[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

[agiorgio@us.ibm.com: Excessive size of s390 vim binary]



I just got this on the vim-dev list. If this is indeed so (strip does
not work on s390) does that mean Debian policy forces us to have 
large unstripped binaries on s390? 

Wichert.

----- Forwarded message from Anthony Giorgio <agiorgio@us.ibm.com> -----

From: "Anthony Giorgio" <agiorgio@us.ibm.com>
To: vim-dev@vim.org
Subject: Excessive size of s390 vim binary
Date: Wed, 19 Sep 2001 10:37:56 -0400

When vim is built on an IBM s390 mainframe, the size of the binary is much 
larger than it needs to be.  This is due to the symbolics information 
remaining within the binary, since it is compiled with the -g flag.  On a 
standard Unix/Linux system this is fine, as one can strip(1) a binary to 
remove the debugging info.  On an s390 system, strip is identical to 
touch(1), and cannot remove the symbols from the file.

Here's an excerpt from the strip man page:


| strip -- Remove unnecessary information from an executable file


  Format

  strip file

| Description

| On some UNIX systems, strip removes debug information from an 
executable.
| On z/OS, the debug information can only be removed by recompiling. strip
| does not modify the contents of any executable file--it is functionally
| equivalent to touch file.


We should change the default behavior of the Makefile to not generate the 
debugging information.

I propose that the makefile on s390 be altered to remove the -g flag from 
CFLAGS.  Replacing -g with -0   (zero) dramtically reduces the size of the 
binary.  It shrinks from 5853184 bytes to 2052096 bytes, a 65% reduction.

What do you all say?


On a different note, any optimizations I turn on (-1 or -2) for the 
compiler causes vim to commit a segmentation violation as soon as it is 
invoked.



Anthony Giorgio
DBX Developer
Phone: (845) 435-9115
Tie Line: 295-9115
Email: agiorgio@us.ibm.com



----- End forwarded message -----

-- 
  _________________________________________________________________
 /       Nothing is fool-proof to a sufficiently talented fool     \
| wichert@wiggy.net                   http://www.liacs.nl/~wichert/ |
| 1024D/2FA3BC2D 576E 100B 518D 2F16 36B0  2805 3CB8 9250 2FA3 BC2D |



Reply to: