Oops, should've sent this to the list too.
----- Forwarded message from Alexander Schrijver alexander.schrijver@gmail.com -----
Date: Mon, 1 Nov 2010 13:28:59 +0100
From: Alexander Schrijver alexander.schrijver@gmail.com
To: James Butler james.butler@edigitalresearch.com
Subject: Re: [PHP-DEV] RE: [SPAM] Re: [PHP-DEV] rename T_PAAMAYIM_NEKUDOTAYIM to T_DOUBLE_COLON
User-Agent: Mutt/1.5.21 (2010-09-15)
-----Original Message-----
From: Alexander Schrijver [mailto:alexander.schrijver@gmail.com]
Sent: 01 November 2010 12:19
To: Stefan Marr
Cc: Dennis Haarbrink; Stan Vass; internals@lists.php.net
Subject: Re: [PHP-DEV] RE: [SPAM] Re: [PHP-DEV] renameT_PAAMAYIM_NEKUDOTAYIMtoT_DOUBLE_COLONIts a minor change and an annoyance to a lot of people. Yes, by not changing
this you'r annoying thousands of people.
Instead of going for this cosmetic nonsense you should help those people on the lemon branch.
I am insulted every time I have to read a parser token name in an error message, instead of a sensible error message.
The cost of understandingT_PAAMAYIM_NEKUDOTAYIMas part of the current mumbo-jumbo is completely insignificant compared to the cost of actually understanding the error message just indicating what the parser would have expected.Changing to lemon is the only way to actually achieve something in the long run...
Right, and be forced to introduce some bullshit hebrew when its done. No, thank you.
Err, the entire point is that it won't matter what the underlying token is. The error as seen can be anything you want it to be, or at least you can have a fight about what the new message looks like and i'm sure there won't really be a compelling reason for it to be in hebrew (unless localized).
Please grow up...
It's the policy:
There are two reasons this term will stay. It is a tip of the hat to
the amount of PHP work that came out of Israel, and it is a good
reminder that there are a lot of other languages in the world. People
whose first language is not English, myself included, are forced to work
with unfamiliar terms every day. I wouldn't mind having a few more
non-English identifiers in PHP actually.Well, and a third reason, I like it.
There is some reason this policy will change after i write this new tokenizer?
----- End forwarded message
Oops, should've sent this to the list too.
----- Forwarded message from Alexander Schrijver alexander.schrijver@gmail.com -----
Date: Mon, 1 Nov 2010 13:28:59 +0100
From: Alexander Schrijver alexander.schrijver@gmail.com
To: James Butler james.butler@edigitalresearch.com
Subject: Re: [PHP-DEV] RE: [SPAM] Re: [PHP-DEV] renameT_PAAMAYIM_NEKUDOTAYIMtoT_DOUBLE_COLON
User-Agent: Mutt/1.5.21 (2010-09-15)-----Original Message-----
From: Alexander Schrijver [mailto:alexander.schrijver@gmail.com]
Sent: 01 November 2010 12:19
To: Stefan Marr
Cc: Dennis Haarbrink; Stan Vass; internals@lists.php.net
Subject: Re: [PHP-DEV] RE: [SPAM] Re: [PHP-DEV] renameT_PAAMAYIM_NEKUDOTAYIMtoT_DOUBLE_COLONIts a minor change and an annoyance to a lot of people. Yes, by not changing
this you'r annoying thousands of people.
Instead of going for this cosmetic nonsense you should help those people on the lemon branch.
I am insulted every time I have to read a parser token name in an error message, instead of a sensible error message.
The cost of understandingT_PAAMAYIM_NEKUDOTAYIMas part of the current mumbo-jumbo is completely insignificant compared to the cost of actually understanding the error message just indicating what the parser would have expected.Changing to lemon is the only way to actually achieve something in the long run...
Right, and be forced to introduce some bullshit hebrew when its done. No, thank you.
Err, the entire point is that it won't matter what the underlying token is. The error as seen can be anything you want it to be, or at least you can have a fight about what the new message looks like and i'm sure there won't really be a compelling reason for it to be in hebrew (unless localized).
