Hi, On Fri, 31 May 2024 17:57:26 +0800 Bo YU <tsu.y...@gmail.com> wrote: > Yeah, Now Debian official buildd machines are still Unmatached[0]: > We have many timeout issues on Debian packages on riscv64 > architectures. So I think we can increase timeout first to fix the > issue.
don't know whether this is an acceptable approach, but I'll point out
that the autotests could of course be disabled, completely. They don't
change anything in the build, and are most immediately useful in our CI.
> ```
> --- a/rkward/autotests/core_test.cpp
> +++ b/rkward/autotests/core_test.cpp
> @@ -57,7 +57,7 @@
> class RKWardCoreTest: public QObject {
> Q_OBJECT
>
> - void runCommandWithTimeout(RCommand *command, RCommandChain*
> chain, std::function<void(RCommand*)> callback, int timeoutms = 1000)
> {
> + void runCommandWithTimeout(RCommand *command, RCommandChain*
> chain, std::function<void(RCommand*)> callback, int timeoutms = 3000)
> {
> QString ccopy = command->command();
> QElapsedTimer t;
> t.start();
> ```
>
> Just is it enough to modify here?
No, that won't be enough. For one thing, the "1000" is only the default
parameter value, and may not apply. To increase all command timeouts,
see tline 68:
```
while (!done && (t.elapsed() < timeoutms)) {
```
you might want to write 3*timeoutms, here, for instance.
The timeout that actually got hit in your case is a different one,
on line 102, however:
```
if (t.elapsed() > 40000) break;
```
you'd have to adjust that, separately.
Regards
Thomas
pgpdMaEeOfwNp.pgp
Description: OpenPGP digital signature
