Skip to contentSkip to conversation

Problem with "already tagged token"

Cocci Rasmus Villemoes 3 participants 4 messages Sep 10, 2014 – Sep 11, 2014

lore
Share & tools
lore
#0 lore raw

Hi,

I'm trying to write a semantic patch to convert consecutive seq_puts() calls into a single one, with the string literals concatenated. My attempt below works fine when there are exactly two seq_puts() calls, but fails when there are more. If I use coccicheck on a directory, coccinelle seems to silently skip files containing three-or-more (e.g. kernel/trace/trace.c), but gives me the "already tagged token" error when I use it on the specific file.

Ideally, I'd of course like coccinelle to apply the spatch iteratively until only a single call remains. I think I understand why I get the error, but I'd like to know if there's some obvious fix I've overlooked.

@concat1 depends on patch@
expression s;
constant c1, c2;
position p1, p2;
@@
seq_puts at p1(s, c1);
seq_puts at p2(s, c2);

@script:python concat2@
c1 << concat1.c1;
c2 << concat1.c2;
c3;
@@

// The indentation probably needs to be fixed manually coccinelle.c3 = c1 + "\n\t" + c2

@concat3 depends on patch@
identifier concat2.c3;
expression concat1.s;
constant concat1.c1, concat1.c2;
position concat1.p1, concat1.p2;
@@
- seq_puts at p1(s, c1);
- seq_puts@p2(s, c2);
+ seq_puts(s, c3);

Thanks,
Rasmus

#1 lore raw

Clever :)

I haven't tried this, but I think you cudl do the following:

@r@
position p;
@@

seq_puts(...);
seq_puts(...);
seq_puts at p(...);

Quoted 4 lines

@concat1 depends on patch@
expression s;
constant c1, c2;
position p1, p2;

Now change this line to

position p1 != r.p, p2 != r.p;

Then you can either use Coccinelle to do the iteration, or just run the semantic patch over and over on the code.

julia

Quoted 33 lines

@@
seq_puts at p1(s, c1);
seq_puts at p2(s, c2);

@script:python concat2@
c1 << concat1.c1;
c2 << concat1.c2;
c3;
@@

// The indentation probably needs to be fixed manually coccinelle.c3 = c1 + "\n\t" + c2

@concat3 depends on patch@
identifier concat2.c3;
expression concat1.s;
constant concat1.c1, concat1.c2;
position concat1.p1, concat1.p2;
@@
- seq_puts at p1(s, c1);
- seq_puts at p2(s, c2);
+ seq_puts(s, c3);

Thanks,
Rasmus

_______________________________________________
Cocci mailing list
Cocci at systeme.lip6.fr
https://systeme.lip6.fr/mailman/listinfo/cocci

#2 lore raw
Quoted 17 lines · Julia Lawall

Clever :)

I haven't tried this, but I think you cudl do the following:

@r@
position p;
@@

seq_puts(...);
seq_puts(...);
seq_puts at p(...);

Quoted 4 lines

@concat1 depends on patch@
expression s;
constant c1, c2;
position p1, p2;

Quoted 7 lines

Now change this line to

position p1 != r.p, p2 != r.p;

Then you can either use Coccinelle to do the iteration, or just run the semantic patch over and over on the code.

Thanks, that seems to work. Can you tell me how I would get coccinelle to do the iteration? It's not really that important, since the whitespace has to be fixed manually anyway, and with your suggestion one finds all places with at least two consecutive seq_puts calls. But it would be nice if one would _only_ have to fix whitespace

Thanks,
Rasmus

#3 lore raw

[Cocci] Conversion of similar function calls into one with a longer parameter

In reply to Rasmus Villemoes
Quoted 4 lines

I'm trying to write a semantic patch to convert consecutive seq_puts() calls into a single one, with the string literals concatenated. My attempt below works fine when there are exactly two seq_puts() calls, but fails when there are more.

How do you think about to process the corresponding syntax and control flow graph by interfaces for the programming language "OCaml"?
Would it be safer to reuse the available API for such an application where several similar function calls should be merged together?

Regards,
Markus