Page MenuHomeFreeBSD

New Port: ftp/php72-fastdfs
ClosedPublic

Authored by joneum on Nov 16 2017, 7:53 PM.
Tags
None
Referenced Files
Unknown Object (File)
Feb 21 2024, 4:49 AM
Unknown Object (File)
Jan 4 2024, 10:53 AM
Unknown Object (File)
Dec 21 2023, 11:36 PM
Unknown Object (File)
Dec 20 2023, 2:35 AM
Unknown Object (File)
Dec 1 2023, 3:52 AM
Unknown Object (File)
Dec 1 2023, 3:52 AM
Unknown Object (File)
Dec 1 2023, 3:52 AM
Unknown Object (File)
Dec 1 2023, 3:52 AM
Subscribers

Details

Summary

New Port: ftp/php72-fastdfs

PHP module for accessing a FastDFS cluster

WWW: https://github.com/happyfish100/fastdfs

PR: 223714
Submitted by: Daniel Ylitalo <daniel@blodan.se> (maintainer)
Approved by: xxx (mentor)
#Differential Revision: https://reviews.freebsd.org/Dxxxxx

Test Plan

Diff Detail

Repository
rP FreeBSD ports repository
Lint
No Lint Coverage
Unit
No Test Coverage
Build Status
Buildable 13046
Build 13299: arc lint + arc unit

Event Timeline

ftp/php72-fastdfs/files/patch-config.m4
8 ↗(On Diff #35341)

instead of doing this, why not pass ROOT=$LOCALBASE to the configure script?
It only falls by to ROOT=/usr if $ROOT is not yet set.

Patching in a hard-coded /usr/local is also not ideal (you should use use the value of $LOCALBASE).

ftp/php72-fastdfs/files/patch-config.m4
8 ↗(On Diff #35341)

this is a copy of ftp/php71-fastdfs, ftp/php70-fastdfs, ftp/php56-fastdfs.
In all the other Ports, has decided to do it the same.
There is no reason to change this.

ftp/php72-fastdfs/files/patch-config.m4
8 ↗(On Diff #35341)

there's always a reason =)

if it is a copy of an existing port, make sure you dont simply svn add, but svn cp :)

ftp/php72-fastdfs/files/patch-config.m4
8 ↗(On Diff #35341)

Sorry, but no one has complained about years. Why you now?

it should be $PREFIX, not $LOCALBASE, as it is used to install the port, I think :)

Well, maybe no one looked at it? :) -- no one complaining does not mean you should not fix it once you see it :)

I think you can use

CONFIGURE_ENV= ROOT=${LOCALBASE}

instead of the patch.

As it is used in the following way

PHP_ADD_INCLUDE($ROOT/include/fastcommon)
PHP_ADD_INCLUDE($ROOT/include/fastdfs)

setting it to $LOCALBASE seems like the correct move

Also like with the other ports i highly doubt that this is a useful addition!
Did somebody evaluate a slave-port, using /etc/make.conf or a flavor? This would be nice instead of just adding the port.

ftp/php72-fastdfs/files/patch-config.m4
8 ↗(On Diff #35341)

Bad answer. Many things are not right but nobody complained. Lack of time, lack of understanding, lack of motivation and more. Choose one.

After update php*/fastdfs, here is an copy for the new php72 port.

svn cp php71-fastdfs/ php72-fastdfs/
A php72-fastdfs

#PHP7.2
[11.1-amd64-PHP72](http://joneumbox.org/data/11-1_amd64-PHP72-ports/2017-11-25_09h07m01s/logs/php72-fastdfs-5.0.11.lo()
11.1-i386-PHP72

Theoretically this is good to go. However, if you'll merge all the fastdfs ports into one soon, I'm not sure if there is much point in adding it :)

If "svn cp php71-fastdfs/ php72-fastdfs/" was correct, and there is nothing to complain about the code, I ask for release. thx :-)

If "svn cp php71-fastdfs/ php72-fastdfs/" was correct, and there is nothing to complain about the code, I ask for release. thx :-)

Technically the code is fine. Its just a bad solution. The correct one would be a version-free port which uses the PHP defined as default either in the ports-tree or in /etc/make.conf.
Also a slave-port would be a better variant - or a flavor.

Its not about getting a technical good solutions, its about getting a good solution. The code is just the implementation. It doesn't make much sense to add this port wrongly, because it just works while planning to do it correctly in the future.

I summarize:

  • 2 weeks ago it was not ok to have ftp/php56-fastdfs, ftp/php70-fastdfs and ftp/php71-fastdfs external patches. I changed it.
  • It was not okay that ftp/php72-fastdfs was created with "avn add". I changed it.
  • Well, after 2 weeks, after all changes, should everything be changed completely? Are you serious?

