On Friday, 30 April 2021 at 19:22:57 UTC, Andrei Alexandrescu wrote:
> I mentioned this in the past, apparently it hasn't been resolved.
>
> Tried to install a dev tree for dmd from complete scratch. So:
>
> $ curl https://raw.githubusercontent.com/dlang/tools/master/setup.sh
> >/tmp/setup.sh
> $ /tmp/setup.sh
>
> The script gets everything from github nice and fast but then fails to download http://downloads.dlang.org/releases/2.x/2.095.0/dmd.2.095.0.linux.tar.xz for some reason.
>
> Anyway, I then try to run:
>
> $ cd dmd
> $ make AUTO_BOOTSTRAP=1 -f posix.mak
>
> As mentioned in the title, the command fails and confusingly suggest using AUTO_BOOTSTRAP=1, which is already there.
>
> If the plan is to give up boostrapping, fine but let's eliminate the code. The way things are now we both pay for the technical debt and we don't profit from it.
>
> Of course it would be much better if bootstrapping actually worked.

Just tried in WSL2 like this at ~:

# First I just cleaned my home
sudo rm -rf *

# And made sure I had c++
apt install g++

# Then I just did this
wget https://raw.githubusercontent.com/dlang/tools/master/setup.sh && chmod +x setup.sh && ./setup.sh

Everything worked as expected

On 5/1/21 4:51 PM, Petar Kirov [ZombineDev] wrote:
> On Saturday, 1 May 2021 at 01:07:33 UTC, Andrei Alexandrescu wrote:
>>
>> WSL2 running Ubuntu under Windows.
>>
>> [..]
> 
> Something is definitely wrong with your computer.

Thanks to all who tested - the problem is indeed with WSL (or my installation thereof).

For some reason, network connectivity from within WSL (at least on my machine) becomes really bad after a while and I need to shutdown and restart WSL in order to restore speed. At one point I couldn't push to git anymore.

Because of that, the download performed by the installation script was not finished. When I tried to manually run make, the system was not provisioned properly.

There would be the argument that that code should belong in the makefile as opposed to the installation script (as it was before) but no matter.

Thanks!

On Sunday, 2 May 2021 at 11:49:06 UTC, Andrei Alexandrescu wrote:
> On 5/1/21 4:51 PM, Petar Kirov [ZombineDev] wrote:
>> [...]
>
> Thanks to all who tested - the problem is indeed with WSL (or my installation thereof).
>
> For some reason, network connectivity from within WSL (at least on my machine) becomes really bad after a while and I need to shutdown and restart WSL in order to restore speed. At one point I couldn't push to git anymore.
>
> Because of that, the download performed by the installation script was not finished. When I tried to manually run make, the system was not provisioned properly.
>
> There would be the argument that that code should belong in the makefile as opposed to the installation script (as it was before) but no matter.
>
> Thanks!

Are you running WSL1 or WSL2 btw?

https://docs.microsoft.com/en-us/windows/wsl/compare-versions

On 5/2/21 10:17 AM, Imperatorn wrote:
> On Sunday, 2 May 2021 at 11:49:06 UTC, Andrei Alexandrescu wrote:
>> On 5/1/21 4:51 PM, Petar Kirov [ZombineDev] wrote:
>>> [...]
>>
>> Thanks to all who tested - the problem is indeed with WSL (or my installation thereof).
>>
>> For some reason, network connectivity from within WSL (at least on my machine) becomes really bad after a while and I need to shutdown and restart WSL in order to restore speed. At one point I couldn't push to git anymore.
>>
>> Because of that, the download performed by the installation script was not finished. When I tried to manually run make, the system was not provisioned properly.
>>
>> There would be the argument that that code should belong in the makefile as opposed to the installation script (as it was before) but no matter.
>>
>> Thanks!
> 
> Are you running WSL1 or WSL2 btw?
> 
> https://docs.microsoft.com/en-us/windows/wsl/compare-versions

WSL2. I had no idea it has these odd behaviors. Thought what our unittests do is fairly common.