Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Uuid Changes #2555

Merged
merged 7 commits into from
Nov 3, 2022
Merged

Uuid Changes #2555

merged 7 commits into from
Nov 3, 2022

Conversation

shaopeng-gh
Copy link
Collaborator

  • BREAKING: For Guid properties defined in SARIF spec, updated Json schema to use uuid, and updated C# object model to use Guid? instead of string.
  1. This is based on the updated JSchema Code v2.1.0 that supported generate nullable Guid?.
  2. NOTE: Had to update the $schema from draft-04 to latest 2020-12 which has the draw back of VS not able to auto-compete. The Uuid introduced in 2019-09, however since VS also not able to auto-compete in that version, if we are updating might as well update to the latest version.

if (compareResult != 0)
{
return compareResult;
}
Copy link
Collaborator Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

  1. I found a bug here, if both left and right are null, TryReferenceCompares will directly return true and all the other properties are not compared down below.

  2. The fix by adding CompareTonot only fix this bug but also able to remove this file entirely.
    #Closed

public bool ShouldSerializeId()
{
return Id >= 0;
}
Copy link
Collaborator Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

The better place for this is to put in core folder, this is not related to this change, moved btw. #ByDesign

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Yes! But see my other comment. I wondered why we preserved this should serialize id if we have 'emit default value' == false.

One reason is that someone could set this number to a garbage value like -37, which wouldn't be removed by the serialization attribute.

If that's true, though, we don't really need the default value and emit default value attributes.

But perhaps we do, because someone could emit serialized json where the -1 is explicit (that's legal!) but we don't want to populate in this case.

OK, so talking it all over, I agree with your implementation. :) We need to preserve this file as it is and retain (and move) ShouldSerializeId().

Thanks for listening/. :)

Copy link
Collaborator Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

why we preserved this should serialize id if we have 'emit default value' == false.
------ I think we need both because
ShouldSerializeXXX is from Newtonsoft.Json and used in Newtonsoft.Json.JsonConvert
EmitDefaultValue in the DataMember is used in dotnet e.g. DataContractSerializer
I tried removing either one of them will make the corresponding one populate id:-1 value in the file.

Copy link

@github-advanced-security github-advanced-security bot left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

CodeQL found more than 10 potential problems in the proposed changes. Check the Files changed tab for more details.

/// <returns>
/// Returns 1 if left is greater than right, -1 if left is less than right, 0 if they are equal.
/// </returns>
internal static int CompareTo<T>(this T? left, T? right) where T : struct, IComparable
Copy link
Member

@michaelcfanning michaelcfanning Oct 17, 2022

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Did we just implement Nullable.Compare here? :)

I think so. If I'm correct, let's zap this helper and use the C# built-in.

https://learn.microsoft.com/en-us/dotnet/api/system.nullable.compare?view=net-6.0 #Resolved

Copy link
Collaborator Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Good catch. I have just updated the extension method to use the C# built-in. Let me close this one and reply on the other so no need to look duplicated

if (compareResult != 0)
{
return compareResult;
}

compareResult = string.Compare(left.RunGuid, right.RunGuid);
compareResult = left.RunGuid.CompareTo(right.RunGuid);
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

left.RunGuid.CompareTo(right.RunGuid)

This is where we would prefer Nullable.Compare(left, right), if possible. Would this require an update to JSchema?

Copy link
Collaborator Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Good catch. I have first updated my extension method to utilize the C# built-in method.

Nullable.Compare(left, right) will need to update the JSchema, it will need to be able to identify if a type is nullable, also struct not ref type and generate different code:
from:
left.RunGuid.CompareTo(right.RunGuid);
to:
Nullable.Compare(left.RunGuid., right.RunGuid)

also all those other nullable in the notyetautogenerated folder, in the JSchema we will not be able to know they are going to be replaced manually and we will read them as not nullable there.
By adding an extension method it works with above. e.g. that file that I removed AddressComparer.cs.

let me know your instruction if to keep the extension method.

@@ -45,11 +45,6 @@ public virtual SarifNodeKind SarifNodeKind
[JsonProperty(DefaultValueHandling = DefaultValueHandling.IgnoreAndPopulate)]
public virtual BigInteger Id { get; set; }

public bool ShouldSerializeId()
Copy link
Member

@michaelcfanning michaelcfanning Oct 17, 2022

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

ShouldSerializeId

Why did this get deleted? Is it because we specify -1 as a default value and 'emit default value' is false, so this is redundant?

Can you explain how this code got left behind, if so? Did we take an autogenerator fix? #Closed

Copy link
Collaborator Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

closing this, replied in the other one.


/// <summary>
/// An array of unique identifies in the form of a GUID by which this report was known in some previous version of the analysis tool.
/// </summary>
[DataMember(Name = "deprecatedGuids", IsRequired = false, EmitDefaultValue = false, Order = 11)]
public virtual IList<string> DeprecatedGuids { get; set; }
public virtual IList<Guid> DeprecatedGuids { get; set; }
Copy link
Member

@michaelcfanning michaelcfanning Oct 17, 2022

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

DeprecatedGuids

Just curious, do we have an explicit test that populates, serializes this list of Guids? I think we do. It just occurs to me this is a pretty specialized code path/utilization, these list of struct Guids... #WontFix

Copy link
Collaborator Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

yes we did have, and I made those test data from string to Guid in the list.

