bug#45146: [PATCH] scripts: substitute: Improve fetch-narinfos progress reporting.

[Christopher Baines <mail@cbaines.net>]

[-- Attachment #1: Type: text/plain, Size: 2271 bytes --]


Christopher Baines <mail@cbaines.net> writes:

> Christopher Baines <mail@cbaines.net> writes:
>
>> Ludovic Courtès <ludo@gnu.org> writes:
>>
>>> Christopher Baines <mail@cbaines.net> skribis:
>>>
>>>> At least in guix weather, these changes make the progress bar actually appear.
>>>>
>>>> * guix/scripts/substitute.scm (fetch-narinfos): Use (guix progress) for
>>>> progress reporting.
>>>
>>> Cool.  I noticed that something was wrong with ‘guix weather’, but I
>>> suspected it had to do with the order in which the erase-line sequence
>>> and \r are sent.
>>>
>>>> -      (lambda ()
>>>> -        (display "\r\x1b[K" (current-error-port)) ;erase current line
>>>> -        (force-output (current-error-port))
>>>> -        (format (current-error-port)
>>>> -                (G_ "updating substitutes from '~a'... ~5,1f%")
>>>> -                url (* 100. (/ done total)))
>>>> -        (set! done (+ 1 done)))))
>>>> +  (define fetch-narinfos-progress-reporter
>>>> +    (progress-reporter/bar (length paths)))
>>>
>>> The problem here is that we’d see a progress bar without knowing what it
>>> represents.
>>>
>>> Besides, currently output from ‘guix substitute’ is printed as is by
>>> client commands, regardless of whether stdout is a tty.  The problem
>>> already exists but it would become a bit more visible as logs get filled
>>> with progress bars.
>>
>> Maybe it's best to circle back to fixing guix weather after trying to
>> restructure some of the guix substitute code.
>>
>> I've made an initial attempt at moving things around in [1]. If the
>> underlying code can live in a module, and then the substitute, weather
>> and challenge scripts use that code with whatever UI stuff they want,
>> maybe that will allow for better addressing this weather specific issue.
>>
>> 1: https://issues.guix.info/45409
>
> I've sent a couple of updated patches for this. They're not particularly
> dependent on the above work to create the (guix substitutes) module, but
> I based the commits on that.
>
> These commits make more careful changes, the substitute script behaviour
> should remain the same.

I've gone ahead and pushed these patches now as
f5ffb3bd9cd59cfff5b4625d6d65e8d116d0a25b.

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 987 bytes --]