I'm closing ihere, telling the maintainer you're not releasing that after 2 weeks, and closing the PR too.

This revision is now accepted and ready to land.Nov 29 2017, 4:24 PM
  • It was not okay that ftp/php72-fastdfs was created with "avn add". I changed it.
  • Well, after 2 weeks, after all changes, should everything be changed completely? Are you serious?

Yes. It wasn't okay to have this ports in the first part. And i am consistent with myself, as you can see in some other PRs:
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=223715
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=216667

Also please note: "after all changes" its kind of bike shredding. I answered exactly one day after tcberners comments. That is not the same like "after 2 weeks, after all changes" - its exact one day, after various very helpful tips received by your main mentor the day before.

The problem here is that other committers doesn't had the insight of this problem. Which is fine - it happens to me in other parts multiple times and is still happening. But when a better solution is found, we should switch them.

Here the /etc/make.conf change would be the best advise to the users requesting the port, because it works in quarterly and head and did for many years.
Flavors are a great feature which will solve the problem of compiling everything by yourself, which is needed with the /etc/make.conf solution. But they will first available next year in quarterly.
I'm currently toying around with flavors and talking with portmgr how to incorporate them to solve this kind of problems.

Not correct. You did not say exactly what you would like to have!

You have done that only today, after 2 weeks!

Not correct. You did not say exactly what you would like to have!

You have done that only today, after 2 weeks!

I did. I wrote at 2017-11-17:

Also like with the other ports i highly doubt that this is a useful addition!
Did somebody evaluate a slave-port, using /etc/make.conf or a flavor? This would be nice instead of just adding the port.

There was no answer to my question. You just came up with another svn cp at 2017-11-25 and asked again at 2017-11-28 if its okay. While clearly never picking up my questions. Which is what i stated at 2017-11-29.

Even if you did miss the importance (to me) of my questions its not a good way you dealt with it. You focus on a lack of response for 2 weeks. Between my answers were 12 days and you wait 8 of them.
Also you just focused on the code to get it in. As committer it is your job to make it right, not to make it into the ports tree. If you think it a good idea to commit it anyway: start to discuss it. You basically ignored raised issues - not the first time.

And that all besides from persisting your frustration in a review forever, instead of just discuss it technically here or personal via email. Its okay for me, but i doubt its okay for you. ;)

Also i provided the correct way to enable the user to get what he wanted. You should have added this to the PR or add me to CC to enable me to do this.

They complain about missing answers, but do not answer themselves to open questions. What exactly is in the make.conf?
If you criticize something, you should do it better and not as bad.
You came, complain, without writing exactly how it is better, and are gone again. They have often done that. Not good.
Here I am waiting since Monday for an answer from you: D13235
Please only complain about something that makes you better yourself!
You can find the PR number here at the top ;-)

The maintainer is working on a master / slave setup. So you should be satisfied.

They complain about missing answers, but do not answer themselves to open questions. What exactly is in the make.conf?
If you criticize something, you should do it better and not as bad.

Please read what i wrote: you complained about missing answers from me without saying something about my raised issues. So you complain about yourself, not about me. ;)

Also as you know i'm not only a mentor but a teacher in real life. I don't provide full insight to give *you* the possibility to dig into it and learn from it. This helps much more than an long explanation (which besides is given in the link PRs).

Happy you finally ask for at least make.conf. That is for example in my make.conf to get this port:
cat /etc/make.conf
DEFAULT_VERSIONS= php=7.2

Same for every pecl port i need. It compiles the port against the defined PHP version. The php*-fastdfs ports don't work this way, because the use PHP_VER. Stripping the PHP_VER and use make.conf reduce it to the need of one port only.

Please do not just look for the error with me, but ask yourself if you have acted correctly. The answer should be "no".
Next I will not go into this discussion here

Please do not just look for the error with me, but ask yourself if you have acted correctly. The answer should be "no".
Next I will not go into this discussion here

I never stated that i did everything right. But i try to teach you were you did wrong. As a mentor its my job to learn for myself, but to teach you. I provided points, solutions and possibility for discussion you did not use. I provided feedback for your communication in the PR you did not use. I pointed to the possibility to talk beside this Review in private you refuse to use and just don't want to discuss further.
It is not the first time you prefer to point to your mentors instead of taking every change to learn. Please work on this attitude.

Please work on this setting.

I will gladly accept this advice, but you too should take your own advice. Work on your attitude.