Benefits of using BufferedReader?

[Neil Stalker]

Hi,

I have been testing gitsharp on linux and trying to clear up some of
the test failures. Ive tracked around 30 failures to the use of
BufferedReader objects, several of which went into infinite loops. The
cause is that the ReadLine method is not implemented in BufferedReader
so it is using the method from TextReader which always returns an
empty string.

What benefits does BufferedReader have over StreamReader? When i have
changed the readers over i am not seeing any regresions in the test
suite.

Neil

[Miguel de Icaza]

When I looked at the code last year, I got the impression that a lot of that code in the Java implementation did not have a very useful purpose to begin with.

That entire cache system (including that "Window"  concept) can probably be removed and a proper one implemented later when we better understand how GIT# actually interacts with the system, and when actual profiling is done on how .NET and GIT interact with each other.

[Meinrad Recheis]

As it has no impact on the tests, it seems to be ok. We have not yet a performance testing suite, so we cannot judge the impact on performance from this change. I pulled in your contributions, performance is not the focus now ... functional "completeness" on mono and .net is the current goal.
-- henon

[Neil Stalker]

I had the feeling it had just been pulled over from java, but i wanted
to make sure that there wouldnt be any issues in removing it.

Neil

On Sep 4, 6:41 am, Meinrad Recheis <meinrad.rech...@gmail.com> wrote:

> -- henon- Hide quoted text -
>
> - Show quoted text -