[PATCH] gdb_ari.sh cleanup

[PATCH] gdb_ari.sh cleanup

[Pierre Muller]

I decided to go along recommendations of
removing all old macros in gdb_ari.sh.

I removed all macros that only appear in 
ChangeLog entries other than current ChangeLog
and ChangeLog-2008.


I did not handle two macros, because
they are unused but are still present in docs:
REGISTER_U_ADDR
PROCESS_LINENUMBER_HOOK


Miscellaneous questions:
1a) Should the PARAMS rule be moved to code section?
1b) Same question for __FUNCTION__ rule.
1c) Idem for ARGSUSED
1d) Idem for
2) LITTLE_ENDIAN and BIG_ENDIAN still exists in configure,
should I still remove the rule?
3) HAVE_VFORK is still present in config.in
should I keep the rule or should we remove it from config.in first?


I send here only the list of
BEGIN lines of the rules that I removed:


Pierre Muller
Pascal language support maintainer for GDB


< BEGIN { doc["FETCH_INFERIOR_REGISTERS"] = "\
< BEGIN { doc["CANNOT_FETCH_REGISTER"] = "\
< BEGIN { doc["CHILD_(|POST_)WAIT"] = "\
< BEGIN { doc["CHILD_POST_ATTACH"] = "\
< BEGIN { doc["CHILD_THREAD_ALIVE"] = "\
< BEGIN { doc["CHILD_PREPARE_TO_STORE"] = "\
< BEGIN { doc["CHILD_POST_STARTUP_INFERIOR"] = "\
< BEGIN { doc["CHILD_ACKNOWLEDGE_CREATED_INFERIOR"] = "\
< BEGIN { doc["CHILD_(INSERT|REMOVE)_FORK_CATCHPOINT"] = "\
< BEGIN { doc["CHILD_(INSERT|REMOVE)_VFORK_CATCHPOINT"] = "\
< BEGIN { doc["CHILD_(INSERT|REMOVE)_EXEC_CATCHPOINT"] = "\
< BEGIN { doc["CHILD_FOLLOW_FORK"] = "\
< BEGIN { doc["CHILD_REPORTED_EXEC_EVENTS_PER_EXEC_CALL"] = "\
< BEGIN { doc["CHILD_HAS_EXITED"] = "\
< BEGIN {
doc["CHILD_(ENABLE_EXCEPTION_CALLBACK|GET_CURRENT_EXCEPTION_EVENT)"] = "\
< BEGIN { doc["CHILD_PID_TO_(EXEC_FILE|STR)"] = "\
< BEGIN { doc["HAVE_OPTIONAL_PROC_FS"] = "\
< BEGIN { doc["VARIABLES_INSIDE_BLOCK"] = "\
< BEGIN { doc["TEXT_SEGMENT_BASE"] = "\
< BEGIN { doc["STATIC_TRANSFORM_NAME"] = "\
< BEGIN { doc["SOFUN_ADDRESS_MAYBE_MISSING"] = "\
< BEGIN { doc["SKIP_PERMANENT_BREAKPOINT"] = "\
< BEGIN { doc["SIGCONTEXT_REGISTER_ADDRESS"] = "\
< BEGIN { doc["SETUP_ARBITRARY_FRAME"] = "\
< BEGIN { doc["NUM_REALREGS"] = "\
< BEGIN { doc["LOWEST_PC"] = "\
< BEGIN { doc["IS_STATIC_TRANSFORM_NAME"] = "\
< BEGIN { doc["GDB_TARGET_IS_HPPA_20W"] = "\
< BEGIN { doc["CLEAR_SOLIB"] = "\
< BEGIN { doc["CHILD_SPECIAL_WAITSTATUS"] = "\
< BEGIN { doc["CHILD_GET_CURRENT_EXCEPTION_EVENT"] = "\
< BEGIN { doc["CHILD_ENABLE_EXCEPTION_CALLBACK"] = "\
< BEGIN { doc["HOST_BYTE_ORDER"] = "\
< BEGIN { doc["USG"] = "\
< BEGIN { doc["EXTRACT_RETURN_VALUE"] = "\
< BEGIN { doc["STORE_RETURN_VALUE"] = "\

Re: [PATCH] gdb_ari.sh cleanup

[Joel Brobecker]

> I did not handle two macros, because
> they are unused but are still present in docs:
> REGISTER_U_ADDR
> PROCESS_LINENUMBER_HOOK

Let's just remove them from the documentation. It's a very simple
change, but let me know if you'd like me to take care of it.

> Miscellaneous questions:
> 1a) Should the PARAMS rule be moved to code section?
> 1b) Same question for __FUNCTION__ rule.
> 1c) Idem for ARGSUSED

Does it really make a difference? If it helps you analyze the results,
then I'd say go for it.

> 1d) Idem for

(name missing?)

> 2) LITTLE_ENDIAN and BIG_ENDIAN still exists in configure,
> should I still remove the rule?

I don't think so. I don't know what the details are, but I'm wondering
whether the macros might be defined by the compiler, thus making it
possible for us to accidently reintroduce this usage again.  I'd say,
let's keep the rule.

> 3) HAVE_VFORK is still present in config.in
> should I keep the rule or should we remove it from config.in first?

We need to keep the rule. That macro is still used by gdb_vfork.h
and it is a valid use of that macro. In a way, this is very similar
to the use of "abort" - only very selected uses are allowed.

-- 
Joel

RE: [PATCH] gdb_ari.sh cleanup

[Pierre Muller]

Hi Joel,

> -----Message d'origine-----
> De : gdb-patches-owner@sourceware.org [mailto:gdb-patches-
> owner@sourceware.org] De la part de Joel Brobecker
> Envoyé : Friday, April 17, 2009 2:19 AM
> À : Pierre Muller
> Cc : gdb-patches@sourceware.org
> Objet : Re: [PATCH] gdb_ari.sh cleanup
> 
> > I did not handle two macros, because
> > they are unused but are still present in docs:
> > REGISTER_U_ADDR
> > PROCESS_LINENUMBER_HOOK

  Yes , please do it as it seems that 
at least for REGISTER_U_ADDR, it is not a 
node that you can just remove directly.

> Let's just remove them from the documentation. It's a very simple
> change, but let me know if you'd like me to take care of it.
> 
> > Miscellaneous questions:
> > 1a) Should the PARAMS rule be moved to code section?
> > 1b) Same question for __FUNCTION__ rule.
> > 1c) Idem for ARGSUSED
> 
> Does it really make a difference? If it helps you analyze the results,
> then I'd say go for it.
> 
> > 1d) Idem for

   Whoops: this was ATTRIBUTE_UNUSED
I will then move all those to ari_code type.
 
> (name missing?)
> 
> > 2) LITTLE_ENDIAN and BIG_ENDIAN still exists in configure,
> > should I still remove the rule?
> 
> I don't think so. I don't know what the details are, but I'm wondering
> whether the macros might be defined by the compiler, thus making it
> possible for us to accidently reintroduce this usage again.  I'd say,
> let's keep the rule.

  OK, I leave these one in.
 
> > 3) HAVE_VFORK is still present in config.in
> > should I keep the rule or should we remove it from config.in first?
> 
> We need to keep the rule. That macro is still used by gdb_vfork.h
> and it is a valid use of that macro. In a way, this is very similar
> to the use of "abort" - only very selected uses are allowed.

  OK, same here.


Thanks for the answers,


Pierre Muller
Pascal language support maintainer for GDB