This is the mail archive of the
mailing list for the libstdc++ project.
Re: RFC: basic_filebuf supporting non-modal I/O without seekpos
>> On 09/19/2010 06:41 PM, Paolo Carlini wrote:
>>> if seeks would become redundant in order to switch from reads to writes and viceversa, I expect that after your draft patch fixing 45628 would be just matter of *simplifying* _M_seek, that would be great. Is it correct?
>> ... at least for the case that matters for mode switching that is, (0,
>> cur), or it can be treated again as part of seekoff and hacing open mode
>> in | out should not cause troubles anymore.
> I don't really follow.
> Having gotten more familiar with the code, I think I can make a slightly more elegant patch for 45628 (adjusting the __dummy_state variable), if you would like another submission. Hmm, looking at that patch, half of it seems to *require* the modality issue to be resolved.
I meant exactly this.
> Should I resubmit after this patch is applied? What's the plan?
The plan, since we now agree about the issue above, would be having an
unified patch. Note, the complete patch would incorporate an actual
testcase from 45628 (you can take mine, which you can find attached to
the PR, part of my small draft), and also a testcase actually testing
the new possibility to switch without seek in the middle, something very
simple, just reading and then writing, and closing or seeking to zero,
and checking that everything went well.
> I don't understand the linker script issue, or really what .tcc files do at all. Do we have to add entries to the baseline_symbols.txt for every platform? What's the impact?
No, no, no. I can do that, it's just that the linker script (gnu.ver)
has to be tweaked to *not* export the new symbols (2, one for char, one
for wchar_t) at the old, baseline version (abi_check is failing, thus is
means that an old pattern is matching the new symbol too,
inadvertently), and instead export it explicitly at the current version
(GLIBCXX_3.4.15). It's something boring the first time, I can do it if