WIP: start writing migration guide #3822
Conversation
fb35efa
to
d4bc815
Compare
|
@cmb69 Do you think it makes sense to add a new dedicated page for new class constants? Because at the moment it is all over the place in the UPGRADING document.... |
Covers section 10 of UPGRADING
Covers section 12 of UPGRADING
This should cover sections 3, 5, 9, 13, and 14 of UPGRADING
Cover section 8 of UPGRADING
Covers section 7 of UPGRADING
d4bc815
to
c44b06c
Compare
Yeah, not sure. Userland developers likely want to ignore some of the extensions, and may not even be interested in new class constants per se, but rather how they fit into a new feature. So the current structure makes some sense. Maybe make a list of the new constants while writing the pages, and then post it somewhere, so they can be documented right away (I know, more work for when writing the migration guide, and if nobody documents the constants, it not of much help.) |
Right, maybe we just need to reorganize how the UPGRADING document is structured, or split the "other changes" page into more distinct ones for functions, classes, and generic extension behaviour. |
This covers section 6 of UPGRADING
Better to ping this PR just to be sure. |
This covers section 4 of UPGRADING
ef23d46
to
3451d40
Compare
|
Yep that is in the guide as I refreshed UPGRADING before writing that section. :) |
This covers section 1 of UPGRADING
dad6ddf
to
f39024b
Compare
| <function>curl_multi_select</function> now throws a | ||
| <exceptionname>ValueError</exceptionname> if the | ||
| <parameter>timeout</parameter> parameter is less than | ||
| <literal>0</literal> or greater than <constant>PHP_INT_MAX</constant>. |
There was a problem hiding this comment.
| <literal>0</literal> or greater than <constant>PHP_INT_MAX</constant>. | |
| <literal>0</literal> or greater than <literal>2147483647</literal>. |
The range is a C int, which is 32 bits on all platforms I am aware of.
There was a problem hiding this comment.
It seems neither is correct, see https://github.com/php/php-src/pull/15594/files. Would need to be fixed in UPGRADING, too.
| to be consistent with the new | ||
| <constant>PHP_ROUND_<replaceable>*</replaceable></constant> modes. |
There was a problem hiding this comment.
These new mode constants do not exist: They have been replaced by the RoundingMode enum.
|
|
||
| <simpara> | ||
| It is now possible to fetch the value of the | ||
| <!-- TODO Is this a class constant of the driver class? --> |
There was a problem hiding this comment.
I believe it might even be a class constant on PDO by accident.
There was a problem hiding this comment.
The FB_ATTR_* are there for historic/BC reasons; we still have a lot more https://heap.space/search?project=PHP-8.4&full=&defs=&refs=REGISTER_PDO_CLASS_CONST_LONG&path=&hist=&type=.
| <constant>FB_ATTR_TIME_FORMAT</constant>, | ||
| <constant>FB_ATTR_TIMESTAMP_FORMAT</constant>, | ||
| attributes with | ||
| <!-- TODO Only for the specific driver class? --> |
There was a problem hiding this comment.
Well, yes, unless the attribute values overlap. There are however, new constants on Pdo\Firebird, see https://github.com/php/php-src/blob/PHP-8.4/ext/pdo_firebird/pdo_firebird.stub.php.
Co-authored-by: Niels Dossche <[email protected]>
Co-authored-by: Niels Dossche <[email protected]>
Co-authored-by: Larry Garfield <[email protected]>
Co-authored-by: Niels Dossche <[email protected]>
|
I might merge the PR as it currently is, this will not show-up online yet until doc-base is updated to include the new appendix |
Co-authored-by: Christoph M. Becker <[email protected]>
TODO:
UPGRADING Sections:
incompatibilities.xml)new-features.xml)deprecated.xml)new-functions.xml)new-classes.xml)