PDM 2005 notes
Past weekend a bunch of the PHP developers (Zeev, Dmitri, Andrei, Jani, Wez, Marcus, Rasmus and me) got together in Paris to discuss PHP 6. On our agenda were Unicode support, the items from Rasmus' wishlist and many other issues.
I finally had the time to finish the report of the discussions (as you can see we discussed quite a lot) and publish the notes .
Comments
Great job, nice list!
That's really interesting Derick! With the exception of dropping curly braces for strings I really love ALL the 63 decisions!
I'm really looking forward to the next major PHP version!
Thanks a lot for your work!
Any chance we can get the var->public alias backported to php4 and 5? Its honestly the largest source of non-BC in php5 code.
Nice to see the suggestion about allow_url_fopen.
Funnily enough, I suggested this last year, but it wasn't picked up.
While I think the list is welcome, but I find the lack of real progress on the PHP object model disheartening. There is a real need for things like true accessor support (rather than the __get/__set error handler stuff) and support for Mixins and other capabilities to make writing things like persistence, dependency injection and AOP really possible in PHP.
These are often discussed in the community, but PHP unlike Perl or Java really has not community process other that sending nagging email. Perhaps a community where this list and other requests could be vetted would show the real needs of the community and the toolbuilders are.
Just saw this: 4.2 Adding "goto".
I don't think you're serious about this?! I mean.. you open a batch file and sometime you loose it.. imagine browsing through a php script using "goto"... back and forth, back and forth...
i mean come on... we suppose you add NEW features to PHP.. Not Fortran features (and many others languages like Fortran)..
This is a really good target feature list for PHP6, and I have to agree just about everything here looks good for the future of PHP. My only concerns/questions were:
Re: 2.14 - removing support for dynamic break level; Does this mean "break 2;" will still work, but "break $var" won't? If so, that's OK with me, but if "break 2;" goes away then that would be mildly annoying, as it is occasionally a useful construct.
Re: 4.3 ifsetor() - As the reporter of http://bugs.php.net/24949 , this sounds good to me, but can the proposed "?:" construct be daisy-chained? (Ideally the answer would be yes). E.g. would this be legal: $foo = $_GET['foo'] ?: $_GET['bar'] ?: 42;
All the best, Nick.
@Nikos: If you actually would have read that full section 4.2, you would have known that our "goto" is not a real goto at all and it limits the amount of jumping. Also, you don't have to use it!
@Nick J: Re: 2.14, yes, only "break $level" will be removed, but not "break 2".
Re: 4.3, I don't see why it can not be daisy chained, so your construct would most likely work.
> Any chance we can get the var->public alias backported to php4 and 5
One more vote for this (I really wondered for a very long time what was the benefit intended in raising an E_STRICT on that...)
Plus also backport separation of allow_url_fopen and allow_url_include for extra security.
Thanks a lot for your work!
@Gaetano: The alias will not be backported to PHP 5, but if you ask Ilia nicely, we might do it for PHP 5.1.1.
I have read the whole paragraph 4.2... I did noticed that it will be "not so widely used" but still i don't think it's a good idea.
I know i don't have to use it. I don't think i am. But i do read a lot of other peoples code.. (usualy open souce projects)
Anyway thanks for your reply, Derick.
I've read that Safe Mode will be removed from PHP 6. Safe Mode, as such, is very annoying indeed. But why not improve it rather than totally remove it ? The main problem, as you wrote it, is that Safe Mode doesn't give file ownership to the owner of the script which creates this file. If this problem could be fix, Safe Mode would become a convenient feature, wouldn't it ? Is it so difficult to fix that issue ?
Life Line
I've finished reading Children of Memory, the third book in the series.
Another interesting take on forms of intelligent life.
A fourth one is going to get released later this year.
Updated a post_box, a beauty shop, and a restaurant; Confirmed 2 clothes shops, 2 pet shops, and a restaurant
I walked 5.9km in 1h40m39s
Updated a bicycle_parking
Updated 2 waste_baskets
I walked 7.9km in 1h37m12s
Created 3 waste_baskets; Updated 3 bus_stops, 2 benches, and 2 waste_baskets
I walked 8.1km in 1h25m53s
I walked 1.2km in 9m31s
I walked 9.4km in 1h39m05s
Merge branch 'xdebug_3_5'
Merged pull request #1071
Fixed issue #2411: Native Path Mapping is not applied to the initial …
Created 2 waste_baskets; Updated 3 waste_baskets, 2 benches, and 2 other objects; Deleted a waste_basket
I walked 7.9km in 1h45m36s
RE: https://phpc.social/@phpc_tv/116274041642323081
Now that phpc.tv and phpc.social are part of the same umbrella, I've upped my yearly contributions to their Open Collective: https://opencollective.com/phpcommunity/projects/phpc-social
Merge branch 'xdebug_3_5'
Merged pull request #1070
I walked 7.2km in 1h10m26s
Fixed issue #2405: Handle minimum path in .xdebug directory discovery
I've published a new blog post: "Human Creations", on the difference in content generation by LLMs, and the creation of text, art and code by humans.
You can find it at https://derickrethans.nl/human-creations.html or at @blog
I walked 7.8km in 1h38m32s
RE: https://phpc.social/@afilina/116274024588235234
It's good to see that more and more people are realising that the Web can be for-good, without all the enshittification.
That's why I'm happy to see endeavours like phpc.tv springing up, and helping out where I can.
Taking back the control of how the Web is for people, by people, without big tech making it all shit.
Created a waste_basket; Updated 5 crossings and a bicycle_parking
I walked 10.7km in 2h35m10s


Shortlink
This article has a short URL available: https://drck.me/pdm-2005-notes-480