanation that the current top-ranked answer to this question is wrong.
Briefly, kubectl apply uses the provided spec to create a resource if it does not exist and update, i.e., patch, it if it does. The spec provided to apply need only contain the required parts of a spec, when creating a resource the API will use defaults for the rest and when updating a resource it will use its current values.
The kubectl replace completely replaces the existing resource with the one defined by the provided spec. replace wants a complete spec as input, including read-only properties supplied by the API like .metadata.resourceVersion, .spec.nodeName for pods, .spec.clusterIP for services, and .secrets for service accounts. kubectl has some internal tricks to help you get that right, but typically the use case for replace is getting a resource spec, changing a property, and then using that changed, complete spec to replace the existing resource.
The kubectl replace command has a --force option which actually does not use the replace, i.e., PUT, API endpoint. It forcibly deletes (DELETE) and then recreates, (POST) the resource using the provided spec.
Updated Answer
My original was rather controversial and I would even say now, in hindsight, half incorrect. So here is an updated answer which I hope will be more helpful:
commands like kubectl patch, replace, delete, create, even edit are all imperative: they tell kubectl exactly what to do
the kubectl apply command is OTOH "declarative" in that it tells kubernetes, here is a desired state (the yaml from the file provided to the apply command), now figure out how to get there: create, patch, replace the object, etc whatever it takes... you get the idea.
So the 2 commands are hugely different.
EG with apply you can give it just the changes you want: it will figure out what properties of the object need to be changed, and leave the other ones alone; if those properties are "immutable" (eg, the nodeName of a pod), it will complain, and if you then repeat the command with --force, it is smart enough to know to do the equivalent of a replace --force.
In general, you should favor apply (with --force when necessary), and only use the imperative commands when the declarative approach does not give the expected result (although I would love to see examples of this -- I'm guessing this would happen only when you would need several steps because of interdependencies that will have negative consequences if done with apply).
The difference between apply and replace is similar to the difference between apply and create.
create / replace uses the imperative approach, while apply uses the declarative approach.
If you used create to create the resource, then use replace to update it. If you used apply to create the resource, then use apply to update it.
Note that both replace and apply require a complete spec, and both create the new resources first before deleting the old ones (unless --force is specified).
you can add option -v=8 when use kubectl, and you will find the log like this
apply --force
patch 422
delete 200
get 200
get 200
get 404
post 201
replace --force
get 200
delete 200
get 404
post 201
kubectl apply .. will use various heuristics to selectively update the values specified within the resource.
kubectl replace ... will replace / overwrite the entire object with the values specified. This should be preferred as you're avoiding the complexity of the selective heuristic update. However some resources like ingresses/load balancers can't really be replaced as they're immutable.
Example of the heuristic update leading to non obvious operation: https://github.com/kubernetes/kubernetes/issues/67135
From: https://github.com/kubernetes/website/blob/master/content/en/docs/concepts/cluster-administration/manage-deployment.md
Disruptive updates
In some cases, you may need to update resource fields that cannot be
updated once initialized, or you may just want to make a recursive
change immediately, such as to fix broken pods created by a
Deployment. To change such fields, use replace --force, which deletes
and re-creates the resource.

Clojure - component and names

I am trying to understand Stuart Sierra's component, specifically the naming convention for the components in order to structure a Clojure application.
If I look into system for instance, I see several components mapped to the :server key :
aleph
immutant
Since both use the same key :server, does that mean that I can only use one of them if I use this library ?
Similarly, I use onyx. Several components are already defined inside onyx system.clj.
Does that mean that some keys are effectively reserved by onyx ?
What will happen to the :port parameter, which seems to be used by many components in the wild ?
Questions
What is get the difference between the keys used when associng in the start method and the keys used in component/system-map ?
Is there a naming convention for those keys, how do we avoid collisions between those keys ?
In which cases (if any) does it make sense to have several systems and can they run at the same time ?
Keys in the system map identify specific components (instances) in that system. You can use whatever key you like for whatever component you need.
Keys in a specific component record can be one of three things:
a configuration value set up at creation time
some internal value that is irrelevant to the user of the component
a dependency (which will refer to another component when the system is started)
1 and 2 are generally set up by the component constructor and users do not need care about the actual key used in the record.
Dependencies are configured by setting a mapping on the depending component from the dependency key (3) to the key in the system map referring to the dependancy component. This is done with the component/using function and giving it a map of component keys to system-map keys as the second argument. That way you can always map any expected key to any actually used key. You can use the short-hand form of component/using with a vector of keys, but only if the keys in the system-map are the same as the keys in the component you're configuring.
I hope that answers the first two questions
The third question I think I'd like to see an example of what you're looking for as a separate post
The last question: yes you can have multiple systems running at the same time. That may or may not make sense depending on what you want to do, but running a test system as well as a development system seems like a fairly obvious setup.

REST services - exposing non-data "actions"

I understand how to use REST for doing general entity interactions - using urls names to map to entities and the HTTP verbs to map to actions on those entities. But what is the generally accepted way of looking at "actions" more like RPC?
For example, let's say I want to send a command for the device to reset? There's no real "entity" here or do I do something like POST to http://mydevice/device/reset?
/device/reset or /system/reset are ok.
The REST "design pattern" does encourage you to NOT use any verbs.. You could do:
POST http://mydevice/system/state
<stateType>RESET</stateType>
Related information:
How to create REST URL’s without verbs?
Threads tagged with restful-url
I don't think that's the case to use POST. The "RESET action" is a idempotent action (if you call it n times you will always get the same result), so IMHO you should use a PUT call instead POST (as POST is not idempotent).
Also, as you are Putting a resource, you can use
PUT http://system
<device>
<status>RESET</status>
</device>
or
PUT http://system/status/reset
But I think the first one is "more restful", since you are putting a resource, while the second one you just use the URL.
I usually name the entity "system" or something like that. So you do "/system/reset". You've chosen device so that works too.
But yea, I usually consider these types of actions to be updates, which would use the POST method. So I think you are right to POST to /device/reset