This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: (patch) pragma pack lossage
- To: Mumit Khan <khan at xraylith dot wisc dot EDU>
- Subject: Re: (patch) pragma pack lossage
- From: Jeffrey A Law <law at cygnus dot com>
- Date: Thu, 01 Jul 1999 01:57:12 -0600
- cc: egcs-patches at egcs dot cygnus dot com
- Reply-To: law at cygnus dot com
> My vote to get rid of this code. If anyone thinks this code does the
> right thing, I want to know why and definitely want to see some
> testcases.
This sound wrong. The model we generally use is to parse pragmas, then
internally turn them into attributes which are attached to the appropriate
tree node. It seems to me that removing that code is counter to the way
we want to handle pragmas.
I would rather see us *fix* the code rather than just rip it out.
> The second bug is simple: pragma pack(<n>) sets the *default* alignment
> to be <n>, and if we're popping out of the alignment stack, set the
> alignment to be the default, not to 0.
Please send this as a separate patch. You should know by now not to include
multiple changes in a single patchfile :-)