perf: skip GameDataParser if there are no subscribers

[alrun3]

No description provided.

[TheLazzoro]

OnGameDataBlock is always subscribed to through W3GReplay.cs contructor, so even though it's nullable it's actually never the case. Is there a particular goal in mind?

[alrun3]

OnGameDataBlock is always subscribed to through W3GReplay.cs contructor, so even though it's nullable it's actually never the case

My bad I use LINQPad and just #load in the classes, which makes internal members accessible, so I completely overlooked the visibility:

#load ".\W3G-NET\W3GNET\Parsers\*.cs"
#load ".\W3G-NET\W3GNET\Extensions\*.cs"
#load ".\W3G-NET\W3GNET\Types\*.cs"
#load ".\W3G-NET\W3GNET\Util\*.cs"
#load ".\W3G-NET\W3GNET\W3GReplay.cs"

using W3GNET.Parsers;
using System.Linq;

async Task Main()
{
    var parser = new ReplayParser();

Is there a particular goal in mind?

If you're just searching through your replay directory for specific metadata, skipping the gamedata parsing significantly improves performance. Eg.:

new ReplayBrowser(UserProfile, Id)
    .EnumerateReplays()
    .Where(r => r.MapName.Contains("<some map name>"))
    .Select(r => r.FileInfo)

[TheLazzoro]

If you're just searching through your replay directory for specific metadata, skipping the gamedata parsing significantly improves performance.

Ah yeah right now it just parses everything. If the goal here is to improve performance for this type of scenario I would recommend adding parameters so you can explicitly tell the parser to skip certain parts.

[alrun3]

If you're just searching through your replay directory for specific metadata, skipping the gamedata parsing significantly improves performance.

Ah yeah right now it just parses everything.

I just took another look, and since the class is part of the public API, the change still seems valid.

If the goal here is to improve performance for this type of scenario I would recommend adding parameters so you can explicitly tell the parser to skip certain parts.

I understand the suggestion. However, relying on both configuration and subscriber presence seems redundant. Emitting data only when there are subscribers helps keep the API surface smaller and avoids potential mismatches between configuration and actual usage