Please grow up...It's the policy:
There are two reasons this term will stay. It is a tip of the hat to
the amount of PHP work that came out of Israel, and it is a good
reminder that there are a lot of other languages in the world. People
whose first language is not English, myself included, are forced to work
with unfamiliar terms every day. I wouldn't mind having a few more
non-English identifiers in PHP actually.Well, and a third reason, I like it.
There is some reason this policy will change after i write this new tokenizer?
Yes, there is a reason:
As it was explained before, lemon would not display token names but
actual token "values". So instead of "Unexpected T_PAABLAH" it would say
"Unexpected '::' ..."
Best,
----- End forwarded message -----
--
--
Etienne Kneuss
http://www.colder.ch
It's the policy:
There are two reasons this term will stay. It is a tip of the hat to
the amount of PHP work that came out of Israel, and it is a good
reminder that there are a lot of other languages in the world. People
whose first language is not English, myself included, are forced to work
with unfamiliar terms every day. I wouldn't mind having a few more
non-English identifiers in PHP actually.Well, and a third reason, I like it.
There is some reason this policy will change after i write this new tokenizer?
Yes, there is a reason:
As it was explained before, lemon would not display token names but
actual token "values". So instead of "Unexpected T_PAABLAH" it would say
"Unexpected '::' ..."
But the lesson Rasmus was telling us about would go away. Yet, this is one of
the reasons the token is being kept. I am confused. Are you telling me this is
a lesson for the programmers to be learned? Not for the users?
On Mon, Nov 1, 2010 at 1:43 PM, Alexander Schrijver <
alexander.schrijver@gmail.com> wrote:
It's the policy:
There are two reasons this term will stay. It is a tip of the hat to
the amount of PHP work that came out of Israel, and it is a good
reminder that there are a lot of other languages in the world.
People
whose first language is not English, myself included, are forced to
work
with unfamiliar terms every day. I wouldn't mind having a few more
non-English identifiers in PHP actually.Well, and a third reason, I like it.
There is some reason this policy will change after i write this new
tokenizer?Yes, there is a reason:
As it was explained before, lemon would not display token names but
actual token "values". So instead of "Unexpected T_PAABLAH" it would say
"Unexpected '::' ..."But the lesson Rasmus was telling us about would go away. Yet, this is one
of
the reasons the token is being kept. I am confused. Are you telling me this
is
a lesson for the programmers to be learned? Not for the users?--
I"ve just brought up in the original thread, maybe we should go back there.
Tyrael
It's the policy:
There are two reasons this term will stay. It is a tip of the hat to
the amount of PHP work that came out of Israel, and it is a good
reminder that there are a lot of other languages in the world. People
whose first language is not English, myself included, are forced to work
with unfamiliar terms every day. I wouldn't mind having a few more
non-English identifiers in PHP actually.Well, and a third reason, I like it.
There is some reason this policy will change after i write this new tokenizer?
Yes, there is a reason:
As it was explained before, lemon would not display token names but
actual token "values". So instead of "Unexpected T_PAABLAH" it would say
"Unexpected '::' ..."But the lesson Rasmus was telling us about would go away. Yet, this is one of
the reasons the token is being kept. I am confused. Are you telling me this is
a lesson for the programmers to be learned? Not for the users?
I believe that what Rasmus meant is that a simple renaming of this
token was not justified.
I don't think that he would be against a parser change that would
bring much more to the table, solely because it would make this gem
disappear (at least I hope).
But then again, even though Felipe did an amazing job with this lemon
switch, few problems still prevent this change from happening in a near
future.
Best,
--
Etienne Kneuss
http://www.colder.ch
Yes, there is a reason:
As it was explained before, lemon would not display token names but
actual token "values". So instead of "Unexpected T_PAABLAH" it would say
"Unexpected '::' ..."
hello,
value of some tokens is not what would be expected either. Think a bit about
T_STRING for example.
There should be a smart algorithm implemented.
BTW bison-createdparsercan use a callback (yytnamerr) to replace tokens with
something human-readable. Why not use it?