| source code repo? [message #5330] |
Sat, 16 December 2017 13:58  |
|
Is there an official source code repository somewhere?
I only found a project on sourceforge, but it didn't seem to have a source code repository.
The libretro guys seem to track the releases on https://github.com/libretro-mirrors/mednafen-git and this project seems to have the website at https://github.com/mednafen/mednafen.github.io . And then there are a ton of forks.
But where is the official source repo?
|
|
| Re: source code repo? [message #5331 is a reply to message #5330 ] |
Sun, 17 December 2017 23:28  |
|
|
There isn't one.
|
|
| Re: source code repo? [message #5332 is a reply to message #5331 ] |
Mon, 18 December 2017 03:53 |
|
|
What's the reason?
|
|
| Re: source code repo? [message #5567 is a reply to message #5331 ] |
Thu, 21 June 2018 15:02 |
|
i wish to track the changes beetween 0.9.36.5 and 0.9.37 to disable stupid usage of C++11 features due now the medanfen in 0.9.38 are very slower rather than 0.9.36, i able to compile the 0.9.36 with gcc 4.3 and the speed are better rather thant compiled with gcc 6 or gcc 7 std libs
due there's no repository or a source history change organization file, i cannot made this easyle, or without C++ skills, its a mess
|
|
| Re: source code repo? [message #5571 is a reply to message #5567 ] |
Thu, 21 June 2018 17:19 |
|
|
Without C++ skills, what leads you to believe that having a public repository would make what you want to do any easier? It's not like every commit would be made with "This code uses C++11 features in <place a> and <place b>, do <something> to convert it to C++03."
|
|
| Re: source code repo? [message #5579 is a reply to message #5571 ] |
Fri, 22 June 2018 07:33 |
|
with a repository we can TRACK the CHANGES, so we can see the chages beetween July 21 and July 17, make a revert and take the result to improve the newer coce
of course seems do you live in the past and does not take advantages of that simple operation!?
please administrator! do not thing that others do not able to make important changes! i do not have C++ or C skills but i enought smart to make changes and improve the performance based on the changes that you made..
i'll explaint to you: due you have enought resources do you not nothed the waste of them! but i have enouoght machines to see it due i used mednafen (and others programs) to provide OS to many childrens.. currently mednafen runs with very poor performance in version 1.2X or lasted 0.9.X
also i see the section https://mednafen.github.io/documentation/#Section_lag sabout laggin. .puff all are guindows only and waste .. laggin comes with newer lines to computing.. also runs mednafen in a win10 will be more difficult rather runs mednafen in win95 due that lasted does not have so many "taks to compute".. so the solution that i made its change the code base and does not use so many lines of code that need more computation... with touched results
|
|
| Re: source code repo? [message #5581 is a reply to message #5579 ] |
Fri, 22 June 2018 08:40 |
|
Code is nuanced and intertwined, you cannot simply piecemeal revert changes and expect things to not break, in subtle or not so subtle ways, especially if you don't really know what you're doing.
In any case, it's highly unlikely that C++11 usage in and of itself is causing performance issues, and using a really old version of gcc is not recommended, as they tend to have unfixed bugs. If gcc 6.x and 7.x are producing slower than ideal binaries, then I'd suggest trying gcc 4.9.
|
|
| Re: source code repo? [message #5583 is a reply to message #5581 ] |
Fri, 22 June 2018 10:00 |
|
before answer that, a very importat note: compiled first versions of 0.9.X mednafen again a gcc 3.4 produces a very high optimized binary, of course i changed some pieces of code..
| Quote: |
Code is nuanced and intertwined, you cannot simply piecemeal revert changes and expect things to not break,
|
yeah of course, but of course i only revert pieces of codes (if a repository will be available of course)
| Quote: |
In any case, it's highly unlikely that C++11 usage in and of itself is causing performance issues
|
some newer features also raises the current cpu requerimeints usage, do you note it?
in any case, i only wants to avoid those C++11 parts in the current code base to be able to build agains gcc 4.3 and libc 3.2.3 in my embebed linux on embebed machines for childrens (wil be ilogical spent a core 2 duo to play games of snes and psx, in some countries its huge the prices)
| Quote: |
using a really old version of gcc is not recommended, as they tend to have unfixed bugs.
|
that's do not affect the binary produced and a machine used to emulated games are not connected to network and do not use to any other task! let's thing about it! if i have a machine for work, install mednafen its not my priority!
in build's case: i have a well patched gcc 4.3 version, very well patched, i track all the red-hat patches to my debian packages, no problems at today!
| Quote: |
If gcc 6.x and 7.x are producing slower than ideal binaries, then I'd suggest trying gcc 4.9.
|
still produces slower binary and the ideal binary that are slower too, i made many tests and changes in mednafen history by my own and concluded that best are with gcc 3.4 but due are too old i used now 4.3... as i said i only wants to avoid the C++11 parts and be able to compiled and use agains stdlibs 4.3 and libc 3.2.3
mednafen 0.9.36.5 runs but i cannot track all the changes to 0.9.48 in that code base.. due there's too many beetween each release...
[Updated on: Fri, 22 June 2018 10:33]
|
|
Members 
Search 
Help 
Register 
Login 
Home 
]