The Sinclair QL Forum

ql_freak wrote:BTW: Is there a chance to port git to QDOS?

I would say "highly unlikely" given the need for network access to, say, a GitHub repository.

However, it is written in C and source is available at https://github.com/git/git - bit I imagine compiling and linking might blow out available RAM on non-emulators.

Cheers,
Norm.

You can have local git repositories though and I'm pretty sure local operations don't use a local network call (when I say pretty sure, what I mean of course is I have no idea!)

Sparrowhawk wrote:You can have local git repositories though and I'm pretty sure local operations don't use a local network call (when I say pretty sure, what I mean of course is I have no idea!)

Git, from what I can find online, in manuals etc, doesn't appear to use any networky stuff for local repositories. But don't quote me on that!


Cheers,
Norm.

Hi,

It is little confusing the Git software devloped by Linus Torvalds, who as you know developed a free operating system and Github is hosted by an American company, whose parent company is Microsoft, who developed a closed operating system.

It is a little ironic that Microsoft use free and open software platform to develop software prohects, when their own closed operating system and other commerical applications are closed source.

I must admit to not seeing all this.

Does Microsoft cherry pick all the good source on their Github platform and add it into ther closed source platform.

Hi Derek,

Derek_Stewart wrote:Does Microsoft cherry pick all the good source on their Github platform and add it into ther closed source platform.

Nope. Much as I'm against Microsoft for other reasons, some personal, and their dubious marketing practices, they are not allowed to cherry pick Open Source content from GitHub, unless, they make publically available their changes. This obvioulsy depends on the original licence in use by the code that has been cherry picked. If it's under the GPL, then Microsoft are required to publish the code under the GPL. If, for example, it's under the MIT licence, then they can use it in commercial packages without attribution or publishing changes.

They are coming around to understanding how Open Source (and Linux in general) is a good thing and have embraced it quite well. They are still dubios though - I could use their VSCode editor on Linux, it's open source and free as well as being cross platform, but I choose to use VSCodium instead. That is simply the VSCode source code, recompiled with all the "phone home to Microsoft" stuff removed.

So, in summary, Microsoft can cherry pick code from GitHub, as indeed can anyone else, but they must adhere to the terms of the licence for the code they choose to use.


Cheers,
Norm.

Sparrowhawk wrote:You can have local git repositories though and I'm pretty sure local operations don't use a local network call (when I say pretty sure, what I mean of course is I have no idea!)

That's what I have meant!

git is an horrible program (written by Linus Torvalds, a former QL user) - REALLY HORRIBLE; I MUST USE IT :-|

B_u_t___i_t___i_s___t h e___S_t_a_n_d_a_r_d_! (now unfortunately –– I think [don't know it really] Mercury was better)

ql_freak wrote:B_u_t___i_t___i_s___t h e___S_t_a_n_d_a_r_d_! (now unfortunately –– I think [don't know it really] Mercury was better)

I think Mercurial is a thousand times better I'm really annoyed that GIT has become the standard.

mk79 wrote:I think Mercurial is a thousand times better :( I'm really annoyed that GIT has become the standard.

THANK YOU for confirmation and correction of the "hg" program name!

Hi,

I have installed RCS and GnuDiff on SMSQ/E. rcslog, ci & co work as advertised but I just can't get rcsdiff to work - I get "rcsdiff error: diff failed" error and an output of diff showing the available command line options when running it. I have verified that diff works correctly when run normally from command line. I have a suspicion this might have something to do with what rcsdiff does behind the scenes. I think it might temporarily copy the checked-in version (in e.g. README,v) as a file called (e.g.) "-r1.1" and compare that to the checked out version (e.g. README). So we end up in a situation where rcsdiff is running "diff -r1.1 README" and the first filename of "-r1.1" is interpreted as a command line argument by diff and that causes the failure.

I tried both having a separate RCS subdirectory for the *,v files, and keeping them in the same directory as the source files. Also tried mapping dev1_ to the win1_subdirectory_ I was using.

Has anyone run rcsdiff successfully? Is there some configuration I might be missing?

Cheers,
Petri

I tried RCS many years ago and never got it working I'm afraid.

Cheers,
Norm.