Mathias LANG

Sounds like a good case for exponential backoff, no ?

> Did you have anything specific in mind when you wrote https://github.com/dlang/phobos/pull/8547#issuecomment-1229814392? I was thinking of doing what you did (`.byUTF` on the string) at the call-site, not to change...

``` ERROR: Newly generated header file (/media/disk1/braddr/sandbox/at-client/pull-5344065-Linux_32_64/dmd/generated/linux/release/64/frontend.h) doesn't match with the reference header file (/media/disk1/braddr/sandbox/at-client/pull-5344065-Linux_32_64/dmd/compiler/src/dmd/frontend.h) DETAILS: diff --git a/media/disk1/braddr/sandbox/at-client/pull-5344065-Linux_32_64/dmd/compiler/src/dmd/frontend.h b/media/disk1/braddr/sandbox/at-client/pull-5344065-Linux_32_64/dmd/generated/linux/release/64/frontend.h index 1626123..32353a1 100644 --- a/media/disk1/braddr/sandbox/at-client/pull-5344065-Linux_32_64/dmd/compiler/src/dmd/frontend.h +++ b/media/disk1/braddr/sandbox/at-client/pull-5344065-Linux_32_64/dmd/generated/linux/release/64/frontend.h @@ -628,7 +628,7...

Ping @maxhaton

Should it, really ? I don't see it as particularly problematic, and the solution is to add one FS access for every module we open, which is IMO pretty bad.

> See the bugzilla issue for rationale. Yeah that rationale is shaky. "If one version of Phobos is unzipped over another" isn't a great start, as it suffers from numerous...

Thanks for the great writeup! > Is my guess likely to be correct? And are there any other measures you can think of? Likely, although I'm not sure why you...

What's the goal here ?

This I need to review from a computer, but reducing the number of scans shouldn't be the goal. Eliminating the scan all together is IMO the way to go.

> Are CI failures unrelated? Why would they be unrelated ?