Re: Replacement for ConnectionSocketFactory

2025-04-11 Thread Ryan Schmitt
DefaultHttpClientConnectionOperator is marked @Internal, as is the associated connection manager constructor that takes an instance of it. I'll take a look at your other suggestions. On Fri, Apr 11, 2025 at 4:34 AM Oleg Kalnichevski wrote: > > On Thu, 2025-04-10 at 19:02 -0700, Ryan Schmitt wrote

Re: Replacement for ConnectionSocketFactory

2025-04-11 Thread Ryan Schmitt
Marking incubating interfaces as @Internal as fair. I sometimes use @Deprecated, since the call sites generate warnings that way, but the @Internal convention is good to know. On Fri, Apr 11, 2025 at 3:00 PM Ryan Schmitt wrote: > > DefaultHttpClientConnectionOperator is marked @Internal, as is th

Re: Gradle migration

2025-04-11 Thread Oleg Kalnichevski
I am comfortable with both Maven and Gradle as build tools. I like them both, though I hate Maven Site Plugin with a burning hatred. Maven has served us well so far. We also have Maven project members on board, which helps. I personally see no good reason to migrate. Oleg On Fri, 2025-04-11 at

Re: [PR] Enhance ProtocolSwitchStrategy with Simplified Protocol Parsing [httpcomponents-client]

2025-04-11 Thread via GitHub
arturobernalg commented on PR #627: URL: https://github.com/apache/httpcomponents-client/pull/627#issuecomment-2797696594 > @arturobernalg I took a closer look at your code and had to refactor a few things. Please take a look at [0ba6102](https://github.com/apache/httpcomponents-client/com

Re: Gradle migration

2025-04-11 Thread Ryan Schmitt
Heh, I didn't think so. On Fri, Apr 11, 2025 at 9:32 AM Gary Gregory wrote: > > Please PLEASE no. > > Gary > > On Fri, Apr 11, 2025, 12:30 Ryan Schmitt wrote: > > > Are we fine with staying on Maven, or would there be any interest in > > moving to Gradle? Please note that I am _not_ trying to st

Re: Regression in HTTP/2 cleartext

2025-04-11 Thread Gary Gregory
Yes, sorry about that confusing bit, I meant JApiCmp, not JaCoCo! Gary On Fri, Apr 11, 2025, 12:28 Ryan Schmitt wrote: > I see. The tool I've actually used for this in the past is japicmp. > > On Tue, Apr 8, 2025 at 2:07 PM Gary Gregory > wrote: > > > > Hi Ryan, > > > > Apache RAT checks sourc

Re: Gradle migration

2025-04-11 Thread Gary Gregory
Please PLEASE no. Gary On Fri, Apr 11, 2025, 12:30 Ryan Schmitt wrote: > Are we fine with staying on Maven, or would there be any interest in > moving to Gradle? Please note that I am _not_ trying to start a debate > about which build tool is better, or in persuading anyone that a > migration i

Gradle migration

2025-04-11 Thread Ryan Schmitt
Are we fine with staying on Maven, or would there be any interest in moving to Gradle? Please note that I am _not_ trying to start a debate about which build tool is better, or in persuading anyone that a migration is worthwhile. I'm just offering to do the work of migrating to Gradle if that's som

Re: Regression in HTTP/2 cleartext

2025-04-11 Thread Ryan Schmitt
I see. The tool I've actually used for this in the past is japicmp. On Tue, Apr 8, 2025 at 2:07 PM Gary Gregory wrote: > > Hi Ryan, > > Apache RAT checks source file licensing. You might be thinking of Clirr > (doesn't work well on Java 8), JaCoCo, or RevApi. > > Gary > > On Tue, Apr 8, 2025, 16:

Re: Replacement for ConnectionSocketFactory

2025-04-11 Thread Oleg Kalnichevski
On Thu, 2025-04-10 at 19:02 -0700, Ryan Schmitt wrote: > 2. Connection count metrics. This is a trivial CSF implementation > that > wraps a delegate and emits a metric whenever connectSocket() is > called. > 3. Unix domain socket support. This is a CSF implementation that > calls > JUnixSocket to c