#368 \foreach command fails with macros as argument

[Apostolos Syropoulos]

The following command does not produce the expected result (the command is defined in
pgffor.code.tex):

\def\mysequence{U,D,L,R,B,F}
\foreach\x in {\mysequence}{\fbox{\x},\ }

I have created a simple patch that solves the problem and of course the command

\foreach\x in {U,D,L,R,B,F}{\fbox{\x},\ }

continues to work as expected.

Please consider upadating the code of pgffor.code.tex

[percusse]

Hi Apostolos.

You have to use a single brace group for this usage, so you need to remove the brace around \mysequence

\foreach\x in \mysequence{\fbox{\x},\ }

[Apostolos Syropoulos]

Well I do not agree! Suppose one defines the following command:

\newcommand{\myspot}{U,D,L,R}

and uses this command to define the command that follows

\newcommand{\mypattern}{x,y,U,\myspot,L,R,y}

Then, \foreach\x in \mypattern{\fbox{\x},\ } will not yield
the expected result. And this is not some theoretical exercise.
Someone needs this facility in package that he is working on.
Thus this patch is the only way to make the \foreach command
work as it should. BTW, when one types

\foreach\x in {U,D,L,R,B,F}{\fbox{\x},\ }

it is more natural to type

\foreach\x in {\myspot}{\fbox{\x},\ }

instead of

\foreach\x in \myspot {\fbox{\x},\ }

[percusse]

Hi Apostolos,

Your analogies are not compatible, one is about stripping off braces to understand the context and the other one is expansion control. Braces are not throw-away containers, they actually mean something hence their usage is not about consistency. It is a TeX quirk but it is what it is. You cannot expect consistent syntax from an in-place replacement engine.

TikZ needs to see explicit characters (via \pgfutil@ifnextchar) to branch off hence your natural syntax will break other things and it is certainly not the only way as I have demonstrated. I don't know what the developers think but I don't agree that your patch is the way to go.

And finally, there is nothing natural about TeX.

[Apostolos Syropoulos]

To make this discussion sort: You simply do not want to understand and therefore you do not want to modify the code. It is your package, your work and finaly your idiosyncrasy.

[percusse]

Sorry that you feel that way but it is NOT my package. I'm a user and I tried to show you an alternative and one possible reasoning this is not a good idea because I understand what you are doing in the patch. There is no need to go sour if someone disagrees with you. You can wait until the developers respond instead.

[Apostolos Syropoulos]

I posted the initial message to the developers. Not to some public forum where people argue about everything. If you are not on of the developers, please wait until they step in and make a decision and if you happen to hate their decision you can fill in a bug report. Inaddition, I posted the patrch as a possible solution and not as the solution. And finally this is a bug and it would be a good idea to fix it. No one is obliged to fix it.

[percusse]

Nice attitude from a person whose previous words were Thus this patch is
the only way to make the \foreach command work as it should. Anyway, let's
not discuss further. There is no point.

[Apostolos Syropoulos]

That's the on;y thing I agree with you. BTW, it would be nice to know to whom I am talking! Hiding your identity is not a good thing to do.

[Henri Menke]

The proposed patch contradicts the documented behaviour. Close as wont-fix.