Good Afternoon All.

I'm attempting to fork the FPC Source repository on GitLab so that I can occasionally make some updates to the M68K Sinclair QL cross compiler RTL and Unit files. However, I get an error messages saying "an error occurred, try again". This is constant with every attempt, after fresh logins etc.

I suspect that this is down to the repository being private? If indeed it is private. There are docs on GitLab which advises that private repos can't be cloned unless a member.

Am I correct in this thinking? Is all I need to do is request access to the official repo using the link on the home page for it?

I have checked the FPC development docs, Wiki and searched with Google/DuckDuckGo to no avail.

Any advice gratefully reeceived. Thanks.

The repo is not private.

One possible reason is that the repo is simply too big: it has 26 years of commits.

We had serious problems importing it for this very reason.

Michael.

Hi,

I had no trouble forking FPC sources on GitLab in the past, and now I tested it again 2 times -- I can fork them without any issues (and my fork is even private, that's OK too). After less than 1 minute of waiting, both forks were created OK. See the screenshot :)

I don't have any constructive advise, and the message you mention "an error occurred, try again" is quite cryptic. But at least for me, it worked OK. Aside from just trying again later, you could submit an issue to the GitLab support.

You can also check where do you create a fork. I tested forking into my default user namespace (michaliskambi) and into a group (michalis-private-tests). Maybe you're trying to create a fork into some namespace that has some limits.

Regards,
Michalis

niedz., 30 kwi 2023 o 18:00 Michael Van Canneyt via fpc-pascal napisał(a):

Michael Van Canneyt wrote

The repo is not private.

Ok, thanks. my assumption was obviously incorrect then.

One possible reason is that the repo is simply too big: it has 26 years of commits. We had serious problems importing it for this very reason.

Again, thanks. I'll attempt a different method.

Cheers,
Norm.

On Mon, May 1, 2023 at 4:24 PM Norman Dunbar via fpc-pascal <

FWIW, there are currently 68 forks of the FPC source repository on GitLab. I forked the repository shortly after it was created, at the time there were no issues. I do not expect the size of the repository to be a problem when creating a fork, there is probably some GitLab related issue.

Christo Crause wrote

FWIW, there are currently 68 forks of the FPC source repository on GitLab. ...

I noticed that when I tried to fork it. Unfortunately I am not able to fork those now either.

Admittedly, I didn't try to fork all of them though! ;-)

Cheers,
Norm.

At what time do you get the error?

I tested, first attempt I used a "project slug" I already had in use => immediate error.
(And you may also have to select the namespace)

2nd attempt went through. Took maybe half a fork started about 1 second after the click. The moment it started, the repo was also listed in my projects.

And, well yes, my account for testing was my account also used for fpc/laz. So I have full rights there. But that should not make any diff.

Norman Dunbar wrote

Any advice gratefully reeceived. Thanks.

I just tried, and had no problems. It took about 40 seconds to clone and checkout the `main` branch. The download seemed to be just over 200MB.

----------------------------------------------------------------

git clone

Cloning into 'FPCSource'...
remote: Enumerating objects: 718390, done.
remote: Counting objects: 100% (363/363), done.
remote: Compressing objects: 100% (292/292), done. remote: Total 718390 (delta 114), reused 320 (delta 71), pack-reused 718027 Receiving objects: 100% (718390/718390), 211.33 MiB | 6.32 MiB/s, done. Resolving deltas: 100% (575104/575104), done.

----------------------------------------------------------------

Regards,
Graeme

Michael Van Canneyt wrote

One possible reason is that the repo is simply too big: it has 26 years of commits.

I can't imagine that being the reason. I just did a fresh clone, and it was just over 200MB download, and the commit count on the main branch is less than 70k.

git rev-list --count main

66732

I've seen repos way larger (megabytes) and with vastly more commits - nearing 500k. There were no problems with those repos. The number of branches will also not make much of a difference - branching is cheap in Git.

I guess the question is more, how did the OP try to clone the repo?

- via Github web interface as a logged-in user
- directly to local PC, then the questions are:
- what is your internet download speed
- do you have a firewall interfering with traffic
- did you use the HTTP protocol or GIT+SSH protocol
- Can the OP post the exact clone command and error output

Regards,
Graeme

Graeme Geldenhuys wrote

I just tried, and had no problems. It took about 40 seconds to clone and checkout the `main` branch. The download seemed to be just over 200MB.

Just tried the Gitlab repo too - same result. No problems.

G.

Good Evening All.

Many thanks for all your replies. I have finally managed to form the GitLab repo. I have no idea what the problem was, but it occurred tonight on my first attempt, but my second attempt with the same settings, worked fine.

To answer your questions:

The error occurred when I clicked on the "Fork project" button, after filling in a name (FPC_Source) and choosing my own namespace to fork into.

I was using the GitLab fork option, logged in using my account.

Anyway, problem resolved, but cause remains a mystery. The error message from GitLab was of little help, simply "An error occurred, please try again later", followed by the "Fork project" button being unavailable and showing a busy cursor when mouse over, and the " Cancel" button being unavailable too. I had to refresh the page to get control back, and start again.

Thanks again.

Cheers,
Norm.