@@ -235,10 +235,7 @@ public int GetHashCode(ReportingDescriptor obj)
foreach (var value_5 in obj.DeprecatedGuids)
{
result = result * 31;
if (value_5 != null)
Copy link
Member

@michaelcfanning michaelcfanning Oct 17, 2022

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

null

Wouldn't the right test here be

if (value_5 != default)
OR
if (value_5 != Guid.Empty)

or do we require/benefit from processing the hashcode for these empty items?

Note that an empty guid is not a real guid. OTOH, we should regard two lists as unique if the only difference between them is that one contains Guid.Empty and the other does not, right?

So maybe there is no change here...
#ByDesign

Copy link
Collaborator Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I thought about this a well,
this change is auto-generated not hand made, I believe this is correct because deprecatedGuids is a list of valid GUID and null is not a valid GUID so it can not be in the list, so each item in the list is not nullable and then the null check is not needed. So the DeprecatedGuids is defined as C# List<Guid> without the question mark, let me know.
the definistion says "an array of zero or more unique (§3.7.3) GUID-valued strings (§3.5.3)":
A reportingDescriptor object MAY contain a property named deprecatedGuids whose value is an array of zero or more unique (§3.7.3) GUID-valued strings (§3.5.3) each of which was used by a previous version of the tool as the value of the guid property (§3.49.5) for this object.

@@ -22,5 +22,10 @@ public LogicalLocation LogicalLocation
}
}
}

public bool ShouldSerializeId()
Copy link
Member

@michaelcfanning michaelcfanning Oct 17, 2022

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

ShouldSerializeId

Ah! you moved this out of the 'NotYetAutogenerated' directory, now I understand. But isn't it true that this isn't required because of the serialization attributes set on the Location 'Id' property? Or am I mistaken?

If I'm correct, let's spend a little time looking for similar errors, i.e., other unnecessary 'Should' serialization methods... #Closed

Copy link
Collaborator Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I believe it is needed, please see the comment on the place I removed it.

@@ -12,8 +12,8 @@ public partial class RunAutomationDetails

public bool ShouldSerializeId() => !string.IsNullOrWhiteSpace(this.Id);

public bool ShouldSerializeGuid() => !string.IsNullOrWhiteSpace(this.Guid);
public bool ShouldSerializeGuid() => this.Guid != null;
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

this.Guid

these comparisons to null are fine, but we could also use this.Guid.HasValue, yes? #Resolved

Copy link
Collaborator Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

yes I actually thought about this as well, I believe most people prefers .HasValue,
the only reason I was thinking about consistency with how string compares with null.

Let me know your thoughts?

Below are some random points in internet:

"The compiler replaces null comparisons with a call to HasValue, so there is no real difference. Just do whichever is more readable/makes more sense to you and your colleagues."

"It's pretty silly to complain about being able to set a Nullable to null or compare it to null given that's called Nullable"

"I prefer (a != null) so that the syntax matches reference types."

using System.CodeDom.Compiler;
using System.Collections.Generic;

namespace Microsoft.CodeAnalysis.Sarif
Copy link
Member

@michaelcfanning michaelcfanning Oct 17, 2022

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

CodeAnalysis

Excellent to see this code go! #Closed

@@ -1,5 +1,5 @@
{
"$schema": "http://json-schema.org/draft-04/schema#",
"$schema": "https://json-schema.org/draft/2020-12/schema",
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

$schema

Sorry for the late-breaking comment! You need to produce an rtm.6 version of this file, we can't modify rtm.5, it's an artifact of the standards process. Ask Chris Meyer for information about how to update everything to use a new flavor of schema.

Sorry again for the delay giving you this advice.

@michaelcfanning
Copy link
Member

"id": "https://raw.githubusercontent.com/schemastore/schemastore/master/src/schemas/json/sarif-external-property-file-2.1.0-rtm.5.json",

This is why you can't just update these files, we need to publish new versions in the schema store.


Refers to: src/Sarif/Schemata/sarif-external-property-file-2.1.0-rtm.5.json:4 in ebaff74. [](commit_id = ebaff74, deletion_comment = False)

@@ -59,7 +59,9 @@ public static IEnumerable<Run> EnumerateRuns(this SarifLog log)

public static Result FindByGuid(this SarifLog log, string guid)
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

FindByGuid

Should we simply have updated this API to accept a GUID argument?? Something to keep in mind moving forward.

Copy link
Member

@michaelcfanning michaelcfanning left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

:shipit:

@michaelcfanning michaelcfanning merged commit c0121cb into main Nov 3, 2022
@michaelcfanning michaelcfanning deleted the users/shaopeng-gh/useuuidpr branch November 3, 2022 19:33
@michaelcfanning
Copy link
Member

@shaopeng-gh, what a fantastic (and enormous!) change. Please see my notes, we need to revert your changes to the rtm.5 schema and create an rtm.6. I didn't want to delay getting this in, we'll fix things up as a follow-on change. @EasyRhinoMSFT is familiar with this process (and actually may be willing to simply take on the work).

@shaopeng-gh
Copy link
Collaborator Author

Thanks for the review!
@EasyRhinoMSFT Talked to Michael I will create a clean PR that reverts RTM.5 and adds RTM.6 (in progress version). We won't publish RTM.6 to schemastore.org or elsewhere until finalize it with the next PR with Big Integer changes.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

None yet

2 participants