fix time-1.4.0

[braingram]

The reorganization in #480 broke a few time formats since I missed that in the time schema in addition to the oneOf in base_format:

asdf-standard/resources/schemas/stsci.edu/asdf/time/time-1.4.0.yaml

Lines 239 to 241 in 92bd63c

	oneOf:
	- $ref: "#/definitions/format"
	- $ref: "#/definitions/other_format"

there is an exclusive consideration of only the formats listed in format in format:

asdf-standard/resources/schemas/stsci.edu/asdf/time/time-1.4.0.yaml

Line 233 in 92bd63c

$ref: "#/definitions/format"

This fixes the broken formats (cxcsec, galexsec, unix_tai) by moving them from other_format to format.

This PR is tested in astropy/asdf-astropy#270 (by setting the asdf-standard pin in that PR to this PR). I will open an issue to discuss moving these "astronomy" schemas to asdf-astropy since this PR highlights the problematic relationship of these two packages (asdf-standard has the astronmy schemas but can't test them without asdf-astropy which needs the schemas to implement support). The "asdf-standard" downstream failure in that PR is another sign of the mess that is these two packages (the job is testing asdf-standard main after install the source branch for this PR).

[perrygreenfield]

In looking at the previous PR I was under the impression that the other_formats were never saved as such, but were indicators that the value stored was derived from them and that the base_format only indicates what it was derived from. But this isn't the case? How does base_format relate to the format?

[braingram]

In looking at the previous PR I was under the impression that the other_formats were never saved as such, but were indicators that the value stored was derived from them and that the base_format only indicates what it was derived from. But this isn't the case? How does base_format relate to the format?

It looks like base_format is always the Time.format:
https://github.com/astropy/asdf-astropy/blob/a53570f6fcb1312445a3986696e635d7e94096fe/asdf_astropy/converters/time/time.py#L21
(although it is only written if base_format != format)
(and format is only written if it's not "guessable").
So for isot the output is:

time: !time/time-1.4.0 2025-08-27T15:38:25.758

(no format, no base_format as this is guessable)
for byear the output is:

time: !time/time-1.4.0 {base_format: byear, value: B2025.656}

(i think because "B2025.656" is "guessable" as byear_str but the provided/round-tripped format is byear)
for byear_str the output is:

time: !time/time-1.4.0 B2025.656

for mjd the output is:

time: !time/time-1.4.0 {format: mjd, value: 60914.65168701704}

(format since this is not "guessable").
Perhaps this can be simplified to:

• always write the format
• list all formats supported by astropy

It also looks like there a bugs (not introduced by this or the previous PR) since for byear above the provided time does not round trip:

>> t
<Time object: scale='utc' format='byear' value=2025.65590263171>
>> asdf.loads(asdf.dumps({"time": t}))['time']
<Time object: scale='tt' format='byear' value=2025.656>

[perrygreenfield]

LGTM